This is going to amuse the cat callers to no end. But reality is
reality. Summary: I'm fleeing in terror from Linux. Which means I
won't be using CMUCL or SBCL after all. Which might cause me to ditch
Common Lisp, because I'm poor. I need to decide whether the lack of
UFFI in GNU CLISP and GNU Common Lisp is a dealbreaker or not.
It's been a week and a half, I've put a *lot* of hours of manual labor
into trying to get Ubuntu to work, and I still only have half a working
system. The major dealbreaker is having to learn how to recompile the
kernel to get my Wacom Graphire tablet mouse to work properly. This OS
is seriously brain dead, that people have to amass so much expertise to
make a device driver change. It says tons about what I can expect for
maintenance hassles in the future. Also, my Linux knowledgeable friend
says none of the other distros will be any better.
When I decided to switch from Windows to Linux, it was on the premise
that this would actually save me labor over using Cygwin or MinGW. I am
disturbed to find out how bad Linux actually is. Near as I can tell,
Linux is not some consistent OS that has come a long ways since my early
forays from 1993 to 1996. Rather, Linux is a gamble. It might work
fine for you, or it might royally suck. Here, "sucks" means "more work
than beating Cygwin or MinGW into shape." I thought it would be a good
trade. I am surprised how bad a trade it turned out to be.
I've also become deeply averse to large stacks of dependent packages.
This is common in open source land, and it forces one to get stuck on a
particular OS. It's also the reason why Longhorn is shipping so late.
Perhaps in time, both open source and Windows will evolve to be equally
stacked and equally bad.
I'm also realizing how irrelevant Linux is to shipping games. It isn't
"an extra platform where I might pick up a few sales." It is an
impossible platform. In terms of consumer friendliness, it is complete
junk, and incapable of being otherwise anytime soon. I can't even begin
to fathom taking the tech support calls on this monstrosity. It makes
the disparate world of Windows 3D drivers look like cake!
Plus Apple is moving to x86, and in time will do all the things that the
Linux world won't do. I believe Linux has lost its opportunity to
become relevant as a consumer desktop. It will continue to grow as a
server OS, with expert administrators always in tow.
So, I'm nukeing my Linux partition and not looking back. Good riddance.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
How I really feel about Ubuntu Linux:
http://www.redlandsfortnightly.org/images/baker00.jpg
> Summary: I'm fleeing in terror from Linux.
Yes, this is a natural reaction, Linux is the only modern OS worse than
Microsoft Windows. You might want to take a look at Mac OS X, both for
the server and the client. And if you are too poor to look in that
direction, there is a good free system, FreeBSD, which does not
necessarily have drivers for all the fancy hardware you need, but is
written and maintained well. Most free lisps run on it (actually, all
running on Linux I am aware of).
David
·············@gmail.com" <············@gmail.com> writes:
>
> > Summary: I'm fleeing in terror from Linux.
>
> Yes, this is a natural reaction, Linux is the only modern OS worse than
> Microsoft Windows. You might want to take a look at Mac OS X,
Nooooooooo! Don't send him to OS X.
Those of us who use and like the Mac don't need him coming to our
community.
Note to OP:
Don't try OS X, you won't like it.
--
Thomas A. Russ, USC/Information Sciences Institute
On 13 Jun 2005 09:15:27 -0700, <···@sevak.isi.edu> wrote:
>
>
> ·············@gmail.com" <············@gmail.com> writes:
>
>>
>> > Summary: I'm fleeing in terror from Linux.
>>
>> Yes, this is a natural reaction, Linux is the only modern OS worse than
>> Microsoft Windows. You might want to take a look at Mac OS X,
>
> Nooooooooo! Don't send him to OS X.
> Those of us who use and like the Mac don't need him coming to our
> community.
Don't worry, Brandon only can do "talk".
--
The LOOP construct is really neat, it's got a lot of knobs to turn!
Don't push the yellow one on the bottom.
"Brandon J. Van Every" <·····················@mycompanyname.com> writes:
>
> It's been a week and a half, I've put a *lot* of hours of manual labor
> into trying to get Ubuntu to work, and I still only have half a working
> system. The major dealbreaker is having to learn how to recompile the
> kernel to get my Wacom Graphire tablet mouse to work properly. This OS
> is seriously brain dead, that people have to amass so much expertise to
> make a device driver change. It says tons about what I can expect for
> maintenance hassles in the future. Also, my Linux knowledgeable friend
> says none of the other distros will be any better.
Why are you fooling around with Ubuntu? Just download and burn a copy
of Fedora Core 3, do a full install and be done with it. I'm sure
compiling an application doesn't bother you, why should compiling the
kernel- its little different.
> When I decided to switch from Windows to Linux, it was on the premise
> that this would actually save me labor over using Cygwin or MinGW. I am
> disturbed to find out how bad Linux actually is. Near as I can tell,
> Linux is not some consistent OS that has come a long ways since my early
> forays from 1993 to 1996. Rather, Linux is a gamble. It might work
> fine for you, or it might royally suck. Here, "sucks" means "more work
> than beating Cygwin or MinGW into shape." I thought it would be a good
> trade. I am surprised how bad a trade it turned out to be.
So you've gotten enough experience to really judge this or are you just
fussing because you're frustrated? If you want a consistent Unix OS,
then go onto ebay and buy a well-configured Ultra 10 and put Solaris on
it.
> I've also become deeply averse to large stacks of dependent
> packages. This is common in open source land, and it forces one to get
> stuck on a particular OS. It's also the reason why Longhorn is shipping
> so late. Perhaps in time, both open source and Windows will evolve to be
> equally stacked and equally bad.
Is there some reason why you feel you need to be dinking around with
packages?
> I'm also realizing how irrelevant Linux is to shipping games. It isn't
> "an extra platform where I might pick up a few sales." It is an
> impossible platform. In terms of consumer friendliness, it is complete
> junk, and incapable of being otherwise anytime soon. I can't even begin
> to fathom taking the tech support calls on this monstrosity. It makes
> the disparate world of Windows 3D drivers look like cake!
As is frequently said, Linux is experienced-friendly. "Consumer
friendly" wastes too much time once you're beyond newbie stage.
> Plus Apple is moving to x86, and in time will do all the things that the
> Linux world won't do. I believe Linux has lost its opportunity to
> become relevant as a consumer desktop. It will continue to grow as a
> server OS, with expert administrators always in tow.
You're kidding right? I've configured OS X as a server & gateway, and
making it work is MISERABLE. Its not as hard to manage as Windows but
the gui is constantly in the way and its extremely hard to configure
things and setup subsystems independently of the user interface. While
the config database thing is a great idea, the layouts and organization
of config data is comprehensively undocumented. And then we can talk
about the the complete lack of drivers for anything other than trendy
usb/firewire gizmos. Just you go ahead and try to configure the
necessary StartupItems to configure the firewall, start
fetchmail/exim/courier-imap and remount Windows shares that have
disappeared- its brutal. Heck, I can't even get a driver for my 10 year
old differential scsi card in OS X and even if I had the driver, I'm not
convinced that I can even use tar to talk to it. No doubt OS X is
appealing to some as a desktop- but its utility is pretty much limited
to that market.
Gregm
Greg Menke wrote:
> Why are you fooling around with Ubuntu? Just download and burn a copy
> of Fedora Core 3, do a full install and be done with it. I'm sure
> compiling an application doesn't bother you, why should compiling the
> kernel- its little different.
Ubuntu was one of the easiest installations I had ever done on my
working place at the university. I do not know how to install Windows
(haven't used it for 10 years) but I can say installing Ubuntu is damn
easy.
Okay, I performed a standard Ubuntu installation without any need to
tailor it to my needs. Downloaded the 650MB to my ibook in about 5
minutes (I have a fast connection at my university); made a ISO CD with
my burner in the ibook and after 30 minutes I had a running Ubuntu on
my stationary machine (Ubuntu detected all my networks).
However, often tailoring Linux is not that hard. On my old laptop there
was Windows installed by default. The modem of that laptop was an
integrated Windows only modem. Although, it was a no name modem but I
red by chance who the producers of the no-name modem might be. The guys
from that company then emailed me a driver for Linux and to my great
surprise the driver in the following worked and I dumped Windows
eventually because the modem then worked under SuSE Linux.
> No doubt OS X is
> appealing to some as a desktop- but its utility is pretty much limited
> to that market.
I use my ibook only in a standard setup. Nevertheless OS X has one of
the best graphical user interfaces and I do not know of any Linux
windows managers which comes close. The old Gnome themes looked a bit
similar.
Schneewittchen
PS:Btw: it would have been suprising if the original poster had started
learning CommonLisp eventually. However, learning Common Lisp will turn
out benefitial independet of Linux or Windows. Brandom, you should take
the chance.
F�rster vom Silberwald wrote:
>
> I use my ibook only in a standard setup. Nevertheless OS X has one of
> the best graphical user interfaces and I do not know of any Linux
> windows managers which comes close. The old Gnome themes looked a bit
> similar.
That's funny, because I feel the exact same way about the Ion window
manager. I find it so much nicer than everything else that I can't see
myself switching to an OS that does not allow me to use Ion. So I'm
"stuck" using one of free Unixes. Not that that's a bad thing...
-- MJF
"F�rster vom Silberwald" <··········@hotmail.com> writes:
> Greg Menke wrote:
>
> > No doubt OS X is
> > appealing to some as a desktop- but its utility is pretty much limited
> > to that market.
>
> I use my ibook only in a standard setup. Nevertheless OS X has one of
> the best graphical user interfaces and I do not know of any Linux
> windows managers which comes close. The old Gnome themes looked a bit
> similar.
To each their own I suppose- I don't use a gui for much other than
colorized xterms, emacs frames and mozilla. I like the window manager
to stay out of my face, so OS X, Gnome, KDE, Windows are all more or
less annoying as far as I'm concerned- windowmaker is the easiest to
deal with so far.
Now I will conceed that OS X is the first Apple OS that doesn't make me
want to throw the computer out the nearest window, so I agree it has
improved. OTOH, the user interface "niceness" of OS X is pretty much
limited to the gui, once you're down in the terminal window its quite
opaque and difficult to manage- especially compared to Solaris which has
excellent & well documented management/config utilities at the
command-line.
Gregm
Greg Menke wrote:
> To each their own I suppose- I don't use a gui for much other than
> colorized xterms, emacs frames and mozilla. I like the window manager
> to stay out of my face, so OS X, Gnome, KDE, Windows are all more or
> less annoying as far as I'm concerned- windowmaker is the easiest to
> deal with so far.
That's me two years ago :) Wmaker, sometimes fvwm(1, not 2), uwm,
sometimes enlightenment, but mostly just wmaker, emacs and xterms
(or rxvts) with vi.
> Now I will conceed that OS X is the first Apple OS that doesn't make me
> want to throw the computer out the nearest window, so I agree it has
> improved. OTOH, the user interface "niceness" of OS X is pretty much
> limited to the gui, once you're down in the terminal window its quite
> opaque and difficult to manage- especially compared to Solaris which has
> excellent & well documented management/config utilities at the
> command-line.
Why's that? I can only say that with my Mac for some reason I
didn't touch most Unix programs at all anymore, because I didn't
need to. But there's documentation and man-pages for everything.
And the GUI is really quite un-in-your-face (unlike Windows,
Gnome, KDE) and comfy.
--
Don't let school interfere with your education. -- Mark Twain
Ulrich Hobelmann <···········@web.de> writes:
> Greg Menke wrote:
> > Now I will conceed that OS X is the first Apple OS that doesn't make me
> > want to throw the computer out the nearest window, so I agree it has
> > improved. OTOH, the user interface "niceness" of OS X is pretty much
> > limited to the gui, once you're down in the terminal window its quite
> > opaque and difficult to manage- especially compared to Solaris which has
> > excellent & well documented management/config utilities at the
> > command-line.
>
> Why's that? I can only say that with my Mac for some reason I didn't
> touch most Unix programs at all anymore, because I didn't need to. But
> there's documentation and man-pages for everything.
No theres not- for one example not even google will cough up the details
of how startupitems are sequenced and configured. There are some
references from mac related online/paper magazines showing how they can
be set up, but nothing comprehensive and nothing at all about how to
approach debugging. There is also no documentation about how to
influence the unixy tasks that preceed startupitems, the user is
basically forced to grep files in /etc and do google searches and induce
what to modify. If you use OS X as a client OS or a very simplistic
server then there isn't much need for the command line- just like most
any other OS.
> And the GUI is really quite un-in-your-face (unlike Windows, Gnome, KDE)
> and comfy.
If you say so- all I see are more things eye candy which needs to be
ignored and more gui stuff that need to be clicked to make it go away
and no obvious way to get multiple desktops. Expose is pretty cool, but
when you have a bunch of stuff going in parallel its still too
cluttered.
Gregm
Greg Menke wrote:
> If you say so- all I see are more things eye candy which needs to be
> ignored and more gui stuff that need to be clicked to make it go away
> and no obvious way to get multiple desktops. Expose is pretty cool, but
> when you have a bunch of stuff going in parallel its still too
> cluttered.
Multiple desktops are a good point. There's some free tools
around for that (forgot the name, Workspace manager?), but in the
end I got used to the Mac way.
--
Don't let school interfere with your education. -- Mark Twain
Ulrich Hobelmann <···········@web.de> wrote:
> Greg Menke wrote:
> > If you say so- all I see are more things eye candy which needs to be
> > ignored and more gui stuff that need to be clicked to make it go away
> > and no obvious way to get multiple desktops. Expose is pretty cool, but
> > when you have a bunch of stuff going in parallel its still too
> > cluttered.
> Multiple desktops are a good point. There's some free tools
> around for that (forgot the name, Workspace manager?), but in the
> end I got used to the Mac way.
I've gotten used to it from using OS9 in publishing for so long. But
now that I think about it, it's pretty galling not to have the multiple
workspaces I used to get when on a Sun 15 years ago. I'm realizing
that, used to the mac way or not, there are still a lot of times that I
have to move windows around to get to what I want. It would be ideal to
have 3 or 4 workspaces configured for various common finder tasks and
switch with a key stroke. You can manage some of this in the single
space with appropriate scripts and key commands, but it would still be
helpful.
I'm going to have to check out that freeware.
Michael
Michael Sullivan wrote:
> I've gotten used to it from using OS9 in publishing for so long. But
> now that I think about it, it's pretty galling not to have the multiple
> workspaces I used to get when on a Sun 15 years ago. I'm realizing
> that, used to the mac way or not, there are still a lot of times that I
> have to move windows around to get to what I want. It would be ideal to
> have 3 or 4 workspaces configured for various common finder tasks and
> switch with a key stroke. You can manage some of this in the single
> space with appropriate scripts and key commands, but it would still be
> helpful.
>
> I'm going to have to check out that freeware.
I hope we are not talking at cross points. However, you are right
without virtual desktops Mac OSX would be unuseable (trash like
Windows).
But the good news: there is a cheap shareware called "Desktop Manager".
I have been using Desktop Manager's free beta version for some months
now and it is rock stable.
Schneewittchen
PS: Note: Desktop Manager creates indeed a virtual space quite to the
contrary of some Mac OSX pseudo Desktop Managers which only trick you
into the believing you are working on seperate spaces. Desktop Manager
works like the virtual desktops you know from Unix or Linux X11.
Ulrich Hobelmann <···········@web.de> writes:
> Greg Menke wrote:
> > If you say so- all I see are more things eye candy which needs to be
> > ignored and more gui stuff that need to be clicked to make it go away
> > and no obvious way to get multiple desktops. Expose is pretty cool, but
> > when you have a bunch of stuff going in parallel its still too
> > cluttered.
>
> Multiple desktops are a good point. There's some free tools around
> for that (forgot the name, Workspace manager?), but in the end I got
> used to the Mac way.
I've been using CodeTek Virtual Desktop. It allows for two desktops
in the free version, up to 100 in the shareware version, IIRC.
--
Rob St. Amant
http://www4.ncsu.edu/~stamant
Förster vom Silberwald wrote:
> Greg Menke wrote:
> > Why are you fooling around with Ubuntu? Just download and burn a copy
> > of Fedora Core 3, do a full install and be done with it. I'm sure
> > compiling an application doesn't bother you, why should compiling the
> > kernel- its little different.
>
> Ubuntu was one of the easiest installations I had ever done on my
> working place at the university. I do not know how to install Windows
> (haven't used it for 10 years) but I can say installing Ubuntu is damn
> easy.
>
> Okay, I performed a standard Ubuntu installation without any need to
> tailor it to my needs. Downloaded the 650MB to my ibook in about 5
> minutes (I have a fast connection at my university); made a ISO CD with
> my burner in the ibook and after 30 minutes I had a running Ubuntu on
> my stationary machine (Ubuntu detected all my networks).
>
> However, often tailoring Linux is not that hard. On my old laptop there
> was Windows installed by default. The modem of that laptop was an
> integrated Windows only modem. Although, it was a no name modem but I
> red by chance who the producers of the no-name modem might be. The guys
> from that company then emailed me a driver for Linux and to my great
> surprise the driver in the following worked and I dumped Windows
> eventually because the modem then worked under SuSE Linux.
My personal experience wasn't so perfect. Let's take the latest Ubuntu
release, from I think May 11. I think Ubuntu is a great project, and I
think you're absolutely correct about your personal experiences, but
there's crazy mainstream PR that Linux is simple to use... a lie.
(And let's relax the demand that my pathnames be lettercase-correct.)
* Had to configure my monitor's resolution in /etc/X11/xorg.conf,
telling it HorizSync and VertRefresh. There's supposedly a
monitor-resolution GUI thingie somewhere, but it didn't really work.
* Had to run pppoeconfig from the cmdline since its GUI thingie doesn't
do ASDL. pppeoconfig seems nice though, more solid than Win2k.
* /etc/network/interfaces doesn't permanently set one of my ethernet
cards, so I just set it manually on bootup. And the GUI thingie most
certainly does not work.
* If you try cutting/copying something from Emacs to a normal window,
it doesn't work. (But oddly enough, it works the other way around.)
* Samba, DHCP... I would like to see the suicide rate among sysadmins.
I think one huge barrier to scalability in Linux is how every config
file has some ersatz new textfile grammar. Why not an ersatz sexp one,
like XML?
"What? XML?!" I hear some Linux fan saying. Yes, it's true -- if your
brethren are too blind to adopt sexps, then take its inferior cousin
who's been locked up in the closet for 5 decades!
"Tayssir John Gabbour" <···········@yahoo.com> writes:
>I think one huge barrier to scalability in Linux is how every config
>file has some ersatz new textfile grammar. Why not an ersatz sexp one,
>like XML?
All the XML config files I've seen so far (and more applications are
moving towards XML, especially GUI-oriented ones) were basically
unreadable gibberish; I very much prefer the imperfect situation that
now exists, with formats being different but normally very easily to
grasp, and simply constructed, to having to rely on GUI tools to
decipher (and write) the config files.
mkb.
Matthias Buelow <···@incubus.de> writes:
> "Tayssir John Gabbour" <···········@yahoo.com> writes:
>
> >I think one huge barrier to scalability in Linux is how every config
> >file has some ersatz new textfile grammar. Why not an ersatz sexp one,
> >like XML?
>
> All the XML config files I've seen so far (and more applications are
> moving towards XML, especially GUI-oriented ones) were basically
> unreadable gibberish; I very much prefer the imperfect situation that
> now exists, with formats being different but normally very easily to
> grasp, and simply constructed, to having to rely on GUI tools to
> decipher (and write) the config files.
>
I agree. Never understood why everyone has such a problem ehre. While
there may be some slight differences in how various developers have
defined their config files, most of them end up just being pretty much
attribute=value pairs and pretty simple to work out by just looking at
the file.
This pre-occupation with XML is really beginning to get frustrating. I
can see some real advantages to XML, especially if we selected a
standard DTD for config files. We could have generic GUI programs
which could configure anyting based on just the DTD and the config
file. But no, instead, everyone is now defining their own DTDs, so we
end up with no real progress - the problem has just been moved form
the config file format to the DTD definition.
Tim
--
Tim Cross
The e-mail address on this message is FALSE (obviously!). My real e-mail is
to a company in Australia called rapttech and my login is tcross - if you
really need to send mail, you should be able to work it out!
Tim X <····@spamto.devnul.com> writes:
> Matthias Buelow <···@incubus.de> writes:
>
>> "Tayssir John Gabbour" <···········@yahoo.com> writes:
>>
>> >I think one huge barrier to scalability in Linux is how every config
>> >file has some ersatz new textfile grammar. Why not an ersatz sexp one,
>> >like XML?
>>
>> All the XML config files I've seen so far (and more applications are
>> moving towards XML, especially GUI-oriented ones) were basically
>> unreadable gibberish; I very much prefer the imperfect situation that
>> now exists, with formats being different but normally very easily to
>> grasp, and simply constructed, to having to rely on GUI tools to
>> decipher (and write) the config files.
>>
> I agree. Never understood why everyone has such a problem ehre. While
> there may be some slight differences in how various developers have
> defined their config files, most of them end up just being pretty much
> attribute=value pairs and pretty simple to work out by just looking at
> the file.
>
> This pre-occupation with XML is really beginning to get frustrating. I
> can see some real advantages to XML, especially if we selected a
> standard DTD for config files. We could have generic GUI programs
> which could configure anyting based on just the DTD and the config
> file. But no, instead, everyone is now defining their own DTDs, so we
> end up with no real progress - the problem has just been moved form
> the config file format to the DTD definition.
This *may* be possible to solve with XForms and (possibly) XML
Schemas, but chances are that these will be deemed to heavyweight.
--
Raymond Wiker Mail: ·············@fast.no
Senior Software Engineer Web: http://www.fast.no/
Fast Search & Transfer ASA Phone: +47 23 01 11 60
P.O. Box 1677 Vika Fax: +47 35 54 87 99
NO-0120 Oslo, NORWAY Mob: +47 48 01 11 60
Raymond Wiker <·············@fast.no> writes:
> Tim X <····@spamto.devnul.com> writes:
>
> > Matthias Buelow <···@incubus.de> writes:
> >
> >> "Tayssir John Gabbour" <···········@yahoo.com> writes:
> >>
> >> >I think one huge barrier to scalability in Linux is how every config
> >> >file has some ersatz new textfile grammar. Why not an ersatz sexp one,
> >> >like XML?
> >>
> >> All the XML config files I've seen so far (and more applications are
> >> moving towards XML, especially GUI-oriented ones) were basically
> >> unreadable gibberish; I very much prefer the imperfect situation that
> >> now exists, with formats being different but normally very easily to
> >> grasp, and simply constructed, to having to rely on GUI tools to
> >> decipher (and write) the config files.
> >>
> > I agree. Never understood why everyone has such a problem ehre. While
> > there may be some slight differences in how various developers have
> > defined their config files, most of them end up just being pretty much
> > attribute=value pairs and pretty simple to work out by just looking at
> > the file.
> >
> > This pre-occupation with XML is really beginning to get frustrating. I
> > can see some real advantages to XML, especially if we selected a
> > standard DTD for config files. We could have generic GUI programs
> > which could configure anyting based on just the DTD and the config
> > file. But no, instead, everyone is now defining their own DTDs, so we
> > end up with no real progress - the problem has just been moved form
> > the config file format to the DTD definition.
>
> This *may* be possible to solve with XForms and (possibly) XML
> Schemas, but chances are that these will be deemed to heavyweight.
>
I suspect you may be right. I looked into the xforms stuff a couple of
years ago, but have not checked back lately to see how it is
progressing. I may appear to be anti-XML, but this is not the
case. While XML can appear to be a bit "verbose", it is easy to
validate and relatively easy to parse. However, I don't like the
apparent growth in the belief it is a "magic bullet" which will solve
all our problems and the growth in management attitude which seems to
believe that if those magic 3 letters are included or associated with
the product, it must be good, is very frustrating.
Tim
--
Tim Cross
The e-mail address on this message is FALSE (obviously!). My real e-mail is
to a company in Australia called rapttech and my login is tcross - if you
really need to send mail, you should be able to work it out!
Tayssir John Gabbour wrote:
> Förster vom Silberwald wrote:
> > Greg Menke wrote:
> > > Why are you fooling around with Ubuntu? Just download and burn a copy
> > > of Fedora Core 3, do a full install and be done with it. I'm sure
> > > compiling an application doesn't bother you, why should compiling the
> > > kernel- its little different.
> >
> > Ubuntu was one of the easiest installations I had ever done on my
> > working place at the university. I do not know how to install Windows
> > (haven't used it for 10 years) but I can say installing Ubuntu is damn
> > easy.
>
> * If you try cutting/copying something from Emacs to a normal window,
> it doesn't work. (But oddly enough, it works the other way around.)
A very helpful fellow pointed out:
M-x clipboard-kill-ring-save
Tayssir
Tayssir John Gabbour wrote:
> Tayssir John Gabbour wrote:
>> F�rster vom Silberwald wrote:
>> > Greg Menke wrote:
>> > > Why are you fooling around with Ubuntu? Just download and burn a
>> > > copy
>> > > of Fedora Core 3, do a full install and be done with it. I'm sure
>> > > compiling an application doesn't bother you, why should compiling the
>> > > kernel- its little different.
>> >
>> > Ubuntu was one of the easiest installations I had ever done on my
>> > working place at the university. I do not know how to install Windows
>> > (haven't used it for 10 years) but I can say installing Ubuntu is damn
>> > easy.
>>
>> * If you try cutting/copying something from Emacs to a normal window,
>> it doesn't work. (But oddly enough, it works the other way around.)
>
> A very helpful fellow pointed out:
> M-x clipboard-kill-ring-save
I have (setq x-select-enable-clipboard t) in ~/.emacs, and I tried the above
command on top of that.
Selecting text in Xterm with the left button and then pasting THAT TEXT with
the middle button or C-y in Emacs still doesn't work. I can do with with
all other applications fine, just not with Emacs. Emacs seems to be a
bastard in the GNU/Linux world.
I never found out how to make this work, I quizzed people more proficient
with Emacs than me, they couldn't help me either.
BTW, has anyone done any studies on whether coding in Emacs or Vim is more
likely to cause the repetitive stress injury? Does frequently twisting your
wrists and fingers in unnatural ways contribute to or alleviate it?
You can rebind your keys if you don't like the standard bindings. If
you have a standard PC keyboard, you have that useless windows key
which can be bound to Super, which is a blank slate - nothing is bound
to it by default, and I don't know of any extensions or modes that use
it. So if you find yourself using something painful like Control-Alt-F
or Control-Alt-% you can rebind it to Super-f or Super-% (or Super-5,
for that matter). I've also swapped the paren keys with the bracket
keys, which is helpful for Lispers, and many people switch the
caps-lock key with control. After doing all of this, Emacs isn't so
painful anymore.
alex goldman wrote:
>
> BTW, has anyone done any studies on whether coding in Emacs or Vim is more
> likely to cause the repetitive stress injury? Does frequently twisting
> your wrists and fingers in unnatural ways contribute to or alleviate it?
I'm a writer so I use Emacs quite heavily and I've never had
repetitive stress injury. But a dental assistant one time told me that
people with hands shaped in a certain way are more prone to it--as she
was. Maybe I'm one of the lucky ones. Much of it also comes from
posture. I type with the keyboard in my lap, not on a desk. I'm under
the impression that typing upwards--like with the keyboard on a
desk--is a factor in bringing it on. Your arms should be parallel to
the floor.
I have had some problem with pain from using the mouse when I'm not
careful with how I angle my arm editing photos with GIMP. But this
probably is more from an old injury when I was chasing a thief and
slipped on some grease and fell at full speed on concrete on that
elbow.
Kind of a non-answer, I know.
I have read that it's claimed that Vi/Vim is less conducive to RSI
because you don't have to reach as far using it. But this was from a
story on Vi advocates, not a study. I don't know if the question has
even been studied. My own experience with Vim is so slight that I
couldn't have gotten RSI if I'd typed with the keyboard above my head.
Yesterday I read the claim that switching from QWERTY keyboards to
Dvorak keyboards can alleviate RSI because the fingers don't work as
hard. I don't know. Just throwing it out. I suppose if you have pain
it's worth a try. But the web pages I was looking at suggested the
Dvorak layout really wasn't friendly to either Emacs or Vi/Vim editing
because of the arrangement of the keys commonly used for editing. The
keys are arranged for English words, not editing commands.
Hope this helps some.
--Rod
--
Author of "Linux for Non-Geeks--Clear-eyed Answers for Practical Consumers"
and "Boring Stories from Uncle Rod." Both are available at
http://www.rodwriterpublishing.com/index.html.
To reply take the extra "o" out of the name.
For me the key is to remap the control key to be just left of the 'A'
key, rather than the lower left of the keyboard.
- Chris
···········@core.com" <··········@core.com> writes:
> alex goldman wrote:
>
> >
> > BTW, has anyone done any studies on whether coding in Emacs or Vim is more
> > likely to cause the repetitive stress injury? Does frequently twisting
> > your wrists and fingers in unnatural ways contribute to or alleviate it?
>
> I'm a writer so I use Emacs quite heavily and I've never had
> repetitive stress injury. But a dental assistant one time told me that
> people with hands shaped in a certain way are more prone to it--as she
> was. Maybe I'm one of the lucky ones. Much of it also comes from
> posture. I type with the keyboard in my lap, not on a desk. I'm under
> the impression that typing upwards--like with the keyboard on a
> desk--is a factor in bringing it on. Your arms should be parallel to
> the floor.
>
> I have had some problem with pain from using the mouse when I'm not
> careful with how I angle my arm editing photos with GIMP. But this
> probably is more from an old injury when I was chasing a thief and
> slipped on some grease and fell at full speed on concrete on that
> elbow.
>
> Kind of a non-answer, I know.
>
> I have read that it's claimed that Vi/Vim is less conducive to RSI
> because you don't have to reach as far using it. But this was from a
> story on Vi advocates, not a study. I don't know if the question has
> even been studied. My own experience with Vim is so slight that I
> couldn't have gotten RSI if I'd typed with the keyboard above my head.
>
> Yesterday I read the claim that switching from QWERTY keyboards to
> Dvorak keyboards can alleviate RSI because the fingers don't work as
> hard. I don't know. Just throwing it out. I suppose if you have pain
> it's worth a try. But the web pages I was looking at suggested the
> Dvorak layout really wasn't friendly to either Emacs or Vi/Vim editing
> because of the arrangement of the keys commonly used for editing. The
> keys are arranged for English words, not editing commands.
>
> Hope this helps some.
>
>
> --Rod
>
>
> --
> Author of "Linux for Non-Geeks--Clear-eyed Answers for Practical Consumers"
> and "Boring Stories from Uncle Rod." Both are available at
> http://www.rodwriterpublishing.com/index.html.
>
> To reply take the extra "o" out of the name.
--
(. .)
=ooO=(_)=Ooo========================
Chris McMahan | cmcmahan-at-one.net
====================================
This is an OpenPGP/MIME signed message (RFC 2440 and 3156)
--------------enigE6346280AC47165866E2DE27
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
<snip>
> Yesterday I read the claim that switching from QWERTY keyboards to
> Dvorak keyboards can alleviate RSI because the fingers don't work as
> hard. I don't know. Just throwing it out. I suppose if you have pain
> it's worth a try. But the web pages I was looking at suggested the
> Dvorak layout really wasn't friendly to either Emacs or Vi/Vim editing
> because of the arrangement of the keys commonly used for editing. The
> keys are arranged for English words, not editing commands.
>
> Hope this helps some.
>
>
> --Rod
>
>
Presumably, if you were serious enough to learn to type on a completely
new keyboard layout, remapping the common commands would be a minor change.
I know that I was completely thrown just trying to code on a German
keyboard after using a UK one all my life because you had to use the
"Alt-Gr" key to get at the curly braces. AAArgh! :-)
Rupert
--------------enigE6346280AC47165866E2DE27
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: OpenPGP digital signature
Content-Disposition: attachment; filename="signature.asc"
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
iQCVAwUBQrCXm0bXf6SW2FaGAQEDsQP/TdMlqf0GEe9T8kOvIGZzEPWFK+P5rG3o
8+/q0sAwm5rTwcFoAQKZ3guum02jzJnm6KbnhzA2Gyk8BNuC1tHTEenqbWJQgr8p
sJQcgy2KiBueJk/K7DygQ2bohD0BGS3bhbvkU4Fnh/ThBoAMX0vimT6lqK4NDk+z
7gh9i8mYgRk=
=OMcs
-----END PGP SIGNATURE-----
--------------enigE6346280AC47165866E2DE27--
Rupert Swarbrick wrote:
> I know that I was completely thrown just trying to code on a German
> keyboard after using a UK one all my life because you had to use the
> "Alt-Gr" key to get at the curly braces. AAArgh! :-)
The German keyboard sucks quite badly even if you grew up with it.
Change it to US, Dvorak, whatever (is UK different from US??),
it's MUCH better for programming, and not any worse for typing German.
--
Don't let school interfere with your education. -- Mark Twain
From: Christopher C. Stacy
Subject: Re: I've thought better of Linux
Date:
Message-ID: <uk6kw0vw1.fsf@news.dtpq.com>
alex goldman <·····@spamm.er> writes:
> BTW, has anyone done any studies on whether coding in Emacs or Vim is more
> likely to cause the repetitive stress injury? Does frequently twisting your
> wrists and fingers in unnatural ways contribute to or alleviate it?
Empirically, I can say that constant use (about 10 hours a day),
every day, for the last quarter century, on a variety of keyboards,
some more suited for Emacs than others, has not caused any RSI in me.
On the other hand, using the mouse, and misaligning the height of the
keyboard with respect to my wrists, has occasionally been a problem.
I suppose this disproves any notion that "Emacs always causes RSI".
My friends who have NOT been using Emacs also do NOT have RSI.
I don't know of any studies about Emacs and RSI. YMMV.
From: Robert Uhl
Subject: Re: I've thought better of Linux
Date:
Message-ID: <m3y89bh938.fsf@4dv.net>
alex goldman <·····@spamm.er> writes:
>
> Selecting text in Xterm with the left button and then pasting THAT TEXT with
> the middle button or C-y in Emacs still doesn't work. I can do with with
> all other applications fine, just not with Emacs. Emacs seems to be a
> bastard in the GNU/Linux world.
Add:
(menu-bar-enable-clipboard)
to your .emacs, or run M-x menu-bar-enable-clipboard. No, it's not very
intuitive--but there's a reason for this: mapping between the kill ring
and X's notion of the current selection is not nearly as straightforward
as one might think.
> BTW, has anyone done any studies on whether coding in Emacs or Vim is
> more likely to cause the repetitive stress injury? Does frequently
> twisting your wrists and fingers in unnatural ways contribute to or
> alleviate it?
I dunno, but given that using emacs results in fewer total keystrokes
(since it can automate away a lot of common tasks), it's probably
better.
--
Robert Uhl <http://public.xdi.org/=ruhl>
Cooking is great--it's a socially acceptable excuse to play with knives
and fire.
Robert Uhl wrote:
>
> Add:
>
> (menu-bar-enable-clipboard)
>
> to your .emacs, or run M-x menu-bar-enable-clipboard. No, it's not very
> intuitive--but there's a reason for this: mapping between the kill ring
> and X's notion of the current selection is not nearly as straightforward
> as one might think.
>
Nope. Doesn't work :-(
From: Robert Uhl
Subject: Re: I've thought better of Linux
Date:
Message-ID: <m3psungzwk.fsf@4dv.net>
alex goldman <·····@spamm.er> writes:
> >
> > Add:
> >
> > (menu-bar-enable-clipboard)
> >
> > to your .emacs, or run M-x menu-bar-enable-clipboard. No, it's not
> > very intuitive--but there's a reason for this: mapping between the
> > kill ring and X's notion of the current selection is not nearly as
> > straightforward as one might think.
>
> Nope. Doesn't work :-(
Dunno, then. Unless you've another setting somewhere screwing with it.
default.el, I've discovered, is the spawn of the devil.
The only copy/paste issue I have with in X is that Unicode doesn't quite
work right, no matter how I invoke emacs's Unicode support.
--
Robert Uhl <http://public.xdi.org/=ruhl>
It's a beautiful hymn--at 5 mins. It looses something at 45.
--Megan, on the Hymn of Cassiana and Vain Cantors
Robert Uhl <·········@NOSPAMgmail.com> writes:
>> BTW, has anyone done any studies on whether coding in Emacs or Vim is
>> more likely to cause the repetitive stress injury? Does frequently
>> twisting your wrists and fingers in unnatural ways contribute to or
>> alleviate it?
>
>I dunno, but given that using emacs results in fewer total keystrokes
>(since it can automate away a lot of common tasks), it's probably
>better.
Which would these tasks be? For ordinary text editing, I find emacs
takes more typing effort than vi per se, since the commands are less
precise, and often more repetitive (given the lack of something
equivalent to the ex command set), and many bindings involve
combinations, thus straining hand muscles, compared to the usually
single-key bindings of vi (many of which are on the home
row). However, all together it doesn't really matter that much, unless
you're a secretary (or otherwise have to type and edit enormous
amounts of text in short time, which I normally don't do.) Much more
straining is having to reach for the mouse all the time, as is
commonplace in Windows (and Mac) development environments. When I do
an hour of "heavy mouse work", my right hand hurts and I have to
pause. Gladly, that doesn't happen often.
mkb.
Matthias Buelow <···@incubus.de> writes:
> Robert Uhl <·········@NOSPAMgmail.com> writes:
>
> >
> >I dunno, but given that using emacs results in fewer total keystrokes
> >(since it can automate away a lot of common tasks), it's probably
> >better.
>
> Which would these tasks be? For ordinary text editing, I find emacs
> takes more typing effort than vi per se, since the commands are less
> precise, and often more repetitive (given the lack of something
> equivalent to the ex command set), and many bindings involve
> combinations, thus straining hand muscles, compared to the usually
> single-key bindings of vi (many of which are on the home
I think you might be neglecting the part about vi where you also need a
speaker to differentiate between the "beep-and-mess-up-text" mode and
the plain "mess-up-text" mode. Emacs doesn't require the keyboard bell
and doesn't have any mess-up-text modes- which makes using emacs very
much less like a bad stand-up comedy act.
Although, maybe you're right, vi probably is easier because no-one can
figure out how to actually navigate the cursor so thinking about related
concepts isn't necessary, and its also a bit more fun than emacs because
you can never be sure if the next character you type is going to insert,
overwrite, begin execution of some unknown command that you can't get
out of, just beep or beep and mess up text....
;)
Gregm
Greg Menke <············@toadmail.com> writes:
>I think you might be neglecting the part about vi where you also need a
[...]
Wow! Better than Brandon could ever have said it! :)
mkb.
Matthias Buelow wrote:
> Greg Menke <············@toadmail.com> writes:
>
>
>>I think you might be neglecting the part about vi where you also need a
>
> [...]
>
> Wow! Better than Brandon could ever have said it! :)
If someone's going to cross-post this to comp.emacs, and completely
change the subject, when the subject was already marginally topical in
comp.lang.lisp, would someone mind terribly changing the SUBJECT line
appropriately as well? Note the followup.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
20% of the world is real.
80% is gobbledygook we make up inside our own heads.
Matthias Buelow <···@incubus.de> writes:
> Greg Menke <············@toadmail.com> writes:
>
> >I think you might be neglecting the part about vi where you also need a
> [...]
>
> Wow! Better than Brandon could ever have said it! :)
>
> mkb.
Well if we're going to be pissing and moaning about operating systems,
we might as well do it about editors too...
And why are we posting to comp.emacs when we have alt.religion.emacs?
Gregm
From: Robert Uhl
Subject: Re: I've thought better of Linux
Date:
Message-ID: <m3u0jzh00s.fsf@4dv.net>
Matthias Buelow <···@incubus.de> writes:
>
> > I dunno, but given that using emacs results in fewer total
> > keystrokes (since it can automate away a lot of common tasks), it's
> > probably better.
>
> Which would these tasks be?
Saving a file:
emacs C-x C-s (control-down, x, s, control-up; three-and-a-half strokes)
vi :w<RET> (shift-down, colon, shift-up, return; four strokes)
I give emacs a half-stroke benefit because the hands are idle as they
resume position to type, whereas with vi the hands aren't.
I've a function in emacs to prefix every line in the current region with
some string; it's used to indent things in straight next, and saves
innumerable keystrokes.
In emacs one isn't constantly hitting ESC.
emacs offers C-M-\ in lisp-mode, which re-indents the entire region;
AFAIK vi has no similar functionality.
Granted, vi is _excellent_ for editing config files (I'm a Unix admin by
trade, which no doubt says worlds about my programming skills), and
emacs is awful at that. But for programming, HTML editing, LaTeX
editing, reading mail, reading news, browsing the web and using IM,
emacs is much better. Granted that only the first three are text
editing:-)
> Much more straining is having to reach for the mouse all the time, as
> is commonplace in Windows (and Mac) development environments.
No doubt. I can really tell the difference between SSHing in and doing
everything in emacs and being on my home box running GNOME. Probably
need to get back to ion and doing everything from the keyboard.
--
Robert Uhl <http://public.xdi.org/=ruhl>
I don't jog. It makes the ice jump right out of my gin & tonic.
In comp.lang.lisp Robert Uhl <·········@nospamgmail.com> wrote:
>Saving a file:
>emacs C-x C-s (control-down, x, s, control-up; three-and-a-half strokes)
>vi :w<RET> (shift-down, colon, shift-up, return; four strokes)
Yeah but I can type :w<ret> faster than C-x C-s. That's probably
due to C-x C-s being both on the left hand. Maybe on dvorak that
would be different. But then again, dvorak doesn't seem to be all
that useful for either emacs, vi, or programming in general. And,
to continue in the absurd, Emacs' "C-x C-s C-x C-c" surely can't
beat vi's "ZZ".
>I've a function in emacs to prefix every line in the current region with
>some string; it's used to indent things in straight next, and saves
>innumerable keystrokes.
That's trivial to do in vi with mappings. And even without a mapping,
%s/^/foobar is probably fast to type. In fact, it's at these bulk
edits where vi rules (because of its ex subset) and emacs sucks. I
wonder why nobody has ever written an ex (or ed) mode for emacs
that operates on the last visited buffer, or something like that.
Would probably raise the usefulness of Emacs quite a bit.
>In emacs one isn't constantly hitting ESC.
Not really. Normally one types some amount of of text (in either
input or command mode) before one switches. In emacs, one is hitting
Control (or Alt/Meta) all the time. And instead of singular
keystrokes, it's mostly combinations, which are more effort (imho).
>emacs offers C-M-\ in lisp-mode, which re-indents the entire region;
>AFAIK vi has no similar functionality.
I won't comment on the original vi's Lisp mode, which is clearly
outdated and hasn't been updated since 1985. The vim clone uses the
"=" command for reindentation. It's to be combined with a move
command so can indent all regions that one can select with vi(m)
commands.
Note that I don't find these examples particularly significant.
Basically, these operations work well in both editors.
mkb.
[comp.emacs* removed from Newsgroups-line, for obvious reason.]
To prepend a string to the beginning of each line in a region, the
obvious method is to use rectangles, C-x r t prefix RET. If you really
want to use a regexp, do M-x replace-regexp RET ^ RET prefix RET. If
that's too much typing and you find yourself prepending strings very
often (I don't), you can define a new keybinding with a few lines of
Emacs lisp to do it in as little as one keystroke. I don't the best
way to do it in vi, because I never use it unless forced to by a
program like visudo, but I can't imagine it could be much more
efficient, unless vi reads your mind and writes the text for you.
Matthias Buelow <···@incubus.de> writes:
> In comp.lang.lisp Robert Uhl <·········@nospamgmail.com> wrote:
>
>>Saving a file:
>>emacs C-x C-s (control-down, x, s, control-up; three-and-a-half strokes)
>>vi :w<RET> (shift-down, colon, shift-up, return; four strokes)
>
> Yeah but I can type :w<ret> faster than C-x C-s. That's probably
> due to C-x C-s being both on the left hand. Maybe on dvorak that
C-x C-s is not "both on the left hand". <control>(right pinkie),
followed by "xs"(left ring). Three key strokes, touch typing.
David
--
David Trudgett
http://www.zeta.org.au/~wpower/
What I don't know is not as much of a problem
as what I am sure I know that just ain't so.
-- Mark Twain
alex goldman <·····@spamm.er> writes:
>Selecting text in Xterm with the left button and then pasting THAT TEXT with
>the middle button or C-y in Emacs still doesn't work. I can do with with
>all other applications fine, just not with Emacs. Emacs seems to be a
>bastard in the GNU/Linux world.
I've found it often doesn't work the other way round; for example,
pasting from Firefox into Emacs sometimes (but not always!) simply
doesn't seem to work; I have to go through the indirection through
xclipboard(1). That may also work in your case, have a look at the
xclipboard tool, which comes with X11. Sadly, cut&waste on X11 with
its several diverse cut buffers, and selection mechanisms is a really
bloody mess...
mkb.
Matthias Buelow <···@incubus.de> writes:
> alex goldman <·····@spamm.er> writes:
>
> >Selecting text in Xterm with the left button and then pasting THAT TEXT with
> >the middle button or C-y in Emacs still doesn't work. I can do with with
> >all other applications fine, just not with Emacs. Emacs seems to be a
> >bastard in the GNU/Linux world.
>
> I've found it often doesn't work the other way round; for example,
> pasting from Firefox into Emacs sometimes (but not always!) simply
> doesn't seem to work; I have to go through the indirection through
> xclipboard(1). That may also work in your case, have a look at the
> xclipboard tool, which comes with X11. Sadly, cut&waste on X11 with
> its several diverse cut buffers, and selection mechanisms is a really
> bloody mess...
>
Entering this thread late and obviously may have missed
something......
but I was just wondering what exactly is the problem so many seem to
be encountering with cut and paste between emacs and other
applications. In 10+ years of Linux use and 8+ years of emacs use,
I've never had a problem - well, with the exception of "staircasing"
when pasting into an exterm at one time and this turned out to be an
xterm config setting[1]. I regularly paste from xterm to emacs and emacs
to xterm as well as other applications. I only use the mouse within
non-emacs apps, preferring to use the keyboard within emacs to the
mouse. I don't have any specific settings under X, xterm or emacs
relating to cut and paste and currently am running Debian testing. My
locale is set to UTF-8, though until a few months ago I was using 'C'.
[1] From memory (it was many years ago), the 'staircase' problem was
related to newlines and pasting between an xterm and vi. From memory,
it was fixed by setting an xterm resource. However, I don't think this
was a Linux problem, rather an xterm/X problem and occured on a number
of different platforms I was using at the time (mainly DEC OSF/X)
Tim
--
Tim Cross
The e-mail address on this message is FALSE (obviously!). My real e-mail is
to a company in Australia called rapttech and my login is tcross - if you
really need to send mail, you should be able to work it out!
In article <··············@athena.pienet>,
Greg Menke <············@toadmail.com> wrote:
> You're kidding right? I've configured OS X as a server & gateway, and
> making it work is MISERABLE.
I have Mac OS X running as a server. Actually it is
Mac OS X 10.4.1 Server. Which is a product from Apple.
See here: http://www.apple.com/server/macosx/
Configuring it as a server
was quite easy. Most of the stuff can be done
easily with the server administration software.
Including setting up the firewall and adding
rules, etc.
I have it running on a Mac mini, where also my personal
CL-HTTP based web server runs.
> Its not as hard to manage as Windows but
> the gui is constantly in the way and its extremely hard to configure
> things and setup subsystems independently of the user interface. While
> the config database thing is a great idea, the layouts and organization
> of config data is comprehensively undocumented. And then we can talk
> about the the complete lack of drivers for anything other than trendy
> usb/firewire gizmos. Just you go ahead and try to configure the
> necessary StartupItems to configure the firewall, start
> fetchmail/exim/courier-imap and remount Windows shares that have
> disappeared- its brutal. Heck, I can't even get a driver for my 10 year
> old differential scsi card in OS X and even if I had the driver, I'm not
> convinced that I can even use tar to talk to it.
So, get something newer. I got rid of my old SCSI stuff and just bought
one of these LaCie Firewire 800 drives.
> No doubt OS X is
> appealing to some as a desktop- but its utility is pretty much limited
> to that market.
That's why there is Mac OS X client and server. Different offerings
for different markets.
Rainer Joswig <······@lisp.de> writes:
> In article <··············@athena.pienet>,
> Greg Menke <············@toadmail.com> wrote:
>
> Configuring it as a server
> was quite easy. Most of the stuff can be done
> easily with the server administration software.
> Including setting up the firewall and adding
> rules, etc.
>
> I have it running on a Mac mini, where also my personal
> CL-HTTP based web server runs.
OK, so how do you set up a customized firewall w/ NAT through a 2nd lan
board, with fetchmail/exim/courier-imap and smb remounts using the gui?
I'd like to know because the startupitem scripts are tedious. While
you're at it, I'd like to be able to configure dhcp- by that I mean
limiting interfaces and specifying dhcpd config items.
I know OS X can be a server, but compared to something like Solaris its
very difficult to configure and manage.
> > Its not as hard to manage as Windows but
> > the gui is constantly in the way and its extremely hard to configure
> > things and setup subsystems independently of the user interface. While
> > the config database thing is a great idea, the layouts and organization
> > of config data is comprehensively undocumented. And then we can talk
> > about the the complete lack of drivers for anything other than trendy
> > usb/firewire gizmos. Just you go ahead and try to configure the
> > necessary StartupItems to configure the firewall, start
> > fetchmail/exim/courier-imap and remount Windows shares that have
> > disappeared- its brutal. Heck, I can't even get a driver for my 10 year
> > old differential scsi card in OS X and even if I had the driver, I'm not
> > convinced that I can even use tar to talk to it.
>
> So, get something newer. I got rid of my old SCSI stuff and just bought
> one of these LaCie Firewire 800 drives.
No, thanks. I have a lot of scsi hardware around, the last thing I need
is some new soon-to-be-disappeared transport. Still looking for a tape
backup solution. At the moment I have an Ultra 1 using smbclient to tar
up portions of the mac's filesystem, backing it up onto a local tape
changer. If you can suggest a differential scsi board that OS X
supports, I'm all ears.
> > No doubt OS X is
> > appealing to some as a desktop- but its utility is pretty much limited
> > to that market.
>
> That's why there is Mac OS X client and server. Different offerings
> for different markets.
Does "OS X server" offer a significant difference beyond the gui?
Gregm
Greg Menke <············@toadmail.com> writes:
>
> OK, so how do you set up a customized firewall w/ NAT through a 2nd lan
> board,
I can't help you with this, as I don't use my OS X boxes as
routers.
> with fetchmail/exim/courier-imap and smb remounts using the gui?
I'm not sure what you're asking about here. I use fetchmail to
grab my mail for me and my other users from a remote MX host. It
gets stored locally on my OS X machine until the mail clients access
it. I use POP3S to grab it from the server and put it on my OS X
laptop, while my wife uses IMAP.
We use SMB for remote file access, because a) I didn't pay for OS
X server (which means I can't get AFP over SSH) and b) there were
permissions problems with using NFS (some apps that were ports of
traditional Mac OS apps don't check group permissions on files).
I wrote a small applescript to mount the SMB shares. I haven't
bothered to add it to startup items because I'm lazy, but doing so
wouldn't be difficult.
> I'd like to know because the startupitem scripts are tedious. While
> you're at it, I'd like to be able to configure dhcp- by that I mean
> limiting interfaces and specifying dhcpd config items.
I also use DHCP to hand out IPs/netmasks/default routers/DNS
servers to my clients. Some have hard-wired assignments based on
MAC address, others get from a pool open to transients.
I'm using an iBook G4 800 MHz laptop as my server. I added a small
number of init scripts to handle things like DHCP (these were
trivial to make by copying from other startup items.)
I also have an * PBX setup for VOIP running on it.
I did spend $10 to get a Tiger compatabile version of Postfix
enabler because I didn't want to have to fool around with migrating
my Postfix setup from scratch.
> I know OS X can be a server, but compared to something like Solaris its
> very difficult to configure and manage.
Sounds to me like you know Solaris better than OS X. No shock,
then, that you find it easier to use Solaris. Frankly, I had the
same experience when I first started using Linux back in 1994, as it
was nothing like the BSD-derived systems I was used to adminning
(Ultrix and OSF/1 from DEC). Moving to a system that doesn't use
the SYS-V style inits is a bit - well, awkward, since I've been so
heavily into using Linux for the last 10 years.
> Does "OS X server" offer a significant difference beyond the gui?
It has better support for some things that are either disabled by
default in the client version (like Postfix for mail) or aren't
available at all (like AFP over SSH). I don't have a complete list
of the differences because I've found using the client version to be
quite acceptable.
--jon
Jon Boone <········@delamancha.org> writes:
> Greg Menke <············@toadmail.com> writes:
>
> >
> > OK, so how do you set up a customized firewall w/ NAT through a 2nd lan
> > board,
>
> I can't help you with this, as I don't use my OS X boxes as
> routers.
Well once you exhaust the capabilities of the built-in gui controled
ipfw configuration and get into nat via a 2nd lan board, then you'll run
into a situation where you need some really bizarro startupitem scripts
to let the system boot, dhclient to the isp, re-hostname the host and
apply the firewall. Ugly ugly- particularly the part where messing
around with hostname at the wrong time essentially stalls the OS X boot
process and jams up all kinds of stuff (and its still the wrong time as
the very last startupitems execute). I had to kludge around it by
making the firewall config script fork off a 2nd script which sleeps 45
seconds or so before starting up the firewall and nat- allowing the OS
to finish booting to the desktop.
And worst of all, none of its directly debuggable because the gui is
omnipresent.
> > with fetchmail/exim/courier-imap and smb remounts using the gui?
>
> I'm not sure what you're asking about here. I use fetchmail to
> grab my mail for me and my other users from a remote MX host. It
> gets stored locally on my OS X machine until the mail clients access
> it. I use POP3S to grab it from the server and put it on my OS X
> laptop, while my wife uses IMAP.
Fetchmail to draw mail from a bunch of accounts, funneled into a maildir
format drop where courier-imap serves it out to multiple clients. The
OS X box hosts the software and the maildir. This was pretty
straightforward after working through some OS X ranlib weirdness. I'm
glad OS X ships with fetchmail as I've been unable to get it to compile
due to crazy .h stuff IIRC.
This OS X box is a workstation and server for a few Windows boxes, and
acts as the firewall & gatway out onto the cable modem.
What I'm getting at is the startupitem xml configuration business is
almost entirely undocumented except by google hits on a couple mac
fanzines and the subtlties of how OS X sequences the startupitems and
how the parameters are expressed are essentially undocumented in any
obvious place.
> I wrote a small applescript to mount the SMB shares. I haven't
> bothered to add it to startup items because I'm lazy, but doing so
> wouldn't be difficult.
What happens when the SMB mounts drop and Finder gets all tangled up,
unable to remount and reconnect- or when some smb mounts show up in the
finder and some don't? I still haven't figured out the latter...
> I also use DHCP to hand out IPs/netmasks/default routers/DNS
> servers to my clients. Some have hard-wired assignments based on
> MAC address, others get from a pool open to transients.
No, I mean how does one configure and control the builtin OS X dhcp
client- for instance I might want to keep a locally defined hostname,
not assume the fqdn from the isp's address. There is no executable or
config file that I've found.
> > I know OS X can be a server, but compared to something like Solaris its
> > very difficult to configure and manage.
OK, compare and contrast the startup items vs Solaris 10's new services
facility. Its like night and day. I spent quite a bit of time puzzling
out how to get startupitems to start properly, and ended up just setting
them to start last since the startupitem parameters are so inadequately
documented. Solaris has extensive, indexed and cross-referenced
documentation for nearly the whole OS available online.
> > Does "OS X server" offer a significant difference beyond the gui?
>
> It has better support for some things that are either disabled by
> default in the client version (like Postfix for mail) or aren't
> available at all (like AFP over SSH). I don't have a complete list
> of the differences because I've found using the client version to be
> quite acceptable.
So its just a few daemons changed around and some additional filesystem
stuff. Not really any different, just a couple additional items in the
OS install- so all the config/control shortcomings of OS X client will
still be there. Its the classic deal where when your config/control
requirements fit in the gui, things work OK, but as soon as your needs
exceed that, you're on your own without any assistance from the OS.
Don't get me wrong, I like OS X a heck of a lot better than Windows- but
its a long ways from being able to be much more than a simplistic
server.
Gregm
Wacom Graphire support is an odd reason to choose a language for a
writing 3d games. I'm sure that'll be the least of your problems,
regardless of your choice of OS or language.
You can be a tech pioneer or you can use proven technologies that work
out of the box. You can't do both.
··········@gmail.com wrote:
> Wacom Graphire support is an odd reason to choose a language for a
> writing 3d games. I'm sure that'll be the least of your problems,
> regardless of your choice of OS or language.
>
> You can be a tech pioneer or you can use proven technologies that work
> out of the box. You can't do both.
It's a question of platform support, and right now money. I only
mentioned the Wacom tablet as the worst problem. There are several
others that have sucked a lot of time out of me. Really the decision is
between Linux, Cygwin, MinGW, or straight up Windows as the programming
environment. I was surprised to find out how much work Linux asked me
to do. It's really not any less work than fighting MinGW, so why bother?
I consider the platform important. When a platform wastes 1.5 weeks of
my time on nothing, it's important. That's 1.5 weeks I could have been
writing an OpenGL binding from scratch, if I had had a mind to do so.
Instead it's compiling kernels for mouse drivers, a totally
uninteresting problem.
So now, the issue is whether lack of UFFI support is a dealbreaker or
not. Neither GNU CLISP, GNU Common Lisp, or Corman Lisp has it. I
suppose I'll have to compare UFFI to whatever GNU Common Lisp does have.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
"The pioneer is the one with the arrows in his back."
- anonymous entrepreneur
Brandon J. Van Every wrote:
>
> So now, the issue is whether lack of UFFI support is a dealbreaker or
> not. Neither GNU CLISP, GNU Common Lisp, or Corman Lisp has it. I
> suppose I'll have to compare UFFI to whatever GNU Common Lisp does have.
>
Right now you are like a hamster on a treadmill. Pick something and run
with it already. Choose a date, a goal by which you have to decide upon
the direction you are taking and then move ahead. Don't second guess
your decisions or you will make no progress at all.
What are your goals?
- Write a game?
- For yourself?
- To sell?
- Write a 3D engine from scratch in Lisp (then forget about writing a game)
- Evaluate technologies until the cows come home?
If you are looking to write a game using Lisp on Windows then you are
going to have to pay actual money for a decent Lisp environment. If you
go with Linux then you are able to make use of the open source solutions
that are available.
I use Corman Lisp in Windows. The Corman IDE is a little quirky but
since SLIME added support for Corman I have made the switch to Emacs and
am having a blast. Using the Corman FFI, I have written bindings to:
- SDL,
- SDL_mixer,
- SDL_gfx,
- OpenRM,
- Cal3D,
- And I am about to write bindings to either the Newton or ODE physics
libraries, depending on what takes my fancy.
There, you have quite enough to write a game in Lisp for Windows, so NOW
what is stopping you?
-Luke
I've been through the "Man called Horse" stage with Linux, and I have
the bone through my nose to prove it, so I know the difficulties you
have to go through to get it to work (at least if you throw random, not
well-supported hardware at it). I probably wouldn't make it now,
because, having since become a grown-up, I just don't have the time.
BTW, it's pretty trouble free once you have it set up. Windows users I
know constantly have to mess with spyware, etc, and Google + install
wizards is primitive way to install software compared to apt-get.
Installing Linux is like compressing all of your OS hassles into a few
weeks up front.
Having said that, I think there must be a bigger differential between
languages than the up-front hassle of getting things built or
configured correctly. Let X be the amount of time you spend on that
crap, and let Y be the amount of time you spend actually developing
software. I think Y >> X, and I'd be far more interested in an
environment that minimized Y than one that minimized X (which, for game
development, would be MSVC++, if I'm not too far out of date with
Windows and my long-forgotten ambitions to design games).
··········@gmail.com wrote:
>
> Having said that, I think there must be a bigger differential between
> languages than the up-front hassle of getting things built or
> configured correctly. Let X be the amount of time you spend on that
> crap, and let Y be the amount of time you spend actually developing
> software. I think Y >> X, and I'd be far more interested in an
> environment that minimized Y than one that minimized X (which, for game
> development, would be MSVC++, if I'm not too far out of date with
> Windows and my long-forgotten ambitions to design games).
I can't stomach C++ anymore. I consider it an evolutionary dead end.
The choice of Common Lisp is not proving as straightforward as I
thought, due to my vow of poverty. But, I'm not done yet. Got editors
and C FFIs to chug through.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
"Trollhunt" - (n.) A searching out for persecution
of persons accused of Trolling. (c.f. witch-hunt)
"Brandon J. Van Every" <·····················@mycompanyname.com> writes:
> I can't stomach C++ anymore. I consider it an evolutionary dead end.
To each his own. C is certainly more universal, but I sometimes have
trouble throttling back from the "better C" aspects of C++.
> The choice of Common Lisp is not proving as straightforward as I
> thought, due to my vow of poverty. But, I'm not done yet. Got
> editors and C FFIs to chug through.
I took a long time to actually get down and program anything in Lisp.
I had a project in mind that I never did get to. I was on Windows NT
back in 1998. I wanted something easy to parse for a scripting
language that could be used as an application extension language.
S-Exp representation seemed ideal. After some research I learned that
there was Lisp and Scheme. (By Lisp I mean ANSI Common Lisp).
Further inquiries convinced me that Lisp was preferable to Scheme for
a big application. I also didn't want to spend a bunch of money on a
development environment only to discover that I didn't care for it.
That was probably a very expensive mistake on my part. Like now, the
best free lisps were on Linux or FreeBSD. CMUCL was really the best
game in town at the time so far as Free goes. So I made the switch to
Linux. If you are up to a Google archive search, you can read
postings I made to cll at the time. Long story short, I ended up
being so distracted by the Linux experience that I never wrote a line
of Lisp.
That was my first false start with Lisp.
Work and life kept me too busy to learn Lisp for sometime. Back in
2003, I purchased a 12" PowerBook G4, fully loaded. OS X was a shiny
new thing for me to play with. I had some more free time, but I
didn't get into the Lisp groove straight away. In the time I was
doing other things, LispWorks PE was available to try. I haven't
tried it yet.
Over time, I got Carbon Emacs to build and run on my Mac. I played
with Fink under Jaguar and switched to DarwinPorts under Panther. I
continued to do everything except Lisp. Then something happened to
change that. Someone told me about SLIME. I grabbed SLIME from CVS
and got Carbon Emacs to use it. I started to play around with OpenMCL
and SBCL. SBCL is cool because it can be compiled from many ANSI
compliant Lisps and GCC for the kernel. I used OpenMCL to bootstrap.
I compile SBCL with itself now. I'm not quite current, but I do have
post 0.9.0 versions on a Debian machine and on my Mac.
Finally I had an environment I could hack Lisp with!
At first, I tried to bite off a little too much. I had that FFI bug
where I just had to use Lisp to make a native app. Eventually I came
to my senses and just did stuff that would work in Emacs and also run
under both OpenMCL and SBCL. I'm still a novice by my standards, but
I have started to develop some confidence with Lisp. That's the real
trick. Don't aim for the moon before you can crawl.
It does help to have programs to work on. Aimless coding doesn't
really accomplish much. I did some fractal stuff which works in any
ANSI Lisp and is totally platform neutral. Lately I've been working
on some programming puzzles. I even started an OS X specific project
to make a CLOS library based on the Carbon API as an alternative to
Cocoa. Mind you, OpenMCL does have a Cocoa bridge so I'm doing a bit
of upstream swimming for that. I also have no spec, so it's pure
cowboy coding. I've also let it lie fallow while working on the
puzzles in my spare time. This is fine because my Lisp familiarity is
improving while doing the puzzles which will help me with my Carbon
project.
The point of this lengthy story is that at some point you have to
settle down and start doing some programming. Your game project is
too big to take on in one bite. Start with the basics. Get familiar
with Lisp by writing lots of small programs. Do basic programming
exercises and puzzles without worrying about a GUI. Use the REPL as
your command line. Use SLIME in Emacs.
Right now, it seems you want to stay with Windows. Fine. Do that.
Download the Lispworks PE and live with the five hour timeout. Or
have another go with Linux and get Emacs + SLIME + SBCL going. Don't
worry if your tablet doesn't work. You just need the keyboard. Even
X is optional although it is nice to have the extra screen real
estate.
Stop casting about, complaining about things in the unproductive way
that you have been doing. Install a Lisp environment and just do it.
You should listen to my advice because I wasted one hell of a lot more
time than ten days. I can assure you that once you are hacking away
at Lisp code, the experience will be totally worth it. Get going.
Now. Don't wait another second.
--
(when (or hope despair)
(error "Deal with life as it is."))
On Sun, 12 Jun 2005 23:44:36 -0700, "Brandon J. Van Every"
<·····················@mycompanyname.com> wrote:
>
>I can't stomach C++ anymore. I consider it an evolutionary dead end.
>
>The choice of Common Lisp is not proving as straightforward as I
>thought, due to my vow of poverty. But, I'm not done yet. Got editors
>and C FFIs to chug through.
Let me add my voice to David's and Luke's. There is no substitute for
getting code written and working. I've spent a couple weeks sifting
through the alternatives for open-source Lisp on Windows. This was a
critical time because it would have been easy to give up. I've settled
on CLISP for the forseeable future, because for my purposes, CLISP has
the right balance between pros and cons.
Seeing bits and pieces of code working now is a big morale booster, much
more important than worrying about benchmarks or multi-threading or
native code compilation, etc.
--
Jack
From: Robert Uhl
Subject: Re: I've thought better of Linux
Date:
Message-ID: <m33brl8pyo.fsf@4dv.net>
Jack Unrue <·······@example.org> writes:
>
> Let me add my voice to David's and Luke's. There is no substitute for
> getting code written and working.
Ich auch.
I've been spending the last week or so rewriting my Tasting Notes app (a
web-accessible database of beer, brewer, bar, restaurant & food reviews)
in Lisp (SLIME + SBCL + araneida + CL-EMB) and it's just amazing what I
can do with Lisp. The instant I notice a pattern I can typically
abstract it away in a macro and never think about it again--I've gotten
several hundred lines of code down to three.
The key is to write, write, write from what I can tell. It's how I
learnt perl; it's how I learnt python; it's how I'm learning Lisp.
--
Robert Uhl <http://public.xdi.org/=ruhl>
Christmas is weird. What other time of the year do you sit
in front of a dead tree and eat candy out of your socks?
Jack Unrue wrote:
>
> Let me add my voice to David's and Luke's. There is no substitute for
> getting code written and working. I've spent a couple weeks sifting
> through the alternatives for open-source Lisp on Windows. This was a
> critical time because it would have been easy to give up. I've settled
> on CLISP for the forseeable future, because for my purposes, CLISP has
> the right balance between pros and cons.
>
> Seeing bits and pieces of code working now is a big morale booster, much
> more important than worrying about benchmarks or multi-threading or
> native code compilation, etc.
I agree. Which is why, performance jock that I am, that I'm willing to
defer my performance requirements. Really my question now is whether
CLISP is easy to set up, and whether SLIME, Jabberwocky, or VisualCLisp
is any good.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
"The pioneer is the one with the arrows in his back."
- anonymous entrepreneur
"Brandon J. Van Every" <·····················@mycompanyname.com> writes:
> ··········@gmail.com wrote:
> > Wacom Graphire support is an odd reason to choose a language for a
> > writing 3d games. I'm sure that'll be the least of your problems,
> > regardless of your choice of OS or language.
> > You can be a tech pioneer or you can use proven technologies that
> > work
> > out of the box. You can't do both.
>
> It's a question of platform support, and right now money. I only
> mentioned the Wacom tablet as the worst problem. There are several
> others that have sucked a lot of time out of me. Really the decision
> is between Linux, Cygwin, MinGW, or straight up Windows as the
> programming environment. I was surprised to find out how much work
> Linux asked me to do. It's really not any less work than fighting
> MinGW, so why bother?
>
> I consider the platform important. When a platform wastes 1.5 weeks
> of my time on nothing, it's important. That's 1.5 weeks I could have
> been writing an OpenGL binding from scratch, if I had had a mind to do
> so. Instead it's compiling kernels for mouse drivers, a totally
> uninteresting problem.
So, you were able to solve this totally uninteresting problem?
One thing I don't understand was all your stuff about having to
rebuild the kernel to get the support you needed - this seems very odd
to me and something I've not needed to do in years. Does the kernel
come with a driver for your mouse tablet thingy? If so, then surely
there was already a module which you simply needed to load? If it
doesn't come with support, then what was the point of rebuilding the
kernel.
Putting all that aside, and even factoring in your terrible state of
poverty, surely, since your real stated aim was to evaluate Linux as a
development platform, the sensible and very simple solution would have
been to just go and buy a cheap mouse - I mean, god, that would have
cost less than $10 and saved you loads of time - then if you decided
Linux was worth putting more effort into, you could have looked into
getting your flash mouse tablet working.
> So now, the issue is whether lack of UFFI support is a dealbreaker or
> not. Neither GNU CLISP, GNU Common Lisp, or Corman Lisp has it. I
> suppose I'll have to compare UFFI to whatever GNU Common Lisp does
> have.
>
This still seems like a weird way to go to me. Have you actually
written some lisp yet? Surely it would make more sense to prototype
one of your games or some simple game in Lisp first and decide if you
find it a good language for solving your problems and then see if you
can find an implementation with the right foreign function interface
once you know its a good development platform?
I don't know, but it just seems to me that every way you turn there is
some out of your control obstacle in your way which is preventing you
from earning all those riches you plan to get in 10 years time -
there must be a point at which you begin to re-examine your approach
and possibly realise its not GPL, Linux, UFFI, editors, GL libraries
or anything else thats getting in the way, but instead could possibly
be you yourself?
You may have heard of two common personality types - one type has a
fear of failure and consequently never actually starts anything
because if you don't start, you can't fail. The other has a fear of
success, they start things, but never finish them. I'm beginning to
suspect you fall into one of these groups - guess which one.
Tim
--
Tim Cross
The e-mail address on this message is FALSE (obviously!). My real e-mail is
to a company in Australia called rapttech and my login is tcross - if you
really need to send mail, you should be able to work it out!
Tim X wrote:
> "Brandon J. Van Every" <·····················@mycompanyname.com> writes:
>>
>>I consider the platform important. When a platform wastes 1.5 weeks
>>of my time on nothing, it's important. That's 1.5 weeks I could have
>>been writing an OpenGL binding from scratch, if I had had a mind to do
>>so. Instead it's compiling kernels for mouse drivers, a totally
>>uninteresting problem.
>
>
> So, you were able to solve this totally uninteresting problem?
I could have, for 1..2 days more additional pain. I declined, on the
grounds that I seriously doubted it would be the end of such problems.
I think these kinds of Linux problems would find a way to go on forever.
The whole point of moving to Linux was to get away from having to
compile up tons of things manually for MinGW, and it turns out I wasn't
actually saving any labor over that drill.
> One thing I don't understand was all your stuff about having to
> rebuild the kernel to get the support you needed - this seems very odd
> to me and something I've not needed to do in years. Does the kernel
> come with a driver for your mouse tablet thingy? If so, then surely
> there was already a module which you simply needed to load? If it
> doesn't come with support, then what was the point of rebuilding the
> kernel.
The idea of explaining all the Wacom Graphire driver problems in detail
bores me. That would be allowing the ghost to continue to make me work.
There is some debate over whether I needed to compile a kernel, or
just compile a driver and link it into the kernel. At any rate, there
was still a learning curve left to go.
>
> Putting all that aside, and even factoring in your terrible state of
> poverty, surely, since your real stated aim was to evaluate Linux as a
> development platform, the sensible and very simple solution would have
> been to just go and buy a cheap mouse -
I have a PS/2 mouse sitting on my desktop, that worked with Linux. I
don't like it. I think the drivers jolly well should work. And it
wasn't only that 1 problem.
I evaluated Linux. It saves labor in software packages, but creates
labor with driver support. Not as good a trade as I thought it would
be. And I really don't like all these stacked packages. For
portability and build sanity I want as few dependencies as possible.
Instead the open source way is to keep piling on the library packages,
and some of those are bound to be cantankerous to get going on Windows.
> I mean, god, that would have
> cost less than $10 and saved you loads of time - then if you decided
> Linux was worth putting more effort into, you could have looked into
> getting your flash mouse tablet working.
It only took 1.5 weeks to decide Linux wasn't worth my time. Remember,
it's irrelevant as a consumer gaming desktop. I hadn't previously
realized that.
> This still seems like a weird way to go to me. Have you actually
> written some lisp yet? Surely it would make more sense to prototype
> one of your games or some simple game in Lisp first and decide if you
> find it a good language for solving your problems and then see if you
> can find an implementation with the right foreign function interface
> once you know its a good development platform?
Why do people talk about getting ready to do things before doing them?
Like prototyping is free time that doesn't take very long or something?
> You may have heard of two common personality types - one type has a
> fear of failure and consequently never actually starts anything
> because if you don't start, you can't fail. The other has a fear of
> success, they start things, but never finish them. I'm beginning to
> suspect you fall into one of these groups - guess which one.
Would you like to go up to Marysville today and make $600 on signature
gathering for me? Do any of you people actually have jobs and financial
pressures apart from your Lisp habit?
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
20% of the world is real.
80% is gobbledygook we make up inside our own heads.
"Brandon J. Van Every" <·····················@mycompanyname.com> writes:
> Why do people talk about getting ready to do things before doing them?
> Like prototyping is free time that doesn't take very long or something?
>
Wow, if you don't understand the benefits of actually trying out a
language before deciding to use it, I have even less confidence in you
ever producing anything other than perhaps a few glorified hello world
scripts. You have obviously never written anything with any real
complexity in it. A prototype allows you to verify both a concept and
the suitability of the platform/language before you end up spending
weeks/months/years in developing the final application. You would be
amazed at the number of apps which come unstuck after months of work
because something suddenly became apparent which wasn't when it was
all at a conceptual stage. A good prototype may take a few weeks to
do, but can save months or years of wasted effort.
> > You may have heard of two common personality types - one type has a
> > fear of failure and consequently never actually starts anything
> > because if you don't start, you can't fail. The other has a fear of
> > success, they start things, but never finish them. I'm beginning to
> > suspect you fall into one of these groups - guess which one.
>
> Would you like to go up to Marysville today and make $600 on signature
> gathering for me? Do any of you people actually have jobs and
> financial pressures apart from your Lisp habit?
Well, I guess I'm lucky. I don't have any financial pressures
anymore. I actually only work because on the whole I like to and
mostly enjoy what I do. However, at one point I had pretty much
nothing and worked in a couple of dead end jobs. However, while I was
doing those dead end jobs, I actually wrote applications, sold them
and managed to change my situation by actually producing something
which people were willing to purchase - I didn't waste weeks of time
on the Internet moaning about how nothing fits my unique situation and
that in 10 years I would be a millionaire!
It is also very apparent you have never met or known anyone who has
been successful enough to become a millionaire. I've known a few who
have started with nothing and ended up with more wealth than they ever
expected and the one thing they all have in common is that not a
single one of them had the aim to be a millionaire - in every case,
their objective was to either do something they enjoyed or to realise
an idea they had - the wealth was just a side effect.
>
> 20% of the world is real.
> 80% is gobbledygook we make up inside our own heads.
Aha - now I'm beginning to understand!
--
Tim Cross
The e-mail address on this message is FALSE (obviously!). My real e-mail is
to a company in Australia called rapttech and my login is tcross - if you
really need to send mail, you should be able to work it out!
Tim X wrote:
> "Brandon J. Van Every" <·····················@mycompanyname.com> writes:
>
>
>
>>Why do people talk about getting ready to do things before doing them?
>>Like prototyping is free time that doesn't take very long or something?
>>
>
>
> Wow, if you don't understand the benefits of actually trying out a
> language before deciding to use it, I have even less confidence in you
> ever producing anything other than perhaps a few glorified hello world
> scripts. You have obviously never written anything with any real
> complexity in it.
I'd just as soon say that you've only worked on big bloated projects
with large teams, large budgets, and lotsa time to plan out what you're
doing. That's just not the reality of a lone wolf indie game developer.
It's rather like the difference between mainstream Hollywood film
production and indie film production. Indies have to be smarter about
these things and pick better tools and processes.
> A prototype allows you to verify both a concept and
> the suitability of the platform/language before you end up spending
> weeks/months/years in developing the final application.
And you believe you're going to get these prototypes done in a day,
rather than spending at least weeks on such things? You're dreaming.
Or else I'd *love* to have your dev environment, and so would countless
other businesses. Such tools would be a significant competitive advantage.
> You would be
> amazed at the number of apps which come unstuck after months of work
> because something suddenly became apparent which wasn't when it was
> all at a conceptual stage. A good prototype may take a few weeks to
> do, but can save months or years of wasted effort.
Sure, I quit DEC in 1998 partly because of such a thing. I was the
bright guy who saw how we were screwed 6 months before anyone else did.
12 months later, we were still screwed. I had made my best efforts to
help extract us from our precarious predicament, but I had drawn a line
in the sand and was determined to get off the sinking ship.
There's another important reality to accept about development, however.
You *cannot* possibly plan for all contingencies. You can duck,
dodge, weave, and prototype all you like. Some things are bigger than
you and will screw you. You can apply incredible tactics and be screwed
by Grand Strategy. You may not have the wisdom and crystal ball to
perceive the correct Grand Strategy. Market forces can change what you
should have done; it is quite possible to make the correct moves "at any
given point in the game" and still end up screwed.
So, all calls for plans and prototypes have to be balanced against the
pressing need to get real work done now. I daresay 90% of the effort
has to be directed to the present, not the nebulous future.
>>Would you like to go up to Marysville today and make $600 on signature
>>gathering for me?
Which, incidentally, did not work out remotely like I hoped. So I'm
still in a struggle for survival.
>>Do any of you people actually have jobs and
>>financial pressures apart from your Lisp habit?
>
> Well, I guess I'm lucky. I don't have any financial pressures
> anymore.
You will have to be the judge of how much of your circumstances are
"luck." When I was in better financial shape, I got irritated with
people who said I was "lucky" to be pursuing my own visions without a
job. I told them, "I decided to quit my job and go for it. What's
stopping you from doing the same?" And now I feel it only more
strongly, as I've been screwed for my own choices just as much as I've
been empowered. The only thing that was definitely "luck" about my
circumstances, as in bad luck, was the dot.com bust. I didn't ask for
that to happen, and I didn't forsee it taking out so many companies, not
just the bozo companies.
> I actually wrote applications,
How about games? Are you familiar with the difficulties of creative
work? Probably not. I bet you did regimented "known to work" kinds of
stuff.
> It is also very apparent you have never met or known anyone who has
> been successful enough to become a millionaire.
That's ridiculous. It's apparent to me you're on the Judger end of the
MBTI, although I don't know how far. http://www.personalitypage.com I
have a conversation with someone like you every 4 months or so. The
script is always pretty similar.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
Does your Myers Briggs Type Indicator determine how you
debate? http://www.personalitypage.com/
··········@gmail.com wrote:
> Do you think you're going to be able to sit down and write a very
> difficult and complex project with heavy optimization needs in a
> language you've never used, and is quite different from anything you've
> ever used before, and not end up, at the very least, throwing out big
> chunks of it and starting over? The point is that you're prototyping
> whether you like it or not. If you don't admit it, you just end up
> wasting more time optimizing the parts that you're going to throw out.
> Saying you don't have the resources doesn't help. Besides, the Lisp
> way of doing things is to rapidly write slow code, and then go back and
> speed it up when and where necessary.
This discussion is becoming tiring. I very much feel it's like teaching
your grandma to suck eggs. Part of the reason I'm so poor is I've gone
through this "prototyping" thing before, with C++. I tried 4 different
planetary models for Ocean Mars, over the course of 9 months, working 60
hours a week. If I had just accepted the inadequacies of the 1st one,
it might have been somewhat painful to work with, but I would have had
something demoable.
You can talk about prototyping all you want. Prototyping is a *stage*
in development, and you have to define the scope of resources that are
going to be used for that stage. Those resources must be proportional
to the scope of the overall project, i.e. significantly less, or there
is no value in prototyping at all. Prototyping is not a magic bullet or
a license to just keep coding forever without actually making hard
decisions about engineering tradeoffs and kicking the functionality out
the door. Prototyping is also useless if you have to do everything all
over again. There isn't time for a lone wolf to do everything all over
again. Most things need to be "good enough" on the 1st try.
Another way of framing this debate, is to say that refactoring is
expensive and should *never* be your first thought as you're working on
your project. Refactoring should be delayed as long as possible, and
only done when you're *absolutely certain* you need it.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
"The pioneer is the one with the arrows in his back."
- anonymous entrepreneur
Brandon J. Van Every wrote:
> ...I've gone
> through this "prototyping" thing before, with C++. I tried 4 different
> planetary models for Ocean Mars, over the course of 9 months, working 60
> hours a week. If I had just accepted the inadequacies of the 1st one,
> it might have been somewhat painful to work with, but I would have had
> something demoable.
Prototyping is not about "accepting inadequacies" before you can move
on; rather it is about "proving adequacy" before you commit to a path.
That is, the prototype is an experiment to confirm that the project
requirements can be adequately met within the project constraints using
a certain strategy or design.
It sounds like you had some requirement for "adequacy" which could not
be met with the resources available. The point is your prototyping
seems to have *proved* to your satisfaction that you could not meet
your original goals, but could meet revised goals. Thats the whole damn
reason to have done a prototype.
The extra effort sounds like it was due to not revising your
requirements immediately, but instead trying different strategies to
meet the initial requirements, which turned out not to be essential
after all.
> ...
> Prototyping is not a magic bullet or
> a license to just keep coding forever without actually making hard
> decisions about engineering tradeoffs and kicking the functionality out
> the door. Prototyping is also useless if you have to do everything all
> over again. There isn't time for a lone wolf to do everything all over
> again. Most things need to be "good enough" on the 1st try.
You seem to be missing the point entirely: the choices being discussed
were
A) Prototype
B) No prototype
You seemed to claim "A" is overrated when compared against "B". That is
manifestly wrong.
With the clarifications above, what you appear to believe is that the
choice is between
B) No prototype
A') Prototype, but with such vague problem definition that the
prototype stage spirals out of control, or with such poor project
management that you never reach your milestones.
That is a false choice. You are confusing "prototyping" with a totally
different class of issues, involving clear definition of problem scope,
requirements, and resource allocation, which can screw up *any* stage
of development.
Yes, prototyping in a bone-headed fashion is bad. But that is because
being bone-headed is bad, not because prototyping is bad.
As for your wish to get criticism from "indie game designers," in
Usenet, you get what you pay for. Offer to pay an indie game designer
to criticize you (in private, please), and I'm sure one will be happy
to oblige, if the price is right. Or, accomplish something substantial
enough to be *visible* to indie game designers, and they may do it on
their own.
Or, you could realize that your problems are easy enough to diagnose
without consulting a specialist, and accept the free advice you are
being generously offered.
··············@hotmail.com wrote:
> Brandon J. Van Every wrote:
>
>
>>...I've gone
>>through this "prototyping" thing before, with C++. I tried 4 different
>>planetary models for Ocean Mars, over the course of 9 months, working 60
>>hours a week. If I had just accepted the inadequacies of the 1st one,
>>it might have been somewhat painful to work with, but I would have had
>>something demoable.
>
>
> Prototyping is not about "accepting inadequacies" before you can move
> on; rather it is about "proving adequacy" before you commit to a path.
> That is, the prototype is an experiment to confirm that the project
> requirements can be adequately met within the project constraints using
> a certain strategy or design.
Well I sure as hell wasn't going to give up on my Ocean Mars planetary
rendering technology, if that's what you're driving at. Call me crazy
for attempting a problem that was more difficult than it appeared.
> It sounds like you had some requirement for "adequacy" which could not
> be met with the resources available.
Rendering a 1.5 million hex model of Mars was somewhat unreasonable,
both in technology and game design terms. It was *possible* on the HW I
was using, but it would absolutely stretch the limit of what the HW
could do, and inflict serious implementation pain in the process. I
never even answered whether the AI could be done on such HW. Never even
started asking that question. Except on paper, and the
back-of-the-envelope math didn't look good.
> The point is your prototyping
> seems to have *proved* to your satisfaction that you could not meet
> your original goals, but could meet revised goals. Thats the whole damn
> reason to have done a prototype.
Yes, fine, so my 'prototype' was 9 months of working very hard 60 hours
a week. And indeed, the game ideas I have planned now are much more
abstract and have far fewer moving parts to them. Unfortunately, open
source toolchains mostly suck on Windows. I've wasted 2 years of my
time on that. What should the 'prototype' of that have been? Just try
some compiler out for 3 days and give up? I wouldn't have learned a
damn thing about Functional Programming if I had done that.
> The extra effort sounds like it was due to not revising your
> requirements immediately, but instead trying different strategies to
> meet the initial requirements, which turned out not to be essential
> after all.
Sure, that's a correct analysis. But it doesn't speak to the issue of
what you actually want to get done. That is the danger of creative
work. Find me any game development house that doesn't get slammed
somehow. I'll be impressed.
Programming at DEC was a helluva lot easier. We only worked on
well-studied problems. Each engineering iteration was severely
incremental over previous generations of 3D device drivers. Nothing
groundbreaking happening there.
> You seem to be missing the point entirely: the choices being discussed
> were
>
> A) Prototype
> B) No prototype
No, that's what *you* want to discuss. I'm saying, there's this whole
other dimension of creative work that completely skews this nice,
rationlized engineering notion of "this is how we're going to get from A
to Z."
> Yes, prototyping in a bone-headed fashion is bad. But that is because
> being bone-headed is bad, not because prototyping is bad.
The only way you can be continuously smart is if you are working on
well-studied problems.
> Or, you could realize that your problems are easy enough to diagnose
> without consulting a specialist, and accept the free advice you are
> being generously offered.
One day, the Judgers of the world - the ones who so freely dispense
their 'advice' to others - *might* realize how they're living out their
own predictable personal psychology. But I won't hold my breath. They
will contiune to teach their grandmas how to suck eggs.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
Taking risk where others will not.
Brandon J. Van Every wrote:
> No, that's what *you* want to discuss. I'm saying, there's this whole
> other dimension of creative work that completely skews this nice,
> rationlized engineering notion of "this is how we're going to get from A
> to Z."
Did you miss my multiple examples of prototyping in creative fields?
Movie makers use storyboards to prototype the dramatic structure before
incurring the massive costs of production.
Animators use crude "animatics" of hand-drawn sketches, or wire-frame
models in the case of CGI, to prototype the animation sequencies and
evaluate the visual effects before incurring the massive costs of
frame-by-frame animating or rendering.
Painters use quick sketches and and sculptors use clay studies to
prototype overall design.
Architects use quick sketches and studies and models before rendering
full prints.
In all of these creative fields, true professionals use prototyping to
estimate and mitigate risk, and reduce overall cost and time in
projects.
In all of these fields, spending too much time exploring blind alleys,
attempting the impossible, or continuously studying and changing tools
instead of practicing one's art and completing projects, will all lead
to failure. There's nothing magic about "being creative" that overcomes
poor planning and lack of discipline or professional skill.
···············@hotmail.com" <············@gmail.com> writes:
>
> There's nothing magic about "being creative" that overcomes poor
> planning and lack of discipline or professional skill.
But when you're a 'lone wolf' 'indie game developer' you're somehow
freed from the constraints that bind lesser men--that is, those who've
somehow managed to create stuff.</sarcasm>
--
Robert Uhl <http://public.xdi.org/=ruhl>
It's not like I burst forth from my father's brow holding a fistful
of cash and 10 years of UNIX sysadmin experience. --Mike Sphar
··············@hotmail.com wrote:
>
> Did you miss my multiple examples of prototyping in creative fields?
Honestly I am tired of this discussion. I discovered, many years ago,
that there is in fact a motive for shutting up and not sharing views /
opinions on Usenet, despite my impulses to the contrary. The problem is
that knowledge dissemination is a cascade. First there's what's in my
mind. Then there's what's written in my notebooks, which you may or may
not refer to as prototypes. Then there is implementation, including
getting tools to build and work. Then there is talking to other people
about what the hell you're doing. The effort of fully articulating and
explaining what the hell you do is a parlor game, fraught with difficulty.
I don't really see a point in debating our mental models of
'prototyping' anymore. Suffice it to say I have one, I use it, and I
know from experience that its uses are limited. Technologies that work
the 1st time you run through them are more important than the
exploratory stuff. If your project is too exploratory, it will never
get done.
> In all of these fields, spending too much time exploring blind alleys,
> attempting the impossible, or continuously studying and changing tools
> instead of practicing one's art and completing projects, will all lead
> to failure. There's nothing magic about "being creative" that overcomes
> poor planning and lack of discipline or professional skill.
So do you actually work in a creative programming field, and have
actually taken the lumps from it, or are all your comments theoretical?
It's very easy to talk about "being a pro" when you haven't been there.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
"The pioneer is the one with the arrows in his back."
- anonymous entrepreneur
Brandon J. Van Every wrote:
> So do you actually work in a creative programming field, and have
> actually taken the lumps from it, or are all your comments theoretical?
> It's very easy to talk about "being a pro" when you haven't been there.
Pot : "Hello Mr Kettle, you sure are black."
--
Drew Crampsie
drewc at tech dot coop
"Never mind the bollocks -- here's the sexp's tools."
-- Karl A. Krueger on comp.lang.lisp
drewc wrote:
> Brandon J. Van Every wrote:
>
>> So do you actually work in a creative programming field, and have
>> actually taken the lumps from it, or are all your comments
>> theoretical? It's very easy to talk about "being a pro" when you
>> haven't been there.
>
>
> Pot : "Hello Mr Kettle, you sure are black."
I have prototype game designs, prototype planetary rendering code...
...no I'm not putting them up on a website for you to verify. But I
know well the difficulties of creative programming work. If you don't
want to hear about me, try reading a few hundred GameDeveloper
postmortems about all the game industry's usual foibles.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
"The pioneer is the one with the arrows in his back."
- anonymous entrepreneur
"Brandon J. Van Every" <·····················@mycompanyname.com> writes:
>exploratory stuff. If your project is too exploratory, it will never
>get done.
If you don't like exploratory stuff, then what the hell are you doing
here?
If you don't like experimentation, then make a strict implementation
plan and go through the phases of software engineering. Make a concept
model with Use Cases, make a complete specification in UML, with a
class model and activity diagrams, sequence and communication diagrams
for the participating objects, and work it all out on paper (or in the
UML editor) as far as possible before you even write a single line of
code.
When it's all worked out pretty well, you can implement (most) of it
straight away in C++, or a similar OOP(-style) language that you
already know, without the hazzle and trouble that goes with
experimenting in those languages. This actually works pretty well for
many thousands of software teams (and individuals) out there. You can,
of course, also use Lisp but I guess that you don't know that language
yet, so you would have years of practice in front of you to actually
learn and get acquainted with the language, and in the software
engineering model, the implementation language is secondary, anyways.
With Lisp, prototyping and experimentation is often preferred to the
above modelling, and many programs are built bottom-up instead of
top-down. This is not because this approach is generally better, but
because the language is quite malleable and lends itself very well to
prototyping. But if you don't want to do prototyping and
experimentation at all and want the first implementation to work
straight away, then by all means go for systematic software design
with proven software engineering tools and procedures. It might not be
as joyful and cool as hacking away but you might at least get
something done in the end.
mkb.
Matthias Buelow wrote:
> If you don't like experimentation, then make a strict implementation
> plan and go through the phases of software engineering. Make a concept
> model with Use Cases, make a complete specification in UML, with a
> class model and activity diagrams, sequence and communication diagrams
> for the participating objects, and work it all out on paper (or in the
> UML editor) as far as possible before you even write a single line of
> code.
If we use the painting-sketch analogy, UML does count as a prototype.
I personally found UML to be not very useful to me in practice.
alex goldman wrote:
> If we use the painting-sketch analogy, UML does count as a prototype.
>
> I personally found UML to be not very useful to me in practice.
I think formal UML *is* overkill, but I find loose
class-arrows-and-methods diagrams somewhat useful, just to picture
the relationships of object dependencies/creation and program flow
(method invocations).
--
Don't let school interfere with your education. -- Mark Twain
Matthias Buelow wrote:
> "Brandon J. Van Every" <·····················@mycompanyname.com> writes:
>
>
>>exploratory stuff. If your project is too exploratory, it will never
>>get done.
>
>
> If you don't like exploratory stuff, then what the hell are you doing
> here?
You mean in Lisp-land? Indeed I have wondered. I see the Lisp cultural
tendency of "everything's a prototype" as a serious engineering
liability. It means people mostly roll their own and don't bring tools
and libraries to real world proven maturity. That means much less
off-the-shelf stuff that's going to work the 1st time or do anything useful.
On the other hand, open source in general isn't very good at supplying
off-the-shelf solutions for my problems, no matter what the language.
For my needs, open source has proven to be almost a complete waste of
time. 2 years of wasted time, <sigh>.
Awhile ago, I accepted a roll-my-own mentality. That was true when once
upon a time I moved from OCaml to Bigloo Scheme, roughly last winter. I
still consider things like Standard ML or Mercury to be possibilities,
seeing as how nobody's got any code of any use to me anyways. In fact,
in the extreme, if I get really sick of all these languages that don't
do what I want, I'll be implementing my own language. But for now,
Common Lisp is the most standard + supported of all the esoteric
languages I've investigated. So that's why I'm still provisionally here.
Financial pressure might be about to drive me straight to Java though.
It is way, way too difficult to find a job with a "good language" in
Seattle. So, between the not-so-good languages, I'm simply not going to
do C++ anymore, and I have tremendous antipathy for all things
Microsoft. That would seem to leave Java. I don't see it being useful
for my game development problems, but right now I have the far simpler
problem of needing an income (this signature gathering season is over in
WA).
Holding nose while contemplating Java QA test... weighing it against
working for $10/hour at a homeless shelter... which will suck less?
> If you don't like experimentation, then make a strict implementation
> plan and go through the phases of software engineering. Make a concept
> model with Use Cases, make a complete specification in UML, with a
> class model and activity diagrams, sequence and communication diagrams
> for the participating objects, and work it all out on paper (or in the
> UML editor) as far as possible before you even write a single line of
> code.
For a lone wolf developer this is nuts. The reason to perform all that
labor is to firmly communicate your designs to *large teams of other
people*. If you don't have such a team, you are wasting your time. A
sane approach is to pick a component you are *certain* you need for your
project, and build it. For instance, I know for fact I need an OpenGL
binding.
> You can,
> of course, also use Lisp but I guess that you don't know that language
> yet, so you would have years of practice in front of you to actually
> learn and get acquainted with the language,
I don't see why. Lisp is just a language. It doesn't require new
paradigms to use it, unlike a hardcore FP language like Haskell or
Clean. Nobody said you *have* to write macros or self-modifying code.
And, there's a lot less syntax to learn than, say, OCaml.
Of course, I'm sure a number of you probably think me too stupid to
learn a new language quickly. You probably don't accept the alternate
explanation, that I've learned portions of many languages, but found
their facilities inadequate to my game development problems. The
sticking point is always "ok, how painful is it going to be to talk to
C, ASM, and OpenGL...."
> and in the software
> engineering model, the implementation language is secondary, anyways.
Indeed.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
I won't spend more than 1 day configuring 1 thing.
Brandon J. Van Every wrote:
>
> Holding nose while contemplating Java QA test... weighing it against
> working for $10/hour at a homeless shelter... which will suck less?
>
Have you thought about panhandling around MS campus, telling people "help,
I've been bankrupted by Java" or something hilarious like that?
"Brandon J. Van Every" <·····················@mycompanyname.com> writes:
>You mean in Lisp-land? Indeed I have wondered. I see the Lisp
>cultural tendency of "everything's a prototype" as a serious
>engineering liability. It means people mostly roll their own and
What do you know about the Lisp cultural tendency? How many Lisp
programs (and their development story) have you seen? You're just
speculating out of your arse, that's all.
>On the other hand, open source in general isn't very good at supplying
>off-the-shelf solutions for my problems, no matter what the
>language. For my needs, open source has proven to be almost a complete
>waste of time. 2 years of wasted time, <sigh>.
Now don't start blaming the open source community for your failures.
>Holding nose while contemplating Java QA test... weighing it against
>working for $10/hour at a homeless shelter... which will suck less?
I'd get any programming job first and work in that for a while, and
then think about realizing projects for which you do not yet have the
abilities and resources.
>For a lone wolf developer this is nuts. The reason to perform all
>that labor is to firmly communicate your designs to *large teams of
>other people*. If you don't have such a team, you are wasting your
>time.
That's one reason. The other is to gather and visualize your thoughts
and ideas in a way that (ideally) lets you reformulate them directly
as object-oriented code. Remember, you didn't want to waste any time
on exploration. You want the first implementation to work straight
away. However, you don't want to invest time on formal design
either. So _how_ on earth do you actually want to accomplish the task?
Just wait until it's grown on your computer disk like some mold?
>A sane approach is to pick a component you are *certain* you
>need for your project, and build it. For instance, I know for fact I
>need an OpenGL binding.
I can't see any correlation between chosing tools and using software
design strategies. I somehow get the nagging thought that all you know
and can do _is_ chosing tools (or at least try to). That's why you're
running around like a foraging mouse, trying all kind of "esoteric"
languages and tools, instead just selecting one and getting to the
actual implementation phase (if you don't like designing your software
first). You seriously have to rethink that approach and question
yourself whether you want to do some programming at all or if it's all
some kind of hobbyist pipe dream.
>I don't see why. Lisp is just a language. It doesn't require new
>paradigms to use it, unlike a hardcore FP language like Haskell or
Of course it requires new paradigms. Lisp is often described as being
a language that opens programmers' minds to new programming paradigms.
You can program C-style in Lisp but that will do no good.
I have the notion that if you actually started to program some Lisp,
you'd come back here within a day and whine how shitty useless the
language is and that you won't put up with this kind of nonsense
anymore. Please do not hesitate in trying to learn lisp. Might save
you some time.
mkb.
Matthias Buelow wrote:
> "Brandon J. Van Every" <·····················@mycompanyname.com> writes:
>
>
>>You mean in Lisp-land? Indeed I have wondered. I see the Lisp
>>cultural tendency of "everything's a prototype" as a serious
>>engineering liability. It means people mostly roll their own and
>
>
> What do you know about the Lisp cultural tendency?
Copious perusal of this forum, its archives, the various Lisp wikis,
several active Lisp mailing lists, numerous web searches, almost the
entireity of what's available at Sourceforge, and the input of the
Lispers who co-founded SeaFunc. http://groups.yahoo.com/group/SeaFunc
> How many Lisp
> programs (and their development story) have you seen? You're just
> speculating out of your arse, that's all.
As per above, no not really.
>>On the other hand, open source in general isn't very good at supplying
>>off-the-shelf solutions for my problems, no matter what the
>>language. For my needs, open source has proven to be almost a complete
>>waste of time. 2 years of wasted time, <sigh>.
>
>
> Now don't start blaming the open source community for your failures.
Failure of various HLLs to be widespread and have healthy development
trajectories with plenty 'o' jobs available are not my failures. Of
course if you believe open source is the cure-all for the world's ills,
I'll just have a good laugh about you at the next SeaFunc meeting. The
reality of open source is that in the vast majority of cases, you get
what you pay for.
I think open source is probably most effective when used by large
organizations with dedicated manpower to fill in the inevitable cracks.
ILM, for instance, profits from Python because it can custom engineer
it to do exactly what they want it to do on various machines. So
they'll run Python 1.5 + back-port any new functionality they actually
need. Eventually some old batch of machines is retired and they can
slowly creep the Python version forwards.
Similarly, Linux is effective in corporations when you have sysadmins
constantly in tow. It sucks as a consumer desktop.
To use a Microsoft-ism, the Total Cost of Ownership (TCO) of open source
is often poor. Now, the TCO of Microsoft toolchains might be equally
poor, or poorer, especially when security issues are considered. But
there's no question that for my problem domains, open source TCO has
been poor. I think Microsoft is correct to make us all think about TCO,
to look at all sources of headache. Not just "wow that sounds good
because it costs $0 and I can fiddle with it!"
>>Holding nose while contemplating Java QA test... weighing it against
>>working for $10/hour at a homeless shelter... which will suck less?
>
>
> I'd get any programming job first and work in that for a while, and
> then think about realizing projects for which you do not yet have the
> abilities and resources.
That's you. I think making a homeless guy's life more bearable has a
point to it. Whereas working on some broken-ass programming language,
in a problem domain irrelevant to game development, doesn't. That said,
I've got my own problems to solve. I've done social work before and
that's an awful lot of emotional drainage for a small amount of money.
If they were paying $20/hour the decision would be easy.
>>For a lone wolf developer this is nuts. The reason to perform all
>>that labor is to firmly communicate your designs to *large teams of
>>other people*. If you don't have such a team, you are wasting your
>>time.
>
>
> That's one reason. The other is to gather and visualize your thoughts
> and ideas in a way that (ideally) lets you reformulate them directly
> as object-oriented code.
That's too heavyweight a process for a lone wolf. You should spend 10%
of your time designing and 90% of your time implementing it. I've done
things the "design, design, design" way, it doesn't work.
Also, I don't believe in OO anymore. Except for well-studied problems,
OO designs tend to break when faced with the harsh light of reality.
Most problems are relational, they don't have clean object boundaries.
> Remember, you didn't want to waste any time
> on exploration.
Actually, you know what I really want? I really want to not waste any
more time debating this. I'm only doing this because jobhunting is
deadly boring and I need some kind of relief. I think it's time to find
a better source of relief. Unsubscribing and disappearing for awhile,
as I am apt to do.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
I won't spend more than 1 day configuring 1 thing.
"Brandon J. Van Every" <·····················@mycompanyname.com> writes:
> To use a Microsoft-ism, the Total Cost of Ownership (TCO) of open
> source is often poor. Now, the TCO of Microsoft toolchains might be
> equally poor, or poorer, especially when security issues are
> considered. But there's no question that for my problem domains, open
> source TCO has been poor. I think Microsoft is correct to make us all
> think about TCO, to look at all sources of headache. Not just "wow
> that sounds good because it costs $0 and I can fiddle with it!"
Ignorance is bliss. But who here wants ignorance?
Freedom too is costly.
Let's all be slave! Ignorant slaves!
--
__Pascal Bourguignon__ http://www.informatimago.com/
Our enemies are innovative and resourceful, and so are we. They never
stop thinking about new ways to harm our country and our people, and
neither do we. -- Georges W. Bush
"Brandon J. Van Every" <·····················@mycompanyname.com> writes:
> Matthias Buelow wrote:
> > "Brandon J. Van Every" <·····················@mycompanyname.com> writes:
> >
> >>exploratory stuff. If your project is too exploratory, it will never
> >>get done.
> > If you don't like exploratory stuff, then what the hell are you doing
> > here?
>
> You mean in Lisp-land? Indeed I have wondered. I see the Lisp cultural
> tendency of "everything's a prototype" as a serious engineering
> liability. It means people mostly roll their own and don't bring tools
> and libraries to real world proven maturity. That means much less
> off-the-shelf stuff that's going to work the 1st time or do anything
> useful.
Says who? So you're capable of judging all this before you've even
written a few systems with it? Or is this one of those things where if
you can conceive it, then it must be true?
Lisp IS handy for dinking around with half-formed ideas- it tends to
take less work. But it also takes less work to get down to it and
implement stuff. The not-invented-here vs library pressures are not
materially different than in any other language.
I suggest you STFU and write some Lisp and THEN come back here to
comment about what the language is & isn't.
Gregm
Brandon J. Van Every wrote:
> Of course, I'm sure a number of you probably think me too stupid to
> learn a new language quickly.
I don't think that's the real reason.
> You probably don't accept the alternate
> explanation, that I've learned portions of many languages, but found
> their facilities inadequate to my game development problems.
You're right about that.
By the way, we somehow managed to write that Pentium
native code generator and get it running without your
help [1]. Thanks for offering, though.
Will
[1] http://www.ccs.neu.edu/home/will/Research/ILC2005/benchmarks.html
On Mon, 20 Jun 2005 21:38:42 -0700, Brandon J. Van Every wrote:
> No, that's what *you* want to discuss. I'm saying, there's this whole
> other dimension of creative work that completely skews this nice,
> rationlized engineering notion of "this is how we're going to get from A
> to Z."
This whole other dimension is one of those string theory ones - to small
to be detected easily. But seriously, creative work is about 1 percent
creative, and 99 percent work. I've had the experience that for the work
part, prototyping helps.
--
Christopher
On Mon, 20 Jun 2005 10:20:20 -0700, Brandon J. Van Every wrote:
> This discussion is becoming tiring. I very much feel it's like teaching
> your grandma to suck eggs. Part of the reason I'm so poor is I've gone
> through this "prototyping" thing before, with C++.
Just a thought: maybe part of the reason you're so poor is you don't do
anything all day but post to newsgroups complaining instead of coding
something useful?
--
Christopher
Christopher Koppler wrote:
> On Mon, 20 Jun 2005 10:20:20 -0700, Brandon J. Van Every wrote:
>
>
>>This discussion is becoming tiring. I very much feel it's like teaching
>>your grandma to suck eggs. Part of the reason I'm so poor is I've gone
>>through this "prototyping" thing before, with C++.
>
>
> Just a thought: maybe part of the reason you're so poor is you don't do
> anything all day but post to newsgroups complaining instead of coding
> something useful?
Well, considering that I just went out for an hour to gather signatures,
and got skunked, and that I'm eating lunch before going to the office to
turn in what I've got, and I worked 6.5 hours to make $25 yesterday, and
I didn't do so well in Marysville on Saturday + lost $20..$40 worth of
signature boards while changing buses, and etc. and so forth and so on
as I've struggled to barely make ends meet and stay one step ahead of my
landlord, because at least the job is/was a bird in the hand, I can very
safely say...
No, that is not the problem. The problem is that signature gathering
often looks like it's going to pay, but often doesn't actually deliver.
It is a high risk industry. Every season I've learned more about what
I should have done. But the problem is, when the season ends in one
state, you have to travel to another state if you want to keep going. I
didn't make enough money to do that. Also, when I went down to LA last
winter to "turn pro," I got skunked, made $0, and a Korean lady totalled
my car for me too. So I have bad feelings about hitting the road.
There have to be a lot more initiatives available than the paltry 1..2
I've been carrying up here, and I figure I need at least $1000 before
even considering hitting the road.
Signature gathering reminds me of open source software. Usually not
working.
Well, I'm quitting today. This went past the point of being a waste of
time awhile ago. I was just hanging on to work the Marysville parade,
and to make enough money to keep my landlord ever so barely placated for
maybe 2 more weeks. I would have made some real money in Marysville if
I had had a car. Never say never, but I probably won't do signature
gathering in the future if I don't have a car. Not unless I tried it
out and found it was a really easy measure to move. Show Me The Money.
The thing that continues to bug me about my various critics, is that I
know for fact they are not walking in my shoes, and I have my doubts
that they could. I'm never getting these criticisms from fellow
struggling indie game developers. Probably they know how hard
everything is, or are busy finding out, and really wouldn't think it's
their place to rag on somebody about all the bombs going off. The truly
brave have even written postmortems about how they failed, so that
others can maybe learn from their mistakes.
For instance,
http://www.bookofhook.com/Article/GameDevelopment/APyrogonPostmortem.html
Brian Hook was a contemporary of mine once upon a time. We used to
argue about low-level rendering stuff in comp.graphics.api.opengl, and
we had this huge excruciating technical debate that lasted a month and
ended in a draw. So I know he wasn't any smarter than I was. But he
went "full hog" into a 3D device driver career and I didn't. I didn't
like device drivers that much, I wanted something more creatively
fulfilling. Then Brian worked with ID Software for awhile. He made
good money, then started his own game studio with a very solid funding base.
He bombed. He pursued derivative work, i.e. things like "Bejeweled"
clones that everyone and their brother was trying to cash in on. He
never did the tough thing of pursuing a more distinctive vision, despite
a nebulous intent to do so. And, who knows, maybe device drivers and ID
Software wasn't the right background for doing terribly original work.
Quake has never been none for being innovative in terms of game desgin,
it's always a technology showcase. Anyways, whatever the reasons, more
skilled and better financed game developers than myself have fallen flat
on their faces.
I'd just love, for once, to be criticized by a successful *indie game
designer*. Just once. I might actually learn something, instead of
seeing the usual predictable primate posturing psychology.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
"The pioneer is the one with the arrows in his back."
- anonymous entrepreneur
Brandon J. Van Every wrote:
> We used to
> argue about low-level rendering stuff in comp.graphics.api.opengl, and
> we had this huge excruciating technical debate that lasted a month and
> ended in a draw. So I know he wasn't any smarter than I was.
Instant classic!
Brandon J. Van Every wrote:
>
> Well, considering that I just went out for an hour to gather signatures,
> and got skunked,
Just wondering: what is "being skunked"?
Quoth ···············@hotmail.com" <············@gmail.com>:
> Brandon J. Van Every wrote:
>>
>> Well, considering that I just went out for an hour to gather signatures,
>> and got skunked,
>
> Just wondering: what is "being skunked"?
It presumably involves being held down by irritated people to
introduce the person being held down to an even more irritated
skunk...
--
wm(X,Y):-write(X),write(·@'),write(Y). wm('cbbrowne','gmail.com').
http://cbbrowne.com/info/lsf.html
"Politics is not a bad profession. If you succeed there are many
rewards, if you disgrace yourself you can always write a book."
-- Ronald Reagan
THINGS YOU SHOULD KNOW ABOUT BRANDON J. VAN EVERY BEFORE REPLYING TO
ONE OF HIS POSTS
1. He has never designed any game, nor contributed to the design of
any game, which has ever seen the light of day, despite referring
to himself as a "game designer." (In rebuttal, he pointed out his
"one complete game" from "1983" on the "Atari 800" which he showed
to his "8th grade math teacher.")
2. He has never been employed in the game industry, in any way,
shape, manner or form. Despite this, for some reason he managed
to get named as an Independent Games Festival judge; a curious
turn of events, since their stated intent is to appoint
"professionals in the game industry" (their quote, not his).
3. In fact, the only programming job he had listed on his resume was
for only "2 years" ending in "1998," working in C and assembly on
a graphics driver, as a "Sr. Software Engineer" -- a curious
title, since this was his first (and only) job in the software
industry. There is no evidence he has used C++, nor any other
language, professionally. (And the company in question is
defunct, anyway, so there is no way to verify his claim.)
4. The other jobs he has mentioned having after this one and only
items on his resume are: "yard maintenance work," "painting
apartments," "scrubbing floors," "sub minimum wage signature
gathering," and working for "$5/hour at a Vietnamese restaurant."
5. The only personal project he actually wrote code for and made
available in some manner was Free3d, a software 3D rendering
engine. Stating that its goals were to be "100% efficient, 100%
portable" and to release it in a "one year time frame," which he
started in "1993" and abandoned in "1996," admitting that it
"barely drew even a single polygon" and "did hardly anything in
the 3D department."
6. Almost every Internet community (Usenet newsgroup, mailing list,
etc.) he has ever introduced himself to has resulted in him
repeating the same pattern: asking leading questions, demanding
people do things his way, becoming hostile, annoying the other
participants, alienating them, and finally leaving in disgust.
7. Of the projects (open source and otherwise) whose communities he
has (briefly) joined, he has never contributed anything tangible
in terms of code or documentation.
8. The project he has intermittently claimed to be working on, Ocean
Mars, is vaporware -- and is one of his admitted "failures." He
allegedly sunk "nine months of full time 60 hours/week" and about
"$80K" into it (at least; he "stopped counting") with only a
"spherical hexified icosahedron" display to show for it (only
allegedly, since it has never been shown or demonstrated
publicly).
9. Since his embarassing frustration with his Ocean Mars project, he
has decided that C and C++ aren't "worth anything as a resume
skill anymore," and embarked on a quest in 2003 to find a
high-level language that will suit his needs. After more than a
year, at least ten languages, and not having even "written a line
of code" in any of them, he still has yet to find a language that
will suit him.
10. Finally, despite vehemently insisting that he is not a troll, many
people quite understandingly have great difficulty distinguishing
his public behavior from that of a troll.
You have been trying out different implementations, toolkits, OSes,
etc. for quite a while now. As long as you are not writing any code,
why not document your findings as you go? It would be interesting to
have a giant chart filled with interesting bits like how fast a
implementation X of language Y was on platform Z, and the fact that
language Y has great libraries, which are not presently maintained. It
would be more interesting if you had written some code in some of the
languages, but at least if you document you will not forget what you
learn a year later, and you would be producing something of at least
arguable value, as opposed to nothing at all.
Brandon J. Van Every wrote:
> When I decided to switch from Windows to Linux, it was on the premise
> that this would actually save me labor over using Cygwin or MinGW. I am
> disturbed to find out how bad Linux actually is.
Well, most hardware vendors don't care enough for you, the
customer, that they don't give you (or anyone else) the details
needed to write a working device driver. So in a political sense,
using Linux sucks, yes, because you're on your own.
> Near as I can tell,
> Linux is not some consistent OS that has come a long ways since my early
> forays from 1993 to 1996. Rather, Linux is a gamble. It might work
> fine for you, or it might royally suck. Here, "sucks" means "more work
> than beating Cygwin or MinGW into shape." I thought it would be a good
> trade. I am surprised how bad a trade it turned out to be.
Nobody ever said Linux was consistent. It's a collection of free
(beer&speech) software tools that many people happen to like.
It's definitely NOT consistent, both in the tools working together
(or even having as much as a similar look and feel) and in terms
of quality.
> I've also become deeply averse to large stacks of dependent packages.
> This is common in open source land, and it forces one to get stuck on a
> particular OS. It's also the reason why Longhorn is shipping so late.
> Perhaps in time, both open source and Windows will evolve to be equally
> stacked and equally bad.
Isn't it normal that one kind of software DEPENDS on another, say,
utilities, data structures, a GUI library, maybe a database?
> I'm also realizing how irrelevant Linux is to shipping games. It isn't
> "an extra platform where I might pick up a few sales." It is an
> impossible platform.
I could have told you that ;)
Linuxers don't buy software. That's why they use Linux. But Unix
in general is a great platform for *development*, like most Lisps
run well on Unix, but not other systems. Just don't expect to
write commercial stuff FOR linux.
> In terms of consumer friendliness, it is complete
> junk, and incapable of being otherwise anytime soon. I can't even begin
> to fathom taking the tech support calls on this monstrosity. It makes
> the disparate world of Windows 3D drivers look like cake!
It's a volunteer-based system, with the advantages (lots of nice
features appear early in some window managers), and the
disadvantages (user-friendly it isn't).
> Plus Apple is moving to x86, and in time will do all the things that the
> Linux world won't do. I believe Linux has lost its opportunity to
> become relevant as a consumer desktop. It will continue to grow as a
> server OS, with expert administrators always in tow.
Maybe. I don't know the future.
> So, I'm nukeing my Linux partition and not looking back. Good riddance.
If you hate Windows, but want consistency and a viable market of
users willing to pay money (but not a huge market), get a (cheap)
Mac. Saves loads of time, in my experience the last 1.5 years
(and watching other people use XP (and suffer) during that time,
poor suckers).
If you want the huge market, you need to invest in the stuff
needed to develop for it, I guess.
--
Don't let school interfere with your education. -- Mark Twain
Ulrich Hobelmann <···········@web.de> writes:
> Linuxers don't buy software. That's why they use Linux. But Unix
Red Hat is hoping differently. :)
--
David Magda <dmagda at ee.ryerson.ca>
Because the innovator has for enemies all those who have done well under
the old conditions, and lukewarm defenders in those who may do well
under the new. -- Niccolo Machiavelli, _The Prince_, Chapter VI
David Magda wrote:
>>Linuxers don't buy software. That's why they use Linux. But Unix
> Red Hat is hoping differently. :)
Well, you can buy support for about any OS, either from the vendor, or
from independent consultants. Many people make their living by selling
support for Linux, BSD or other open-source software (like databases,
web application servers, etc.)
mkb.
Matthias Buelow wrote:
> David Magda wrote:
>
>
>>>Linuxers don't buy software. That's why they use Linux. But Unix
>>
>>Red Hat is hoping differently. :)
>
>
> Well, you can buy support for about any OS, either from the vendor, or
> from independent consultants. Many people make their living by selling
> support for Linux, BSD or other open-source software (like databases,
> web application servers, etc.)
But this kind of business only works as long as these softwares
NEED consulting (i.e. are crappy). If it were all drag-and-drop
to install, with user-friendly config panels and documentation
(like some Linux and Mac programs are), then nobody would need to
buy consulting. Note how nobody makes any money on games that are
open-source, or on web browsers.
--
Don't let school interfere with your education. -- Mark Twain
Ulrich Hobelmann wrote:
> Matthias Buelow wrote:
>
>> David Magda wrote:
>>
>>
>>>> Linuxers don't buy software. That's why they use Linux. But Unix
>>>
>>>
>>> Red Hat is hoping differently. :)
>>
>>
>>
>> Well, you can buy support for about any OS, either from the vendor, or
>> from independent consultants. Many people make their living by selling
>> support for Linux, BSD or other open-source software (like databases,
>> web application servers, etc.)
>
>
> But this kind of business only works as long as these softwares NEED
> consulting (i.e. are crappy).
Not necessarily. The software could be well suited to task, but require
a difficult learning curve that someone doesn't have time for. I agree,
however, that not being able to easily install a driver is crappy, and I
don't much care for a business model of "ask an expert or do it manually
with guru skills" when competitors like Windows and OSX handle the job
just fine.
> If it were all drag-and-drop to install,
> with user-friendly config panels and documentation (like some Linux and
> Mac programs are), then nobody would need to buy consulting.
Well, that's not true. There are still Windows consultants to train
people in the basics, after all. Also, no matter how good an
ease-of-use veneer, there's always an ugly OS in there somewhere.
Corporations in particular want the comfort level of knowing they can
turn to someone to deal with their problems.
It's much less of a business model in the consumer space. Consumers
expect things to be simple and just work. Actually, the whole GPL is
often silly in the consumer space. I can't see selling shrinkwrap
desktop games that way, for instance. It's as though the GPL was
designed in an era of big corporations and "big iron" with teams of
administrators running around holding the keys to the temple.
> Note how
> nobody makes any money on games that are open-source, or on web browsers.
The sad thing is, 90% of games don't make money at all. :-(
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
"The pioneer is the one with the arrows in his back."
- anonymous entrepreneur
Brandon J. Van Every wrote:
> Corporations in particular want the comfort level of knowing they can
> turn to someone to deal with their problems.
Heh. I think that's because they invariably can't attract good
people, who could do that work on their own. After all all those
really bad CS students, who get their degrees nonetheless, have to
end up somewhere. For the HUGE consulting fees you *could* employ
decent people full-time capable of useful work.
> It's much less of a business model in the consumer space. Consumers
> expect things to be simple and just work. Actually, the whole GPL is
> often silly in the consumer space. I can't see selling shrinkwrap
> desktop games that way, for instance. It's as though the GPL was
> designed in an era of big corporations and "big iron" with teams of
> administrators running around holding the keys to the temple.
It was mostly RMS working in the MIT media lab, and pissed off
because he didn't have the source to modify the system on his
computer, and because he couldn't pass the source to his friends.
In a way he was right. Sometimes I wish I had *full* access to my
Mac, for instance to change Dock, Finder etc.
>> Note how nobody makes any money on games that are open-source, or on
>> web browsers.
>
>
> The sad thing is, 90% of games don't make money at all. :-(
That's because all those games today are yet-another-3D game,
unlike all those *great* games they had for the Super Nintendo or
Mega Drive/Genesis. They are only expensive effects, with the fun
missing.
--
Don't let school interfere with your education. -- Mark Twain
"Brandon J. Van Every" <·····················@mycompanyname.com> writes:
>I agree, however, that not being able to easily install a driver is
>crappy, and I don't much care for a business model of "ask an expert
>or do it manually with guru skills" when competitors like Windows and
>OSX handle the job just fine.
OSX and Windows are not "competitors".
Gnu/Linux (and FreeBSD/NetBSD etc.) do not have "business models".
They are running outside of market competition, they do not strive to
make money for anyone (it doesn't matter if some companies like RedHat
try to siphon money from it, they don't make or break open-source),
and they do not have customers, since they are produced by the very
people who are the most eager users, and if anyone else likes the
software, that's ok; if not, who cares. Free software is written to
solve specific problems, and not for making profits.
You need to get that commercial crap out of your mind, or you'll get
nowhere with the free software community.
>I can't see selling shrinkwrap
>desktop games that way, for instance.
Oh noes! Tell Stallman right away, I bet he hasn't noticed yet! :)
It seems you don't quite understand the whole modus operandi of free
software (GPL or not).
mkb.
Matthias Buelow wrote:
> OSX and Windows are not "competitors".
Not direct competitors on the PC market, but I sold my PC running
XP and NetBSD for a Mac, so that makes them indirect competitors
at least.
> Gnu/Linux (and FreeBSD/NetBSD etc.) do not have "business models".
GNU has one: world domination ;)
(in fact that was/is Linus Torvalds' goal, too)
--
Don't let school interfere with your education. -- Mark Twain
Matthias Buelow wrote:
>
> OSX and Windows are not "competitors".
> Gnu/Linux (and FreeBSD/NetBSD etc.) do not have "business models".
At face value, your comment is nothing short of bizzare.
> Free software is written to
> solve specific problems, and not for making profits.
I don't think you've sufficiently studied the GPL or its history.
Freedom as in "free speech," not "free beer."
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
20% of the world is real.
80% is gobbledygook we make up inside our own heads.
"Brandon J. Van Every" <·····················@mycompanyname.com> writes:
>> OSX and Windows are not "competitors".
>> Gnu/Linux (and FreeBSD/NetBSD etc.) do not have "business models".
>
>At face value, your comment is nothing short of bizzare.
Aha. What's the "business model" of Linux?
>I don't think you've sufficiently studied the GPL or its history.
>Freedom as in "free speech," not "free beer."
Firstly, free software is a strict superset of GPL software. Then, of
course it is also the intention of the GPL to give things away gratis,
that is free of charge. This is a direct implication of the fact that
you are allowed to copy it without having to pay anything. Few people
would pay for what they can get for free, or wouldn't they? Of course,
you can make payments in the form of donations, or pay for
value-adding services. But not for the right to use and copy the
software. This is granted free of any charge (and hence, also free as
in "free beer").
mkb.
Ulrich Hobelmann wrote:
> But this kind of business only works as long as these softwares NEED
> consulting (i.e. are crappy). If it were all drag-and-drop to install,
> with user-friendly config panels and documentation (like some Linux and
> Mac programs are), then nobody would need to buy consulting. Note how
That's utopical. Many programs are complicated not because they are
badly programmed but because the problems they solve are inherently
complicated (see sendmail, for example, or any database system.)
Graphical configuration only goes so far. If the problem is complicated
enough, there's little difference, from a point of intellectual effort,
if you specify the rules graphically, or in a textual representation. In
many cases, the latter might actually be more convenient.
> nobody makes any money on games that are open-source, or on web browsers.
That's because open-source games are mostly played by, well, enthusiasts
(see Nethack for a typical example). Note that, for example, Doom3,
which sells for money, _does_ have tech-support hotline, and even a hint
line you can call (both for money, of course). And games are hardly on
the same level of configuration complexity as most professional
(especially server) software, or otherwise they wouldn't be a great
success. Some people even pay for web browsers; apparently Opera still
exists.
mkb.
Matthias Buelow wrote:
> Ulrich Hobelmann wrote:
>
>
>>But this kind of business only works as long as these softwares NEED
>>consulting (i.e. are crappy). If it were all drag-and-drop to install,
>>with user-friendly config panels and documentation (like some Linux and
>>Mac programs are), then nobody would need to buy consulting. Note how
>
>
> That's utopical. Many programs are complicated not because they are
> badly programmed but because the problems they solve are inherently
> complicated (see sendmail, for example, or any database system.)
> Graphical configuration only goes so far. If the problem is complicated
> enough, there's little difference, from a point of intellectual effort,
> if you specify the rules graphically, or in a textual representation. In
> many cases, the latter might actually be more convenient.
But with the right specification language something of sendmail
dimensions could actually be easy. See OpenBSD's pf for a *nice*
firewall, for instance. (Lisp syntax also comes to mind)
Most people of some education can probably create a lex spec,
maybe a yacc spec. Why shouldn't they be able to write sendmail
specs etc. without expensive consulting? Why couldn't a
reasonably structured grammar be represented in a graphical way?
--
Don't let school interfere with your education. -- Mark Twain
Ulrich Hobelmann <···········@web.de> writes:
>But with the right specification language something of sendmail
>dimensions could actually be easy. See OpenBSD's pf for a *nice*
>firewall, for instance. (Lisp syntax also comes to mind)
The purported inadequacy of sendmail's configuration language is a
myth on the proportions of "Lisp sucks because it's all parentheses".
For most sites, using the m4 templates is more than sufficient, and
that's basically the same like, for example, Postfix' config language.
The ruleset language gives sendmail the dynamic flexibility it needs
(or needed) to deal with the great variety of mail setups and mail
routings (it has been worse 20 years ago than it is now, when almost
everyone now is using SMTP over TCP/IP.)
mkb.
Matthias Buelow <···@incubus.de> writes:
> Ulrich Hobelmann <···········@web.de> writes:
>
> >But with the right specification language something of sendmail
> >dimensions could actually be easy. See OpenBSD's pf for a *nice*
> >firewall, for instance. (Lisp syntax also comes to mind)
>
> The purported inadequacy of sendmail's configuration language is a
> myth on the proportions of "Lisp sucks because it's all parentheses".
> For most sites, using the m4 templates is more than sufficient, and
> that's basically the same like, for example, Postfix' config language.
> The ruleset language gives sendmail the dynamic flexibility it needs
> (or needed) to deal with the great variety of mail setups and mail
> routings (it has been worse 20 years ago than it is now, when almost
> everyone now is using SMTP over TCP/IP.)
>
I couldn't agree more. I've never understood why everyone has such an
issue over sendmail's configuration. As you point out, the m4 macros
will do the job for nearly all sites. Even if you had to write all the
rulesets from scratch, its not that bad. I did it as an exercise about
10 years ago when I wanted to learn sendmail and get to understand
it. A few days withthe Oreilly sendmail book and I had a working
sendmail configuration which worked fine and did all those added bits,
like masquerading, spam checking with RBLs, blocking relaying etc.
That couple of days has proven to be very valuable in some many other
ways. It gave me a very good solid understanding of how SMTP works and
I've been able to move to configuring other systems, like postfix,
exim and qmail with no problems (though I never liked qmail for some
reason - just never felt like I really understood it in the sense I
could truely predict the full impact of config changes).
Tim
--
Tim Cross
The e-mail address on this message is FALSE (obviously!). My real e-mail is
to a company in Australia called rapttech and my login is tcross - if you
really need to send mail, you should be able to work it out!
Tim X <····@spamto.devnul.com> writes:
> I couldn't agree more. I've never understood why everyone has such an
> issue over sendmail's configuration. As you point out, the m4 macros
> will do the job for nearly all sites.
Personnally, I never could make the m4 stuff work on my hosts when I
used sendmail. But the rules were easy to understand, only hard to
read for the cryptic tokens.
> Even if you had to write all the
> rulesets from scratch, its not that bad. I did it as an exercise about
> 10 years ago when I wanted to learn sendmail and get to understand
> it. A few days withthe Oreilly sendmail book and I had a working
> sendmail configuration which worked fine and did all those added bits,
> like masquerading, spam checking with RBLs, blocking relaying etc.
>
> That couple of days has proven to be very valuable in some many other
> ways. It gave me a very good solid understanding of how SMTP works and
> I've been able to move to configuring other systems, like postfix,
> exim and qmail with no problems (though I never liked qmail for some
> reason - just never felt like I really understood it in the sense I
> could truely predict the full impact of config changes).
>
> Tim
--
__Pascal Bourguignon__ http://www.informatimago.com/
Litter box not here.
You must have moved it again.
I'll poop in the sink.
Tim X wrote:
> Matthias Buelow <···@incubus.de> writes:
>
>
>>Ulrich Hobelmann <···········@web.de> writes:
>>
>>
>>>But with the right specification language something of sendmail
>>>dimensions could actually be easy. See OpenBSD's pf for a *nice*
>>>firewall, for instance. (Lisp syntax also comes to mind)
>>
>>The purported inadequacy of sendmail's configuration language is a
>>myth on the proportions of "Lisp sucks because it's all parentheses".
>>For most sites, using the m4 templates is more than sufficient, and
>>that's basically the same like, for example, Postfix' config language.
>>The ruleset language gives sendmail the dynamic flexibility it needs
>>(or needed) to deal with the great variety of mail setups and mail
>>routings (it has been worse 20 years ago than it is now, when almost
>>everyone now is using SMTP over TCP/IP.)
>>
>
> I couldn't agree more. I've never understood why everyone has such an
> issue over sendmail's configuration. As you point out, the m4 macros
> will do the job for nearly all sites. Even if you had to write all the
> rulesets from scratch, its not that bad. I did it as an exercise about
> 10 years ago when I wanted to learn sendmail and get to understand
> it. A few days withthe Oreilly sendmail book and I had a working
> sendmail configuration which worked fine and did all those added bits,
> like masquerading, spam checking with RBLs, blocking relaying etc.
Just two words: DAYS, and BOOK. ;)
But I agree, some of it is inherent complexity, too. But still
there should IMHO be easier ways to set up a simple mail server.
Maybe I should write one someday (yay, another thing on my
infinite todo list...). But maybe the standard needs reworking,
too. Yesterday I read the FTP spec and kept thinking "what were
they smoking?" Half of it is old useless baggage, and then there
are problems that are totally unnecessary. Whatever happened to
designing a protocol that solves one particular problem, not the
kitchen sink?
I bet HTTP and SMTP are even worse.
> That couple of days has proven to be very valuable in some many other
> ways. It gave me a very good solid understanding of how SMTP works and
> I've been able to move to configuring other systems, like postfix,
> exim and qmail with no problems (though I never liked qmail for some
> reason - just never felt like I really understood it in the sense I
> could truely predict the full impact of config changes).
Sounds good.
--
Don't let school interfere with your education. -- Mark Twain
Ulrich Hobelmann <···········@web.de> writes:
> Tim X wrote:
> > Matthias Buelow <···@incubus.de> writes:
> >
> >>Ulrich Hobelmann <···········@web.de> writes:
> Just two words: DAYS, and BOOK. ;)
>
> But I agree, some of it is inherent complexity, too. But still there
> should IMHO be easier ways to set up a simple mail server. Maybe I
> should write one someday (yay, another thing on my infinite todo
> list...). But maybe the standard needs reworking, too. Yesterday I
> read the FTP spec and kept thinking "what were they smoking?" Half of
> it is old useless baggage, and then there are problems that are totally
> unnecessary. Whatever happened to designing a protocol that solves one
> particular problem, not the kitchen sink?
>
Sometimes complexity is unavoidable. OTOH the ftp spec is a royal PITA.
Gregm
Greg Menke <············@toadmail.com> writes:
> Ulrich Hobelmann <···········@web.de> writes:
>
> > Tim X wrote:
> > > Matthias Buelow <···@incubus.de> writes:
> > >
> > >>Ulrich Hobelmann <···········@web.de> writes:
> > Just two words: DAYS, and BOOK. ;)
> >
> > But I agree, some of it is inherent complexity, too. But still there
> > should IMHO be easier ways to set up a simple mail server. Maybe I
> > should write one someday (yay, another thing on my infinite todo
> > list...). But maybe the standard needs reworking, too. Yesterday I
> > read the FTP spec and kept thinking "what were they smoking?" Half of
> > it is old useless baggage, and then there are problems that are totally
> > unnecessary. Whatever happened to designing a protocol that solves one
> > particular problem, not the kitchen sink?
> >
>
> Sometimes complexity is unavoidable. OTOH the ftp spec is a royal PITA.
>
Its very very rare to read a protocol spec which isn't a
pain. Even relatively simple protocols, like http are a pain
when described within a formal spec. The problem is with trying to
balance clarity/brevity without ambiguity - something human language
is not great at. This is often made worse by the unfortunate reality
that often those who are very talented in a technical sense are often
somewhat less talented with written expression.
The FTP protocol specification does seem overly complex in today's
world of high speed connections and fast servers. However, if you put
it in context with the time it was defined - in a world of slow and
less reliable networks, slow cpus and little security concerns, it
sort of makes sense.
The other thing to wonder is that if it is such a bad protocol, why is
it still so widely in use? Is it just due to so many utilities etc
which support it or that it works and nothing better has come along?
Tim
--
Tim Cross
The e-mail address on this message is FALSE (obviously!). My real e-mail is
to a company in Australia called rapttech and my login is tcross - if you
really need to send mail, you should be able to work it out!
Tim X <····@spamto.devnul.com> writes:
> Greg Menke <············@toadmail.com> writes:
>
> >
> > Sometimes complexity is unavoidable. OTOH the ftp spec is a royal PITA.
> >
> Its very very rare to read a protocol spec which isn't a
> pain. Even relatively simple protocols, like http are a pain
> when described within a formal spec. The problem is with trying to
> balance clarity/brevity without ambiguity - something human language
> is not great at. This is often made worse by the unfortunate reality
> that often those who are very talented in a technical sense are often
> somewhat less talented with written expression.
Got that right. The IEEE 802.* specs will make your brains leak out
your ears. They approach write-ohly documentation but if you keep
slogging away its possible to extract content from them.
> The FTP protocol specification does seem overly complex in today's
> world of high speed connections and fast servers. However, if you put
> it in context with the time it was defined - in a world of slow and
> less reliable networks, slow cpus and little security concerns, it
> sort of makes sense.
The big complicated horribleness of FTP isn't related to IO speeds, but
is more associated with its built-in conversion features that attempt to
correctly transfer datasets between systems with very different concepts
of what files are- blocks here, bytes there, different coding etc.. Now
that we have a nearly monocultural concept of filesystems, all those
conversion features are pretty much vestigal.
>
> The other thing to wonder is that if it is such a bad protocol, why is
> it still so widely in use? Is it just due to so many utilities etc
> which support it or that it works and nothing better has come along?
Its bad because the command/data protocols occur over multiple TCP
connections that are set up in different directions, which are not well
suited to modern firewalled networks. Authentication happens in plain
text, its also difficult- or at least subtle to secure. Its popularity
is due to its ubiquity and a very minimally conformant implementation
will pretty much work on very similar architectures.
We don't allow telnet, ftp or any of the r* tools to traverse the
routers on our enterprise networks- terminal and file transfer
activities are ssh-based or nothing, and our network isn't on the
security bleeding edge, so I'd say ftp is pretty well gone as far as
we're concerned.
Greg
Greg Menke <············@toadmail.com> writes:
> Tim X <····@spamto.devnul.com> writes:
>
> > Greg Menke <············@toadmail.com> writes:
> >
> > >
> > > Sometimes complexity is unavoidable. OTOH the ftp spec is a royal PITA.
> > >
> > Its very very rare to read a protocol spec which isn't a
> > pain. Even relatively simple protocols, like http are a pain
> > when described within a formal spec. The problem is with trying to
> > balance clarity/brevity without ambiguity - something human language
> > is not great at. This is often made worse by the unfortunate reality
> > that often those who are very talented in a technical sense are often
> > somewhat less talented with written expression.
>
> Got that right. The IEEE 802.* specs will make your brains leak out
> your ears. They approach write-ohly documentation but if you keep
> slogging away its possible to extract content from them.
>
>
> > The FTP protocol specification does seem overly complex in today's
> > world of high speed connections and fast servers. However, if you put
> > it in context with the time it was defined - in a world of slow and
> > less reliable networks, slow cpus and little security concerns, it
> > sort of makes sense.
>
> The big complicated horribleness of FTP isn't related to IO speeds, but
> is more associated with its built-in conversion features that attempt to
> correctly transfer datasets between systems with very different concepts
> of what files are- blocks here, bytes there, different coding etc.. Now
> that we have a nearly monocultural concept of filesystems, all those
> conversion features are pretty much vestigal.
>
hmmm, why is this voice in my head saying "pathnames" right now?
> > The other thing to wonder is that if it is such a bad protocol, why is
> > it still so widely in use? Is it just due to so many utilities etc
> > which support it or that it works and nothing better has come along?
>
> Its bad because the command/data protocols occur over multiple TCP
> connections that are set up in different directions, which are not well
> suited to modern firewalled networks. Authentication happens in plain
> text, its also difficult- or at least subtle to secure. Its popularity
> is due to its ubiquity and a very minimally conformant implementation
> will pretty much work on very similar architectures.
Yes, that is a royal pain, but in context, the firewall and security
issue was barely even considered back when the ftp protocol was
defined - back then, it was still a community cnsidered to be made up
pretty much of trusted techies
> We don't allow telnet, ftp or any of the r* tools to traverse the
> routers on our enterprise networks- terminal and file transfer
> activities are ssh-based or nothing, and our network isn't on the
> security bleeding edge, so I'd say ftp is pretty well gone as far as
> we're concerned.
>
Oh how I envy your position. We keep trying to do this, but keep
having to "soften" our stance because of "critical business needs". In
the main, its still bloody Windows applicaitons which still rely on
telnet/ftp/rsh type protocols. All the Unix/Linux apps we run are fine
and have decent "security awareness". However, many of the windows
apps seem to use DLLs which provide telnet or ftp type protocol
support, but not ssh/ssl/tls etc. When we contact the vendors and ask
them to provide secure comms channels in their software, they sort of
go silent and then ask something like "what do you mean".
In fact, I can almost guess what sort of developer and what platform
whenver they ask to access our network to help track down a
problem. The Unix/Linux ones generally ask for VPN access or at least
ssh. The windows developers usually ask for telnet and then for
instructions on how to setup putty!
Meanwhile, I have constant arguments with management - I'm responsible
for ensuring data security and integrity, but then get forced to open
up firewalls etc to allow insecure applications because they are
"business critical". Consequently, I'm considered by them to be a tech
head with no real understanding of business and the real world and I
consider them to be unthinking fools who think Gartner is god!
Tim
--
Tim Cross
The e-mail address on this message is FALSE (obviously!). My real e-mail is
to a company in Australia called rapttech and my login is tcross - if you
really need to send mail, you should be able to work it out!
Tim X <····@spamto.devnul.com> writes:
> Greg Menke <············@toadmail.com> writes:
>
> > Tim X <····@spamto.devnul.com> writes:
> >
> > > Greg Menke <············@toadmail.com> writes:
> > > The FTP protocol specification does seem overly complex in today's
> > > world of high speed connections and fast servers. However, if you put
> > > it in context with the time it was defined - in a world of slow and
> > > less reliable networks, slow cpus and little security concerns, it
> > > sort of makes sense.
> >
> > The big complicated horribleness of FTP isn't related to IO speeds, but
> > is more associated with its built-in conversion features that attempt to
> > correctly transfer datasets between systems with very different concepts
> > of what files are- blocks here, bytes there, different coding etc.. Now
> > that we have a nearly monocultural concept of filesystems, all those
> > conversion features are pretty much vestigal.
> >
>
> hmmm, why is this voice in my head saying "pathnames" right now?
I couldn't say. Perhaps its time to experiment with a hat made of
aluminum foil? ;)
> >
> > Its bad because the command/data protocols occur over multiple TCP
> > connections that are set up in different directions, which are not well
> > suited to modern firewalled networks. Authentication happens in plain
> > text, its also difficult- or at least subtle to secure. Its popularity
> > is due to its ubiquity and a very minimally conformant implementation
> > will pretty much work on very similar architectures.
>
> Yes, that is a royal pain, but in context, the firewall and security
> issue was barely even considered back when the ftp protocol was
> defined - back then, it was still a community cnsidered to be made up
> pretty much of trusted techies
True, so I guess I'd put the difficult to secure aspect in as a 2nd
order issue. OTOH, plaintext authentication is certainly a problem from
day 1. If you're going to authenticate, then there must be provision
for doing better than that.
> > We don't allow telnet, ftp or any of the r* tools to traverse the
> > routers on our enterprise networks- terminal and file transfer
> > activities are ssh-based or nothing, and our network isn't on the
> > security bleeding edge, so I'd say ftp is pretty well gone as far as
> > we're concerned.
> >
> Oh how I envy your position. We keep trying to do this, but keep
> having to "soften" our stance because of "critical business needs". In
> the main, its still bloody Windows applicaitons which still rely on
> telnet/ftp/rsh type protocols. All the Unix/Linux apps we run are fine
> and have decent "security awareness". However, many of the windows
> apps seem to use DLLs which provide telnet or ftp type protocol
> support, but not ssh/ssl/tls etc. When we contact the vendors and ask
> them to provide secure comms channels in their software, they sort of
> go silent and then ask something like "what do you mean".
Too true. Thankfully the "advanced business software" which can't
handle crypto and insists on incomplete or partially secured
infrastructure can be dumped behind a firewall. But then there are the
wankers who really MUST export their broken-ass protocols...
> In fact, I can almost guess what sort of developer and what platform
> whenver they ask to access our network to help track down a
> problem. The Unix/Linux ones generally ask for VPN access or at least
> ssh. The windows developers usually ask for telnet and then for
> instructions on how to setup putty!
I think my Windows users are similar to yours. I have been fortunate
enough for the moment to give a resounding NO to the ftp and telnet
question. Watching Windows people dink around with notepad and telnet
is good for a laugh though- they're IT Knowledge Workers... wheee!
> Meanwhile, I have constant arguments with management - I'm responsible
> for ensuring data security and integrity, but then get forced to open
> up firewalls etc to allow insecure applications because they are
> "business critical". Consequently, I'm considered by them to be a tech
> head with no real understanding of business and the real world and I
> consider them to be unthinking fools who think Gartner is god!
Gah.. well its got to be that or some UML idiot...
Gregm
Greg Menke <············@toadmail.com> wrote:
+---------------
| Tim X <····@spamto.devnul.com> writes:
| > Even relatively simple protocols, like http are a pain
| > when described within a formal spec. The problem is with trying to
| > balance clarity/brevity without ambiguity - something human language
| > is not great at. This is often made worse by the unfortunate reality
| > that often those who are very talented in a technical sense are often
| > somewhat less talented with written expression.
|
| Got that right. The IEEE 802.* specs will make your brains leak out
| your ears. They approach write-ohly documentation but if you keep
| slogging away its possible to extract content from them.
+---------------
Whereas the Xerox-written [despite being a "DEC/Intel/Xerox" co-effort]
Ethernet 1.0 & 2.0 specs that pre-dated 802.3 were *EXTREMELY* readable,
and presented the same information three different ways: plain English,
Pascal-like pseudo-code, and engineering-style electrical specs. Yet for
all their clarity and redundancy, they were a *lot* shorter than 802.3.
Hint: The Ethernet specs weren't written by "A Standards Committee"...
-Rob
-----
Rob Warnock <····@rpw3.org>
627 26th Avenue <URL:http://rpw3.org/>
San Mateo, CA 94403 (650)572-2607
····@rpw3.org (Rob Warnock) writes:
> Greg Menke <············@toadmail.com> wrote:
> +---------------
> | Tim X <····@spamto.devnul.com> writes:
> | > Even relatively simple protocols, like http are a pain
> | > when described within a formal spec. The problem is with trying to
> | > balance clarity/brevity without ambiguity - something human language
> | > is not great at. This is often made worse by the unfortunate reality
> | > that often those who are very talented in a technical sense are often
> | > somewhat less talented with written expression.
> |
> | Got that right. The IEEE 802.* specs will make your brains leak out
> | your ears. They approach write-ohly documentation but if you keep
> | slogging away its possible to extract content from them.
> +---------------
>
> Whereas the Xerox-written [despite being a "DEC/Intel/Xerox" co-effort]
> Ethernet 1.0 & 2.0 specs that pre-dated 802.3 were *EXTREMELY* readable,
> and presented the same information three different ways: plain English,
> Pascal-like pseudo-code, and engineering-style electrical specs. Yet for
> all their clarity and redundancy, they were a *lot* shorter than 802.3.
>
> Hint: The Ethernet specs weren't written by "A Standards Committee"...
>
And they probably weren't encumbered by a formal standards procedure
which elevates process to the same level (or higher) as content. Once
you get the wranglers involved theres no upper limit to the officious
formalism and arguments for simplicity are discarded out of hand. It
happens all over, not just IEEE.
Gregm
Rob Warnock wrote:
> Greg Menke <············@toadmail.com> wrote:
> +---------------
> | Tim X <····@spamto.devnul.com> writes:
> | > Even relatively simple protocols, like http are a pain
> | > when described within a formal spec. The problem is with trying to
> | > balance clarity/brevity without ambiguity - something human language
> | > is not great at. This is often made worse by the unfortunate reality
> | > that often those who are very talented in a technical sense are often
> | > somewhat less talented with written expression.
> |
> | Got that right. The IEEE 802.* specs will make your brains leak out
> | your ears. They approach write-ohly documentation but if you keep
> | slogging away its possible to extract content from them.
> +---------------
>
> Whereas the Xerox-written [despite being a "DEC/Intel/Xerox" co-effort]
> Ethernet 1.0 & 2.0 specs that pre-dated 802.3 were *EXTREMELY* readable,
> and presented the same information three different ways: plain English,
> Pascal-like pseudo-code, and engineering-style electrical specs. Yet for
> all their clarity and redundancy, they were a *lot* shorter than 802.3.
>
> Hint: The Ethernet specs weren't written by "A Standards Committee"...
>
Are the Xerox written specs available online somewhere?
Kristian
Kristian Elof S�rensen <····@image.dk> wrote:
+---------------
| Rob Warnock wrote:
| > Whereas the Xerox-written [despite being a "DEC/Intel/Xerox" co-effort]
| > Ethernet 1.0 & 2.0 specs that pre-dated 802.3 were *EXTREMELY* readable,
|
| Are the Xerox written specs available online somewhere?
+---------------
Sorry, no, not to my knowledge. The DEC/Intel/Xerox (DIX) "Blue Book"
Ethernet spec *should* still be available from whatever successor entity
used to be the Xerox PARC Library or Document Center. You would want
"Version 2.0 (1982)", not "Version 1.0 (1980)". [Sorry I don't have the
exact document number -- my copy's at home & I'm not, at the moment.]
Or get both, and compare. One of the main main differences
is that v.2 is somewhat closer to 802.3 in the behavior of
repeaters w.r.t. regenerating lost preamble bits.
-Rob
-----
Rob Warnock <····@rpw3.org>
627 26th Avenue <URL:http://rpw3.org/>
San Mateo, CA 94403 (650)572-2607
····@rpw3.org (Rob Warnock) writes:
> Greg Menke <············@toadmail.com> wrote:
> +---------------
> | Tim X <····@spamto.devnul.com> writes:
> | > Even relatively simple protocols, like http are a pain
> | > when described within a formal spec. The problem is with trying to
> | > balance clarity/brevity without ambiguity - something human language
> | > is not great at. This is often made worse by the unfortunate reality
> | > that often those who are very talented in a technical sense are often
> | > somewhat less talented with written expression.
> |
> | Got that right. The IEEE 802.* specs will make your brains leak out
> | your ears. They approach write-ohly documentation but if you keep
> | slogging away its possible to extract content from them.
> +---------------
>
> Whereas the Xerox-written [despite being a "DEC/Intel/Xerox" co-effort]
> Ethernet 1.0 & 2.0 specs that pre-dated 802.3 were *EXTREMELY* readable,
> and presented the same information three different ways: plain English,
> Pascal-like pseudo-code, and engineering-style electrical specs. Yet for
> all their clarity and redundancy, they were a *lot* shorter than 802.3.
>
> Hint: The Ethernet specs weren't written by "A Standards Committee"...
>
>
Since a camel is a horse designed by a committee, maybe we should
classify specifications as horse-based and camel-based to destinguish
well written clear specifications from confusing, bloated, hard to
follow ones. :-)
Actually, I suspect the real problem is that often the types of people
who get involved in committees are exactly the sorts of people who
should NOT be there, while those who really understand the goal, have
the right deep technical understanding and are able to communicate
well, are not interested because by the time they understand the
problem domain the spec applies to, they are already a bit bored and
have moved on to something else. Of course this is a major
generalisation and there are some very dedicated and talented people
who do get involved in trying to formulate clear specifications and a
tip my hat to them. Just look at the history associated with getting
an accepted standard specification for something like CL - I don't
think I would have had the patience to stick it out (putting aside my
lack of deep understanding of lisp).
Tim
--
Tim Cross
The e-mail address on this message is FALSE (obviously!). My real e-mail is
to a company in Australia called rapttech and my login is tcross - if you
really need to send mail, you should be able to work it out!
Tim X wrote:
>>Sometimes complexity is unavoidable. OTOH the ftp spec is a royal PITA.
>>
Actually the spec is *really* ok, the protocol is dumb, though,
IMHO (not simple (and specified!) enough for what it does, and
full of garbage).
> The FTP protocol specification does seem overly complex in today's
> world of high speed connections and fast servers. However, if you put
> it in context with the time it was defined - in a world of slow and
> less reliable networks, slow cpus and little security concerns, it
> sort of makes sense.
Not sure about that. Lots of stuff applies to other "file
systems" with records etc. Ok, only I wouldn't put that into a
FILE transfer thing in the first place. Let the server sort it
out what to do with the data it is sent. Then the connection
thing: why not just send the data in-line, after telling the
length of it? Why create new connections (with all that
active/passive hassle) and with the problems mentioned it the spec
regarding "image mode" (lost connection looks just like EOF)?
> The other thing to wonder is that if it is such a bad protocol, why is
> it still so widely in use? Is it just due to so many utilities etc
> which support it or that it works and nothing better has come along?
If you only implement the basic handful of commands, it's ok, and
there's lots of software for it. Nobody writes a new protocol,
with a BSD-licensed open-source implementation, so nobody cares.
It might be a pain, but then it's not *that* hard to just
implement it and shut up ;)
--
Don't let school interfere with your education. -- Mark Twain
Ulrich Hobelmann <···········@web.de> writes:
>Not sure about that. Lots of stuff applies to other "file systems"
>with records etc. Ok, only I wouldn't put that into a FILE transfer
>thing in the first place. Let the server sort it out what to do with
One should remember that the FTP protocol was not invented on Unix.
>the data it is sent. Then the connection thing: why not just send the
>data in-line, after telling the length of it? Why create new
>connections (with all that active/passive hassle) and with the
>problems mentioned it the spec regarding "image mode" (lost connection
>looks just like EOF)?
Maybe that's because the initiator (which opens the control channel)
is not necessarily the same as the target for the data.
mkb.
Matthias Buelow wrote:
> One should remember that the FTP protocol was not invented on Unix.
Sure, but as far as files make sense, they should be used,
different protocol for different purposes (records etc.), IMHO.
> Maybe that's because the initiator (which opens the control channel)
> is not necessarily the same as the target for the data.
That's a point, but I'd rather have a simpler protocol, and if you
want to transfer stuff from B to C, log in to one of them, and
start the Simple FTP ;)
That's even simpler than logging in to both and arranging a
connection between them.
--
Don't let school interfere with your education. -- Mark Twain
Ulrich Hobelmann <···········@web.de> writes:
> Tim X wrote:
> > Matthias Buelow <···@incubus.de> writes:
> >
> >>Ulrich Hobelmann <···········@web.de> writes:
> >>
> >>
> >>>But with the right specification language something of sendmail
> >>>dimensions could actually be easy. See OpenBSD's pf for a *nice*
> >>>firewall, for instance. (Lisp syntax also comes to mind)
> >>
> >>The purported inadequacy of sendmail's configuration language is a
> >>myth on the proportions of "Lisp sucks because it's all parentheses".
> >>For most sites, using the m4 templates is more than sufficient, and
> >>that's basically the same like, for example, Postfix' config language.
> >>The ruleset language gives sendmail the dynamic flexibility it needs
> >>(or needed) to deal with the great variety of mail setups and mail
> >>routings (it has been worse 20 years ago than it is now, when almost
> >>everyone now is using SMTP over TCP/IP.)
> >>
> > I couldn't agree more. I've never understood why everyone has such an
> > issue over sendmail's configuration. As you point out, the m4 macros
> > will do the job for nearly all sites. Even if you had to write all the
> > rulesets from scratch, its not that bad. I did it as an exercise about
> > 10 years ago when I wanted to learn sendmail and get to understand
> > it. A few days withthe Oreilly sendmail book and I had a working
> > sendmail configuration which worked fine and did all those added bits,
> > like masquerading, spam checking with RBLs, blocking relaying etc.
>
> Just two words: DAYS, and BOOK. ;)
You have got to be kidding! You expect to be able to configure
something as powerful as sendmail without spending a little time doing
some research? Either you have no clue about what is involved in
configuring server software or you live in some fairy land where
everything just happens without thinking about it.
In fact, the more I think about it, the more I realise you have not a
clue what your talking about. Have you ever looked at sendmail? Note
that I am talking about writing the whole config file from scratch,
which you DO NOT NEED TO DO to configure sendmail. As mentioned, the
whole system can be configured using the m4 macros in about 30
minutes, which includes reading the documentation on the various
options you can use in the master m4 config file. The few days and a
book was what I needed to write a whole config file, including all the
sendmail rule sets, from scratch, which was done as a learning
experience not as a necessity.
Tim
--
Tim Cross
The e-mail address on this message is FALSE (obviously!). My real e-mail is
to a company in Australia called rapttech and my login is tcross - if you
really need to send mail, you should be able to work it out!
Tim X wrote:
> You have got to be kidding! You expect to be able to configure
> something as powerful as sendmail without spending a little time doing
> some research? Either you have no clue about what is involved in
> configuring server software or you live in some fairy land where
> everything just happens without thinking about it.
I haven't read the mentioned book, so no, I don't know the
innermost details of sendmail config.
> In fact, the more I think about it, the more I realise you have not a
> clue what your talking about. Have you ever looked at sendmail? Note
> that I am talking about writing the whole config file from scratch,
> which you DO NOT NEED TO DO to configure sendmail. As mentioned, the
> whole system can be configured using the m4 macros in about 30
> minutes, which includes reading the documentation on the various
> options you can use in the master m4 config file.
I did that once, yes, and it didn't take long. However, just like
when configuring an Apache, I feel lost inbetween the options. A
much better example is Linux. With the good old Tcl/Tk Xconfig
even the first time I configured my own kernel didn't take too
long (ok, reading all the docu) and was quite comfortable. This
would be what I'd like for * servers too.
After all, what huge amounts of work does a mail server do? Not
THAT much, accepting SMTPs, putting it into files for local users,
having an outgoing queue that gets delivered if the remote server
is responsive, then some firewall-like filtering options, like
what mail to reject. Where's the immense inherent complexity of
the application domain? Where would it be in a web server, or in
an FTP?
> The few days and a
> book was what I needed to write a whole config file, including all the
> sendmail rule sets, from scratch, which was done as a learning
> experience not as a necessity.
But for all non-exotic setups I expect to be able to config a
server for *anything* in an hour. We're not talking setting
obscure Oracle flags here, to optimize your system ;)
--
Don't let school interfere with your education. -- Mark Twain
Ulrich Hobelmann <···········@web.de> writes:
>After all, what huge amounts of work does a mail server do? Not THAT
>much, accepting SMTPs, putting it into files for local users, having
>an outgoing queue that gets delivered if the remote server is
>responsive, then some firewall-like filtering options, like what mail
>to reject. Where's the immense inherent complexity of the application
>domain? Where would it be in a web server, or in an FTP?
Ok, you have no clue. Come back to this topic, when you have set up a
couple mail servers in complex environments. Until then, don't let
your naive peconceptions about how things (ought) to work misguide
you.
mkb.
From: Bulent Murtezaoglu
Subject: Re: I've thought better of Linux
Date:
Message-ID: <87fyvfy72g.fsf@p4.internal>
>>>>> "UH" == Ulrich Hobelmann <···········@web.de> writes:
[...]
UH> After all, what huge amounts of work does a mail server do?
I wouln't say huge, but it does more than you imply.
UH> Not THAT much, accepting SMTPs,
Except sendmail was supposed to handle other protocols and do the right
thing including things like accepting UUCP mail, and relaying it
through SMTP gateway to a bitnet host etc. While I am not fond of the
odd sendmail.cf language (which M4 does hide, but M4 itself is no monument
to elegance either) there is a reason why it is there.
UH> putting it into files for local users,
Except it when it doesn't. The mail can be properly local in the
lowest visible MX sense with the spool being elsewhere, or the
localness might be decided by some failure-prone mechanism (LDAP, split
DNS MX records, NIS etc.). Or maybe there is no spool for the user
indicated? (eg an alias, autoresponder with its own rejection
mechanisms, or a list etc.)
UH> ... having an outgoing queue that gets delivered if
UH> the remote server is responsive
Except you can't do this in a straightforward manner. You have to
refrain from creating a thundering herd problem, deal with flaky or
misconfigured DNS (so perhaps you know the name of the machine you'd
like to send to, but you cannot get the IP. What now? AFAIR Qmail
would fail in odd ways in this sceario and would bounce deliverable
mail.) etc. etc.
UH> then some firewall-like
UH> filtering options, like what mail to reject.
Which of course is a non-trivial problem if you want to do it right.
And remember my 'right' is different than your 'right.' The customer
is always right but it doesn't mean they know what they want or
communicate it properly even when they do. And there is more than one
of them.
UH> Where's the
UH> immense inherent complexity of the application domain? Where
UH> would it be in a web server, or in an FTP?
I don't know if I'd call it _immense_ but managing to get mail to be
realiable in a hostile network where your peers might suffer from
all manner of brokenness is more complex than ftp or http.
[...]
UH> But for all non-exotic setups I expect to be able to config a
UH> server for *anything* in an hour. [...]
I don't expect any such thing unless I have had some success or,
preferably some well-understood _failures_ in doing so previously.
My fear would be getting something working _by accident_ and having to
learn what that accident was when the thing fails at a critical time.
I am _not_ saying there isn't any gratuitous complexity in sendmail, or
that the present mail architecture is well-designed. But for the world
we are living in such things don't 'just work' after being configured in
a DWIM fashion. Otherwise reasonable expectations don't change that
fact.
cheers,
BM
Ulrich Hobelmann <···········@web.de> writes:
> After all, what huge amounts of work does a mail server do? Not THAT
> much, accepting SMTPs, putting it into files for local users, having
> an outgoing queue that gets delivered if the remote server is
> responsive, then some firewall-like filtering options, like what mail
> to reject. Where's the immense inherent complexity of the application
> domain? Where would it be in a web server, or in an FTP?
Who needs anything else than telnet remote.host 25 to send email?
And to receive:
#!/bin/bash
#*****************************************************************************
#FILE: smtpd.sh
#LANGUAGE: sh
#SYSTEM: POSIX
#USER-INTERFACE: NONE
#DESCRIPTION
#
# A very SIMPLE mail transfer agent.
#
# Put in inetd.conf:
# smtp stream tcp nowait mail /usr/local/bin/smtpd.sh smtpd.sh
#
#AUTHORS
# <PJB> Pascal Bourguignon <···@informatimago.com>
#MODIFICATIONS
# 2003-09-04 <PJB> Created.
#BUGS
#LEGAL
# GPL
#
# Copyright Pascal Bourguignon 2003 - 2003
#
# This program is free software; you can redistribute it and/or
# modify it under the terms of the GNU General Public License
# as published by the Free Software Foundation; either version
# 2 of the License, or (at your option) any later version.
#
# This program is distributed in the hope that it will be
# useful, but WITHOUT ANY WARRANTY; without even the implied
# warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR
# PURPOSE. See the GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public
# License along with this program; if not, write to the Free
# Software Foundation, Inc., 59 Temple Place, Suite 330,
# Boston, MA 02111-1307 USA
#*****************************************************************************
mailboxes=/tmp/var/spool/mail
mkdir -p "$mailboxes" >/dev/null 2>&1 || true
echo "220 `hostname` SMTP SIMPLE mail transport agent Version 1.0"
done=0
while [ $done -eq 0 ] ; do
read command arguments
command="$(echo "$command" | tr -d '\015')"
arguments="$(echo "$arguments" | tr -d '\015')"
case "$command" in
[hH][eE][lL][oO]) echo "250 Hi $arguments" ; done=1 ;;
*) echo "503 Bad command" ;;
esac
done
s_idle=0 ; s_from=1 ; s_rcpt=2
state=$s_idle
while true ; do
read command arguments
command="$(echo "$command" | tr -d '\015')"
arguments="$(echo "$arguments" | tr -d '\015')"
case "$command" in
[Mm][Aa][Ii][Ll])
if [ 0 -eq $(( $state & $s_from )) ] ; then
from=$(echo "$arguments" | sed -e 's/^ *[Ff][Rr][Oo][Mm] *: *//')
state=$(( $state | $s_from ))
echo "250 OK"
else
echo "503 Bad sequence of commands."
fi
;;
[Rr][Cc][Pp][Tt])
if [ $s_from -eq $(( $state & ( $s_from | $s_rcpt ) )) ] ; then
to=$(echo "$arguments" | sed -e 's/^ *[Tt][Oo] *: *//')
state=$(( $state | $s_rcpt ))
echo "250 OK"
else
echo "503 Bad sequence of commands."
fi
;;
[Dd][Aa][Tt][Aa])
if [ $(( $s_from | $s_rcpt )) -eq $(( $state & ( $s_from | $s_rcpt ) )) ] ; then
file="${mailboxes}/$(echo "$to" | sed -e 's/^ *< *//' -e 's/ *> *$/\1/' -e ···@.*//')"
echo "From $from `date '+%a %b %d %H:%M:%S %Y'`" >> "$file"
echo "354 Enter mail, end with <CR><LF>.<CR><LF>"
data=1
while [ $data -eq 1 ] ; do
read line
line="$(echo "$line" | tr -d '\015')"
if [ "$line" = "." ] ; then
data=0
state=$s_idle
echo "" >> "$file"
echo "250 OK"
else
echo "$line" >> "$file"
fi
done
else
echo "503 Bad sequence of commands."
fi
;;
[Nn][Oo][Oo][Pp]) echo "250 OK" ;;
[Rr][Ss][Ee][Tt]) state=$s_idle ; echo "250 OK" ;;
[Vv][Rr][Ff][Yy]) echo "252 Cannot VRFY." ;;
[Qq][Uu][Ii][Tt]) echo "221 Bye" ; exit 0 ;;
*) echo "502 Command not implemented." ;;
esac
done
#### smtpd.sh -- 2003-09-04 19:20:50 -- pascal ####
--
__Pascal Bourguignon__ http://www.informatimago.com/
Kitty like plastic.
Confuses for litter box.
Don't leave tarp around.
Ulrich Hobelmann <···········@web.de> writes:
>
> After all, what huge amounts of work does a mail server do? Not THAT
> much, accepting SMTPs, putting it into files for local users, having
> an outgoing queue that gets delivered if the remote server is
> responsive, then some firewall-like filtering options, like what mail
> to reject. Where's the immense inherent complexity of the application
> domain? Where would it be in a web server, or in an FTP?
Actually, there is quite a lot a mail server *may* need to do, which
is why it *can* be complicated. Note that there are simpler mail
servers, like exim on one side and very powerful ones like sendmail on
the other. You also overlook many quite complex operations which a
mail server may need to support, such as masquerading based on sender,
checking .forward, aliases, dns and MX records, delivery over various
different interfaces such a uucp, fax, SMS, handling of mail loops and
bounced messages, virtual mail domains, various different mail box
formats (though sendmail and many other SMTP servers don't actually
deliver mail), mail relaying, serving only authenticated users,
address re-writing, spam filtering (RBL etc) etc etc.
I think the real issue is about selecting the right application for
the right job. I don't run sendmail or postfix for my small LAN at
home because I don't need most of the features it offers - instead I
use a simple basic mail server, but I do run sendmail and postfix at
work because in those environments we have to support multiple domains
and over 20,000 users (nearly 1/3 on the local LAN) with a diverse
range of requirements. The same goes for things like web servers -
apache is great with many powerful modules and able to support just
about anything you may need/want to do via the web. However, at home,
I just use a very lightweight httpd server which has only the basics I
need and takes about 10 minutes to configure.
I think the point is that you cannot judge the required complexity of
a piece of software just based on your own need or understanding. None
of these protocols are inherently complicated when you just consider
basic functinality. However, the devil is always in the details.
I do not mean to appear rude, but I get the feeling you have not
written many "real world" applications designed to be used by a large
number of users. I say this as your views seem to have the flavor of
being more based on academic and theoretical reasoning than on
knowledge acquired through experience in writing applications (neither
is better than the other BTW - I just believe both are needed for a
good balance and understanding).
> > The few days and a
> > book was what I needed to write a whole config file, including all the
> > sendmail rule sets, from scratch, which was done as a learning
> > experience not as a necessity.
>
> But for all non-exotic setups I expect to be able to config a server
> for *anything* in an hour. We're not talking setting obscure Oracle
> flags here, to optimize your system ;)
>
and this is quite possible, but you need to select the correct server
software for your needs. Its not reasonable to expect that just
because your requirements are simple that it will be simple to
configure a powerful and flexible application. Even if the app has
reasonable defaults, it will still take effort to check that those
defaults really do suit your setup (or you can just throw caution to
the wind and accept the person who packaged the system chose
reasonable defaults and risk it).
Tim
--
Tim Cross
The e-mail address on this message is FALSE (obviously!). My real e-mail is
to a company in Australia called rapttech and my login is tcross - if you
really need to send mail, you should be able to work it out!
Tim X wrote:
> I think the real issue is about selecting the right application for
> the right job.
I fully agree there.
> I do not mean to appear rude, but I get the feeling you have not
> written many "real world" applications designed to be used by a large
> number of users. I say this as your views seem to have the flavor of
> being more based on academic and theoretical reasoning than on
> knowledge acquired through experience in writing applications (neither
> is better than the other BTW - I just believe both are needed for a
> good balance and understanding).
I haven't administered any big networks, so I never needed the
sendmail functionality, right. For big systems of course a
full-time employee makes more sense, and then a couple days for
learning a complex sendmail config isn't bad. Maybe we were just
talking about different things here...
--
Don't let school interfere with your education. -- Mark Twain
> (Lisp syntax also comes to mind)
And I bet you still find Lisp consultants and support people.
Jon
----
Learn to program using Linux assembly language
http://www.cafeshops.com/bartlettpublish.8640017
> But this kind of business only works as long as these softwares NEED
> consulting (i.e. are crappy).
Have you ever worked in IT?
> If it were all drag-and-drop to install,
> with user-friendly config panels and documentation (like some Linux and
> Mac programs are), then nobody would need to buy consulting.
Totally incorrect. A good, efficient IT department will have the IT
system match the COMPANY, and not the other way around. Linux is great
because it is easy to adapt to new environments and new ways of doing
business. Anytime you adapt something, whether proprietary or free, to
do something new, you have problems. Consulting helps you:
1) make a good plan
2) fix any problems that arise
3) find inefficiencies and fix them
If you can name me any system that does not have need of those
components, let me know! Even TeX, though being bug-free for years
(anyone remember when the last bug was found), still needs such things
for TeX users.
In the Linux model, the people you pay are co-Linuxers who know more
than you do and thus have better experience to answer
questions/problems. In the proprietary model, you have to pay first for
the right to use the software, and then again to have someone come
alongside and help you implement. And, if for any reason you need the
software modified, you probably aren't going to get any help from anyone
unless you are talking in the hundreds of thousands of dollars.
Jon
----
Learn to program using Linux assembly language
http://www.cafeshops.com/bartlettpublish.8640017
Ulrich Hobelmann wrote:
> Brandon J. Van Every wrote:
>
>> I'm also realizing how irrelevant Linux is to shipping games. It
>> isn't "an extra platform where I might pick up a few sales." It is an
>> impossible platform.
>
>
> I could have told you that ;)
>
> Linuxers don't buy software. That's why they use Linux. But Unix in
> general is a great platform for *development*, like most Lisps run well
> on Unix, but not other systems. Just don't expect to write commercial
> stuff FOR linux.
I had *heard* all of that, but I didn't *believe* it, because I'd heard
so much about Linux having come a long ways etc. blah blah blah. I
thought people were just overstating the difficulties. Turns out, they
weren't making the difficulties / impossibilities sufficiently clear.
Well, primary experience sorted that out.
>> So, I'm nukeing my Linux partition and not looking back. Good riddance.
>
>
> If you hate Windows, but want consistency and a viable market of users
> willing to pay money (but not a huge market), get a (cheap) Mac. Saves
> loads of time, in my experience the last 1.5 years (and watching other
> people use XP (and suffer) during that time, poor suckers).
I may do that someday. Right now I punt with MinGW. It's a certain
amount of work to maintain, and as far as I can see, there are no
solutions that are less work.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
How I really feel about Ubuntu Linux:
http://www.redlandsfortnightly.org/images/baker00.jpg
"Brandon J. Van Every" <·····················@mycompanyname.com> writes:
> Ulrich Hobelmann wrote:
> > Brandon J. Van Every wrote:
> >
> >> I'm also realizing how irrelevant Linux is to shipping games. It
> >> isn't "an extra platform where I might pick up a few sales." It is
> >> an impossible platform.
> > I could have told you that ;)
> > Linuxers don't buy software. That's why they use Linux. But Unix
> > in general is a great platform for *development*, like most Lisps
> > run well on Unix, but not other systems. Just don't expect to write
> > commercial stuff FOR linux.
While I am not convinced Linux will ever be a real contender in the
market for desktop solutions, there is a rapidly growing market for
commercial software running on Linux within the server marketplace. In
fact, many organisations are moving away from expensive IBM/Sun/HP
"boom boxes" to smaller (and often clustered) Linux based
systems. Some companies, like Oracle, have moved all their developers
onto Linux as their development platform. Where I work, all but one of
our major systems are now supported on a Linux platform and we are
moving away from Alpha based servers to 64 bit Opterons running Linux
and this is software representing considerable investment (in the
order of $1.5 Million just to purchase it and then you have licensing,
maintenance etc).
It is incorrect to imply that you cannot be successful with commercial
software for a Linux platform. While I would agree that attempting to
be successful in the desktop market would be very difficult (but not
impossible witht he right app), commercial software for a Linux based
server market is certainly possible.
I also don't agree with the statement that Linux users don't pay for
software and thats why they use Linux. Therre are plenty of Linux
users who purchase their distro rather than download it. I've run
commercial paid for software on Linux since the mid 90's. While I
would agree many people try out Linux because it is free and
represents no real investment to try it out, my experience has been
that those who stick with it do so for a lot of reasons other than it
being free - mainly its stability, fewer issues with viruses and to
some extent its freedom to explore.
[Of course, the real reason is that we are all really just Communists
with an evil agenda to over throw this utopia we call Capitalism and
destroy everyones freedoms and ability to earn squillions Whahahahah]
--
Tim Cross
The e-mail address on this message is FALSE (obviously!). My real e-mail is
to a company in Australia called rapttech and my login is tcross - if you
really need to send mail, you should be able to work it out!
Tim X <····@spamto.devnul.com> writes:
> While I am not convinced Linux will ever be a real contender in the
> market for desktop solutions, there is a rapidly growing market for
> commercial software running on Linux within the server marketplace.
Well, I think a lot of the chatter regarding "desktop linux" is based on
a very sloppy, and very bad definition of what linux is. This results
in a heck of a lot of complaints about look and feel issues that have
very little to do with linux per se.
Linux is just a kernel, drivers, and a fairly small set of basic
utilities. That's it.
Now, on top of that you can build desktop environments, servers, thin
clients, cell phones, and POS systems. You can bundle a set of related
software together with a given configuration and call it something like,
Mandriva, Knoppix, Debian, or SuSE.
But a bit too often, I see complaints about Linux that have very little
to do with actually using Linux.
--
Kirk Job-Sluder
"The square-jawed homunculi of Tommy Hilfinger ads make every day an
existential holocaust." --Scary Go Round
Kirk Job Sluder wrote:
>
> But a bit too often, I see complaints about Linux that have very little
> to do with actually using Linux.
One of my complaints is that Linux adherants usually have a different
mentality for parcelling application and system functionality than the
rest of us "mere mortals" who simply expect things to work. So you'll
say things like, "well that's not really a *Linux* problem..." and I'll
say the hell it isn't. There used to be a gazillion Unixes, and it was
certainly valid to talk about "a Unix problem" because of the patchwork,
always-slightly-incompatible way that the Unix world did things.
I'm sure we can agree on these things being "Linux distro" problems.
You just think this doesn't implicate Linux, and I think it does.
Furthermore, all the lay people coming from a Windows world (for
instance, consumers) consider these sorts of things to be "Linux
problems." People talk about "Windows problems" even though there's
W2K, XP, Win98SE, ME, and so forth... doesn't make a problem any less of
a "Windows problem."
You can try to give people a massive re-education / brainwashing, that
somehow if you put the label "Linux" on the right components and
categories, it suddenly doesn't have problems. But I think that's both
pissing in the wind and a lot of malarky. Consumers aren't interested
in hearing it, they just want things to work.
It remains to be seen if any Linux distro will rise above its peers,
provide things that just work, and capture mindshare as "the reliable
Linux." I doubt it's going to happen. I think Apple will kill any
consumer interest in such a beast, and Linux will remain a server system
with lotsa admins in tow.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
I won't spend more than 1 day configuring 1 thing.
"Brandon J. Van Every" <·····················@mycompanyname.com> writes:
>
> It remains to be seen if any Linux distro will rise above its peers,
> provide things that just work, and capture mindshare as "the reliable
> Linux." I doubt it's going to happen. I think Apple will kill any
> consumer interest in such a beast, and Linux will remain a server system
> with lotsa admins in tow.
Show me an OS that "just works" and I'll show you the deed to the
Brooklyn Bridge that I have.
Gregm
Greg Menke <············@toadmail.com> writes:
> "Brandon J. Van Every" <·····················@mycompanyname.com> writes:
>
> >
> > It remains to be seen if any Linux distro will rise above its peers,
> > provide things that just work, and capture mindshare as "the reliable
> > Linux." I doubt it's going to happen. I think Apple will kill any
> > consumer interest in such a beast, and Linux will remain a server system
> > with lotsa admins in tow.
>
> Show me an OS that "just works" and I'll show you the deed to the
> Brooklyn Bridge that I have.
>
Sure, it "Just Works" after someone else does all the hacking to get the
kinks out. Since you're so obsessed about not actually having to learn
what you wish to use, then XP is also excluded because a fresh XP
install can quite easily take hours to get right if you have to track
down drivers and work your way through all the reboots. SHow me the
quick way around needing the network drivers so you can get online to
download the network & modem & USB memory stick drivers. Frankly the
quickest is sometimes to pull the hard disk out of the machine, stick it
in a Linux box, download the drivers and save them to the hard disk,
reinstall boot and go.
Theres not an OS around that Just Works freshly installed, there is
always a greater or lesser degree of hacking to get the OS configured to
match the hardware. OS X and Solaris on SPARC probably comes closest
because of the limited selection of hardware on which they will run.
I've had the occasional Linux install that works end-to-end on first
boot too, but then theres always recompiling the kernel for local
optimizations, adding vga=ext to the console and disabling all sorts of
helpful features.
Gregm
Greg Menke wrote:
> Frankly the
> quickest is sometimes to pull the hard disk out of the machine, stick it
> in a Linux box, download the drivers and save them to the hard disk,
> reinstall boot and go.
That sounds quite desperate. Are you sure the HW you're using isn't
just pure junk?
> Theres not an OS around that Just Works freshly installed, there is
> always a greater or lesser degree of hacking to get the OS configured to
> match the hardware.
Most people buy their HW preconfigured with an OS. Granted, you could
win doing that with Linux. Shifting over to a new OS, yes, people chase
around drivers. The question is whether it's relatively easy to add new
drivers. I find that as long as HW is "sufficiently modern," on Windows
it's pretty easy. But, if the HW is old, then nobody is interested in
supporting it and you have problems.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
Brandon's Law (after Godwin's Law):
"As a Usenet discussion grows longer, the probability of
a person being called a troll approaches one RAPIDLY."
"Brandon J. Van Every" <·····················@mycompanyname.com> writes:
> Greg Menke wrote:
> > Frankly the
> > quickest is sometimes to pull the hard disk out of the machine, stick it
> > in a Linux box, download the drivers and save them to the hard disk,
> > reinstall boot and go.
>
> That sounds quite desperate. Are you sure the HW you're using isn't
> just pure junk?
Yes, I'm sure. Windows doesn't ship with most of the drivers necessary
for laptops for instance. "Desperate" isn't the right word- annoying
beyond belief, but it solves the problem quickly. THis should actually
be appealing to you, because of your 1 hour rule - or is it 1 day....
> > Theres not an OS around that Just Works freshly installed, there is
> > always a greater or lesser degree of hacking to get the OS configured to
> > match the hardware.
>
> Most people buy their HW preconfigured with an OS. Granted, you could
> win doing that with Linux. Shifting over to a new OS, yes, people chase
> around drivers. The question is whether it's relatively easy to add new
> drivers. I find that as long as HW is "sufficiently modern," on Windows
> it's pretty easy. But, if the HW is old, then nobody is interested in
> supporting it and you have problems.
Sure Windows is easiest because you don't actually have to <think> about
the problem. Works great as long as the problem is simple. The funny
thing about most Windows people is when they're faced with difficult
config problems, they just start reinstalling the OS and ordering new
hardware at the same time they proclaim that Windows "Just Works".
I don't think I have every yet seen a fresh Windows install have all the
drivers necessary to come up first try (not counting the 3 or 4
reboots).
I suggest you go get a P.O.S. Walmart elcheapo PC w/ XP preinstalled on
it, stick your chest out and proclaim yourself an IT master.
Gregm
Greg Menke wrote:
> "Brandon J. Van Every" <·····················@mycompanyname.com> writes:
>
>
>>Greg Menke wrote:
>>
>>>Frankly the
>>>quickest is sometimes to pull the hard disk out of the machine, stick it
>>>in a Linux box, download the drivers and save them to the hard disk,
>>>reinstall boot and go.
>>
>>That sounds quite desperate. Are you sure the HW you're using isn't
>>just pure junk?
>
>
> Yes, I'm sure. Windows doesn't ship with most of the drivers necessary
> for laptops for instance. "Desperate" isn't the right word- annoying
> beyond belief, but it solves the problem quickly. THis should actually
> be appealing to you, because of your 1 hour rule - or is it 1 day....
Ok, I have no laptop experience really. I've not encountered these
problems with desktop Windows. I've configured many machines for
friends, and the rule is always "the troublesome driver is the one for
the old cheapass piece of junk." I won't do those configuration jobs
for friends anymore, I tell them to upgrade their HW.
>>Most people buy their HW preconfigured with an OS. Granted, you could
>>win doing that with Linux. Shifting over to a new OS, yes, people chase
>>around drivers. The question is whether it's relatively easy to add new
>>drivers. I find that as long as HW is "sufficiently modern," on Windows
>>it's pretty easy. But, if the HW is old, then nobody is interested in
>>supporting it and you have problems.
>
> Sure Windows is easiest because you don't actually have to <think> about
> the problem. Works great as long as the problem is simple. The funny
> thing about most Windows people is when they're faced with difficult
> config problems, they just start reinstalling the OS and ordering new
> hardware at the same time they proclaim that Windows "Just Works".
Well, they don't know how to use Google and ask questions in
msnews.microsoft.com. I do. Windows driver problems generally lead to
timely solutions *if* the HW isn't old third-string crap. Windows
answers to such problems are generally available in quick cookbook form.
Finding the answers generally takes more time than applying them. In
contrast, Linux answers often have long learning curves.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
I won't spend more than 1 day configuring 1 thing.
"Brandon J. Van Every" <·····················@mycompanyname.com> writes:
> Greg Menke wrote:
> > "Brandon J. Van Every" <·····················@mycompanyname.com> writes:
> >
>
> >>Most people buy their HW preconfigured with an OS. Granted, you could
> >>win doing that with Linux. Shifting over to a new OS, yes, people chase
> >>around drivers. The question is whether it's relatively easy to add new
> >>drivers. I find that as long as HW is "sufficiently modern," on Windows
> >>it's pretty easy. But, if the HW is old, then nobody is interested in
> >>supporting it and you have problems.
> > Sure Windows is easiest because you don't actually have to <think>
> > about
> > the problem. Works great as long as the problem is simple. The funny
> > thing about most Windows people is when they're faced with difficult
> > config problems, they just start reinstalling the OS and ordering new
> > hardware at the same time they proclaim that Windows "Just Works".
>
> Well, they don't know how to use Google and ask questions in
> msnews.microsoft.com. I do. Windows driver problems generally lead to
> timely solutions *if* the HW isn't old third-string crap. Windows
> answers to such problems are generally available in quick cookbook
> form. Finding the answers generally takes more time than applying them.
> In contrast, Linux answers often have long learning curves.
>
>
Thats surely going to be a lot of help when some app blows up enough of
the registry to make the OS bootable but more or less unusable.
Google's no good there. I find it amazing that you think you know
enough about Linux to make judgements about it. "Cookbook answers"
indeed. LOL!
Gregm
Ingvar <······@hexapodia.net> writes:
> Greg Menke <············@toadmail.com> writes:
>
> [ SNIP ]
> > > Most people buy their HW preconfigured with an OS. Granted, you could
> > > win doing that with Linux. Shifting over to a new OS, yes, people chase
> > > around drivers. The question is whether it's relatively easy to add new
> > > drivers. I find that as long as HW is "sufficiently modern," on Windows
> > > it's pretty easy. But, if the HW is old, then nobody is interested in
> > > supporting it and you have problems.
> >
> > Sure Windows is easiest because you don't actually have to <think> about
> > the problem. Works great as long as the problem is simple. The funny
> > thing about most Windows people is when they're faced with difficult
> > config problems, they just start reinstalling the OS and ordering new
> > hardware at the same time they proclaim that Windows "Just Works".
> >
> > I don't think I have every yet seen a fresh Windows install have all the
> > drivers necessary to come up first try (not counting the 3 or 4
> > reboots).
>
> I think I didn't need to do any reboots (other than what might've been
> required by the install process) last time I instaleld Xp on my
> gamesbox (I was a bit happy when I managed to convince it to get back
> to a bootable state Monday, asves me tracking down some stuff on
> there, for playing with lisp development in a Windows environment).
>
Just yesterday I set up a Toshiba laptop w/ XP preinstall, had to
intersperse 3 or so reboots between various mysterious downloads before
the antivirus/anitspyware software would quit nagging. And then
sometimes XP wants to reboot when the IP config changes and sometimes it
doesn't. And then while uninstalling the shovelware, I had to reboot a
few more times.
Gregm
From: Christopher C. Stacy
Subject: Re: I've thought better of Linux
Date:
Message-ID: <uoea62f81.fsf@news.dtpq.com>
Greg Menke <············@toadmail.com> writes:
> Just yesterday I set up a Toshiba laptop w/ XP preinstall, had to
> intersperse 3 or so reboots between various mysterious downloads
> before the antivirus/anitspyware software would quit nagging.
Is your claim that, if Linux ever attracts lots of users,
that installing that same software will be less mysterious?
My experience on Windows is that when more updates (for either
the OS, or for the well-integrated Norton security) are needed,
icons appear in the tray, I click on them, and then sometimes
afterwards it tells me to reboot. Totally painless.
My experience on Debian is that the security icon appears,
I click on it, and then it brings up Synaptic, and then
it doesn't work. I have no idea what "critical security
updates" are needed, why I am not getting them, and there
doesn't seem to be any straightforward way to figure it out.
Then I delve into the mysteries of command-line apt and
configuration files, and _maybe_ I can figure it out.
If I were Joe Luser, or his mother, on Linux, there
is absolutely no chance that I would figure it out.
Windows is infinitely superior to Linux in that way.
And do please bear in mind that I rather hate Windows.
> And then sometimes XP wants to reboot when the IP config changes and
> sometimes it doesn't. And then while uninstalling the shovelware,
> I had to reboot a few more times.
When you are installing OS patches, is rebooting a big deal?
I have to do this about every third week or so.
My Windows machine does not require using key switches to input
the octal bootloader program I've memorized, nor do I have to read
in a paper tape after that, nor type commands on the TTY to finally
load the OS from disk and start it. On mine, I just press "OK",
and then it comes back up. What kind of machine do you have?
From: Robert Uhl
Subject: Re: I've thought better of Linux
Date:
Message-ID: <m31x72j71z.fsf@4dv.net>
······@news.dtpq.com (Christopher C. Stacy) writes:
>
> My experience on Windows is that when more updates (for either the OS,
> or for the well-integrated Norton security) are needed, icons appear
> in the tray, I click on them, and then sometimes afterwards it tells
> me to reboot. Totally painless.
[snip]
> If I were Joe Luser, or his mother, on Linux, there is absolutely no
> chance that I would figure it out. Windows is infinitely superior to
> Linux in that way.
Ummm...security updates are that easy with Red Hat. And on Fedora it's
a simple 'yum update'. Which reminds me...
And of course a reboot is never required.
> When you are installing OS patches, is rebooting a big deal?
Yeah, because on a single-user machine one may have quite a lot of
state: open web-pages, half-completed projects; chat sessions. Yes,
with better OSes this wouldn't matter: all state would be constantly
saved, and a reboot (or power outage, for that matter) would just reload
the kernel and restore the state. But we don't have those better OSes;
we're stuck with Windows & Unix; and thus rebooting is a hassle.
And of course in a production environment rebooting means interrupting
production. Yeah, it'd be nice to have clusters and thus not need to
worry about it. But, again, we don't live in a perfect world.
--
Robert Uhl <http://public.xdi.org/=ruhl>
An environment that puts women unwillingly at the disposal of men is
never an argument for equality. It is quite vividly the opposite.
--Kathleen Parker
From: Christopher C. Stacy
Subject: Re: I've thought better of Linux
Date:
Message-ID: <u64wd3odl.fsf@news.dtpq.com>
Robert Uhl <·········@NOSPAMgmail.com> writes:
> ······@news.dtpq.com (Christopher C. Stacy) writes:
> >
> > My experience on Windows is that when more updates (for either the OS,
> > or for the well-integrated Norton security) are needed, icons appear
> > in the tray, I click on them, and then sometimes afterwards it tells
> > me to reboot. Totally painless.
>
> [snip]
>
> > If I were Joe Luser, or his mother, on Linux, there is absolutely no
> > chance that I would figure it out. Windows is infinitely superior to
> > Linux in that way.
>
> Ummm...security updates are that easy with Red Hat. And on Fedora
> it's a simple 'yum update'. Which reminds me...
I used to run RedHat 9 for a server.
It worked well for me.
That was supposed to be the stable system,
but given it's poor reputation, when they
dropped it and announced that either very
expensive "Enterprise" charges would ensue,
or else that "bleeding edge" Fedora was the
replacement, I decided to try something else.
Debian Woody was the answer.
Then I wanted a desktop, so I got Sarge.
I like apt and Synaptic.
> And of course a reboot is never required.
Well, that's just not true.
(You admitted as much in the subsequent paragraph).
The reason that you "have to" reboot Windows after
applying packages is so that the running applcations
will pick up the new shared libraries. I think you
reboot so that it's all coherent -- the registered COM
components and DLLs will all be in syncg in the next
"CurrentConfiguration" part of the registry.
In Linux, you usually have the option of just restarting
the application. In either case, there's an outage,
and you pick when that will occur in your production schedule.
On either operating system, you can usually just continue to
run the existing applications until it's convenient to boot.
That's why most patches on Windows don't make you reboot,
some of them offer, "No, I will reboot later" or "Click Cancel
to Reboot later", and only the grossest changes to the main
part of the OS try to insist that you reboot immediately.
······@news.dtpq.com (Christopher C. Stacy) writes:
> Greg Menke <············@toadmail.com> writes:
> > Just yesterday I set up a Toshiba laptop w/ XP preinstall, had to
> > intersperse 3 or so reboots between various mysterious downloads
> > before the antivirus/anitspyware software would quit nagging.
>
> Is your claim that, if Linux ever attracts lots of users,
> that installing that same software will be less mysterious?
It will be a lot less necessary. And btw, Linux has LOTS of users- they
just don't show up so often on marketing surveys and sales figures are a
poor way to estimate their numbers. Don't get me wrong, I think the
Linux-on-desktop as Windows-killer is a ways off yet, but a heck of a
lot closer than it was a couple years ago.
>
> If I were Joe Luser, or his mother, on Linux, there
> is absolutely no chance that I would figure it out.
> Windows is infinitely superior to Linux in that way.
> And do please bear in mind that I rather hate Windows.
After seeing secretaries with nothing much more than Word Perfect and
Excel experience learn SQL and roll their own queries against complex
financial systems, I don't talk about "Joe Lusers" very much. Some
people choose not to learn how to operate and maintain their computers
and some don't- and its not so easy to predict which way a given user
will end up.
I personally know some people with Math and Compsci Masters degrees who
can't write C programs much less set up a network. And a few years ago
I worked with a guy who had some vague associates degree from a
community college who was lead programmer of a 8 or 10 person team
developing a big financial system for a federal gov't agency.
So I'm sure not going to attempt to predict what a given person can
choose to learn or not.
>
> > And then sometimes XP wants to reboot when the IP config changes and
> > sometimes it doesn't. And then while uninstalling the shovelware,
> > I had to reboot a few more times.
>
> When you are installing OS patches, is rebooting a big deal?
> I have to do this about every third week or so.
I do my Solaris patches a couple times a year when my users are offline,
I just init 1 the system, run smpatch to automatically download &
install the patches, followed by a single (and optional) reboot just to
cover the case of kernel patches. On Linux I simply recompile new
versions of software as required. Don't need anti-spyware, anti-virus
or anti-anything updates because I don't use a broken-ass, ill-designed
toy operating system which is an enterprise-wide IT security risk from
email & web-borne worms/virii or simply from poorly designed network
services. On bad Windows-virus days its sort of fun to watch people
scurrying around while I just keep on working.
I'll add one more thing, the biggest pain in my ass IT-wise is all the
security measures put in place in an attempt to keep Windows problems
from wreaking havoc on enterprise systems, applied across the board to
all systems vulnerable or not. Now I have to rename .zip file
attachments to emails because of ·@*&ing stupid Windows email clients
will try and execute whatever crud they receive- so the standard
procedure is to change the name of the extension to .fred or some such
crap.
> My Windows machine does not require using key switches to input
> the octal bootloader program I've memorized, nor do I have to read
> in a paper tape after that, nor type commands on the TTY to finally
> load the OS from disk and start it. On mine, I just press "OK",
> and then it comes back up. What kind of machine do you have?
I don't have machines like that either, they boot up to a nice tty so I
can log in and run startx or not as required. I just don't like having
to reboot an OS just because some software is added/removed or when some
subtle aspect of IP configuration changes. Or when the OS just starts
going wonky and theres nothing else to try. Its funny how often Windows
people reboot their systems...
Gregm
From: Christopher C. Stacy
Subject: Re: I've thought better of Linux
Date:
Message-ID: <u1x713ni2.fsf@news.dtpq.com>
Greg Menke <············@toadmail.com> writes:
> ······@news.dtpq.com (Christopher C. Stacy) writes:
>
> > Greg Menke <············@toadmail.com> writes:
> > > Just yesterday I set up a Toshiba laptop w/ XP preinstall, had to
> > > intersperse 3 or so reboots between various mysterious downloads
> > > before the antivirus/anitspyware software would quit nagging.
> >
> > Is your claim that, if Linux ever attracts lots of users,
> > that installing that same software will be less mysterious?
>
> It will be a lot less necessary. And btw, Linux has LOTS of users- they
> just don't show up so often on marketing surveys and sales figures
We're talking about desktop users.
Linux is not yet a significant player.
I am sure that Linux use has increased greatly recently.
However, if you think that Linux is comparable to Windows
in its adoption in any desktop market, I think you're living
in fantasy land. I would be interested in some proof of your
assertion. (So far, your proof consists of your claiming that
all the available information is wrong.)
> > If I were Joe Luser, or his mother, on Linux, there
> > is absolutely no chance that I would figure it out.
> > Windows is infinitely superior to Linux in that way.
> > And do please bear in mind that I rather hate Windows.
>
> After seeing secretaries with nothing much more than Word Perfect and
> Excel experience learn SQL and roll their own queries against complex
> financial systems, I don't talk about "Joe Lusers" very much.
> Some people choose not to learn how to operate and maintain their
> computers and some don't- and its not so easy to predict which way a
> given user will end up.
Oh, you're talking about the "corporate desktop" world,
not the personal/home computer market. Those secretaries
have a professional system administrator for their machines.
> security measures put in place in an attempt to keep Windows
> problems from wreaking havoc on enterprise systems, applied across
> the board to all systems vulnerable or not. Now I have to rename
> .zip file attachments to emails because of ·@*&ing stupid Windows
> email clients will try and execute whatever crud they receive- so
> the standard procedure is to change the name of the extension
> to .fred or some such crap.
Someday when lots of people are using Linux, why won't they
have stupid email clients that attempt to open attachments?
I think that you're actually complaining about particular
Microsoft applications. I use Windows for my email, and
I never, ever have the slightest problem with viruses.
But of course I don't use Microsoft programs for email.
The Mozilla applications that I use lately are subject to
many of the same vulnerabilities, on either operating system.
> I'm sure not going to attempt to predict what a given
> person can choose to learn or not.
I will continue to make assertions about what I think
the "average home user" can deal with. And I'm right!
The quality of desktop Linux is currently poorer than Microsoft.
Even if a home user purchases Linux pre-installed on the machine,
so that they don't have to do the initial administration,
it's still not going to work as well as Microsoft Windows.
It is possible that Linux could get to be good.
One can hope.
The core technology is in many ways better than Windows.
I do not know if this will true of Longhorn, however.
> Its funny how often Windows people reboot their systems...
Speaking of either Windows or Linux,
see "it's a good thing it boots fast!"
in the Unix Haters Handbook.
Except of course, that it's no longer true.
(That is, neither one boots terribly fast.)
······@news.dtpq.com (Christopher C. Stacy) writes:
> Greg Menke <············@toadmail.com> writes:
>
> > ······@news.dtpq.com (Christopher C. Stacy) writes:
> >
> > It will be a lot less necessary. And btw, Linux has LOTS of users- they
> > just don't show up so often on marketing surveys and sales figures
>
> We're talking about desktop users.
> Linux is not yet a significant player.
>
> I am sure that Linux use has increased greatly recently.
> However, if you think that Linux is comparable to Windows
> in its adoption in any desktop market, I think you're living
> in fantasy land. I would be interested in some proof of your
> assertion. (So far, your proof consists of your claiming that
> all the available information is wrong.)
Linux and Solaris are ubquitous in our organization. Lots of user
workstations run Windows but <all> the build machines, groundstations
and much of the infrastructure are Linux or Solaris boxes. In our lab
alone, we have one purchased copy of Redhat Enterprise 3 (and 2 more
site licensed installs) and one downloaded copy of Fedora Core 3 (which
was an from Redhat 9 which was an upgrade of Redhat 7.2), all of which
yield about 15 Linux machines. As far as external accounting for
installed base is concerned, the best estimate would be 1 Redhat
Enterprise purchase, 1 Fedora Core 3 download, 1 Redhat 9 download and
no Redhat 7.2 download (I got copies of somebody else's cd's).
Of course I'm not proposing Linux has made a large number of percentage
point gains on the desktop- but it has penetrated very deeply elsewhere.
> > > If I were Joe Luser, or his mother, on Linux, there
> > > is absolutely no chance that I would figure it out.
> > > Windows is infinitely superior to Linux in that way.
> > > And do please bear in mind that I rather hate Windows.
> >
> > After seeing secretaries with nothing much more than Word Perfect and
> > Excel experience learn SQL and roll their own queries against complex
> > financial systems, I don't talk about "Joe Lusers" very much.
> > Some people choose not to learn how to operate and maintain their
> > computers and some don't- and its not so easy to predict which way a
> > given user will end up.
>
> Oh, you're talking about the "corporate desktop" world,
> not the personal/home computer market. Those secretaries
> have a professional system administrator for their machines.
As far as turning it on and booting it to a usable state, sure- however
they chose to learn how to do SQL from the ground up, which means
learning relational database concepts in the abstract and as implemented
in their data tables. Do you propose to tell me they could not just as
easily choose to learn how to put together a PC? Heck my parents put
together an Access database and front end for their business which was
the primary "MIS" tool for 7 years- they had no IT experience
whatsoever, it was good enough for contact management, sales, billing
and financials reporting- without resorting to some fo the more horrible
kludges so often seen in accounting apps. All I did for them was make
sure the backups ran (and fix the Windows machines whenever they went
nuts- and nuke the NT server, replacing it with Linux when NT
self-destructed one day).
> Someday when lots of people are using Linux, why won't they
> have stupid email clients that attempt to open attachments?
Because theres no broken default applications being inflicted on
end-users year after year, version after version. The busted stuff is
quickly culled and upgraded.
> I think that you're actually complaining about particular
> Microsoft applications. I use Windows for my email, and
> I never, ever have the slightest problem with viruses.
> But of course I don't use Microsoft programs for email.
> The Mozilla applications that I use lately are subject to
> many of the same vulnerabilities, on either operating system.
<Any> Windows app that passes a file off to exec() or system() or
whatever for processing based on the extension is subject to this bug-
and its even worse because most users run as Administrator so the damage
from virii is essentially unbounded at least locally. And guess how
many Windows users just use Explorer and Exchange because its there. In
our org we have pretty much 2 different Windows email clients- Exchange
or webmail. And when you say "mozilla" to the Windows people, the drool
leaking out of their mouths gets your feet wet.
Now if you can get Microsoft to actually fix the design flaws in their
default email and web clients, then I'll stop bitching about the
problem.
> The quality of desktop Linux is currently poorer than Microsoft.
> Even if a home user purchases Linux pre-installed on the machine,
> so that they don't have to do the initial administration,
> it's still not going to work as well as Microsoft Windows.
If you say so.
> The core technology is in many ways better than Windows.
> I do not know if this will true of Longhorn, however.
What Longhorn? Oh you mean Windows 2007 or whatever year it is they
finally release something that looks remarkably like XP with not much
new but different graphics and a yet-again mixed around Control Panel?
> > Its funny how often Windows people reboot their systems...
>
> Speaking of either Windows or Linux,
> see "it's a good thing it boots fast!"
> in the Unix Haters Handbook.
> Except of course, that it's no longer true.
> (That is, neither one boots terribly fast.)
True- Solaris 10 is excellent in this regard. Myself, I want a Unix
Barf Bag.
Gregm
······@news.dtpq.com (Christopher C. Stacy) writes:
> Greg Menke <············@toadmail.com> writes:
>
>
> > security measures put in place in an attempt to keep Windows
> > problems from wreaking havoc on enterprise systems, applied across
> > the board to all systems vulnerable or not. Now I have to rename
> > .zip file attachments to emails because of ·@*&ing stupid Windows
> > email clients will try and execute whatever crud they receive- so
> > the standard procedure is to change the name of the extension
> > to .fred or some such crap.
>
> Someday when lots of people are using Linux, why won't they
> have stupid email clients that attempt to open attachments?
> I think that you're actually complaining about particular
> Microsoft applications. I use Windows for my email, and
> I never, ever have the slightest problem with viruses.
> But of course I don't use Microsoft programs for email.
> The Mozilla applications that I use lately are subject to
> many of the same vulnerabilities, on either operating system.
>
Well, there is a significant difference your over looking here. The
main reason Windows has such a problem with security is due to
differences in the design compared to Linux. Linux (and most other
Unix like OS) have a much more defined seperation between user space
and kernel space. This seperation provides a lot more security. Its
true that even now there are mail clients for Linux which can be
configured to automatically do things based on the mail they recieve
and could just as easily expose the user to similar dangers as clients
running under Windows. However, the impact of such malicious exploits
is much more limited due to the cleaner seperation between user space
and kernel space. Microsoft has improved a lot in this respect and
more recent releases of the OS have more defined seperation of kernel
and user space. However, the remaining problem is with the GUI itself
- under windows, this is still very heavily linked into the kernel
space. Under Linux, this isn't the case - though there are some
distributions which have blurred this distinction in order to provide
more "user friendly" interface options (such as being able to
manipulate network interfaces, mount media etc as a normal user.
Security is always a play off between user friendliness/functionality
and safety. MS have emphasized the user friendly aspect at the cost of
security. Many windows users are often frustrated when they first
start to use a Linux or BSD OS because there are certain things they
cannot do unless they login as root - which can become even more
complex if the program they want to run as root requires access to X
windows. However, all of this is manageable and configurable, so
ultimately, under Linux/BSD you can choose which you want. Under
windows, your usually stuck with what MS has decided is the best for you.
Tim
--
Tim Cross
The e-mail address on this message is FALSE (obviously!). My real e-mail is
to a company in Australia called rapttech and my login is tcross - if you
really need to send mail, you should be able to work it out!
From: Christopher C. Stacy
Subject: Re: I've thought better of Linux
Date:
Message-ID: <u4qbx72bi.fsf@news.dtpq.com>
Tim X <····@spamto.devnul.com> writes:
> ······@news.dtpq.com (Christopher C. Stacy) writes:
>
> > Greg Menke <············@toadmail.com> writes:
> >
> >
> > > security measures put in place in an attempt to keep Windows
> > > problems from wreaking havoc on enterprise systems, applied across
> > > the board to all systems vulnerable or not. Now I have to rename
> > > .zip file attachments to emails because of ·@*&ing stupid Windows
> > > email clients will try and execute whatever crud they receive- so
> > > the standard procedure is to change the name of the extension
> > > to .fred or some such crap.
> >
> > Someday when lots of people are using Linux, why won't they
> > have stupid email clients that attempt to open attachments?
> > I think that you're actually complaining about particular
> > Microsoft applications. I use Windows for my email, and
> > I never, ever have the slightest problem with viruses.
> > But of course I don't use Microsoft programs for email.
> > The Mozilla applications that I use lately are subject to
> > many of the same vulnerabilities, on either operating system.
> >
> Well, there is a significant difference your over looking here. The
> main reason Windows has such a problem with security is due to
> differences in the design compared to Linux. Linux (and most other
> Unix like OS) have a much more defined seperation between user space
> and kernel space.
It doesn't need to invade kernel space and become root
in order to do arbitrary things to me, or delete all
my files, or install programs in PATH, etc.
Besides, I am completely unable to convince otherwise
seemingly intelligent people that they should NOT
always log into their Linux machines as root...
> Security is always a play off between user friendliness/functionality
> and safety. MS have emphasized the user friendly aspect at the cost of
> security. Many windows users are often frustrated when they first
> start to use a Linux or BSD OS because there are certain things they
> cannot do unless they login as root - which can become even more
> complex if the program they want to run as root requires access to X
> windows.
Funny you should mention that...
> However, all of this is manageable and configurable, so ultimately,
> under Linux/BSD you can choose which you want. Under windows, your
> usually stuck with what MS has decided is the best for you.
What, the users' brains are configurable?
You must be running a different distro than me!
······@news.dtpq.com (Christopher C. Stacy) writes:
>
> It doesn't need to invade kernel space and become root
> in order to do arbitrary things to me, or delete all
> my files, or install programs in PATH, etc.
I wasn't trying to say that it couldn't do damage - in fact, of course
it can do damage to a users own files. However, it cannot
damage/change system files which the user does not own. It cannot
change programs in the PATH except when the path includes the users
own directories. Note also that I acknowledge that Windows has
improved in this aspect and in fact has adopted a model closer to the
concepts of Unix like seperation of user and kernel space, but with a
major limitation in that there is no clear seperation with respect to
core components like the GUI, which means potential areas of exploit
can still exist.
> Besides, I am completely unable to convince otherwise
> seemingly intelligent people that they should NOT
> always log into their Linux machines as root...
Yes, this is always a problem. I see the same thing emerging with
Windows and people configuring their normal user account to also have
administrator privs. I'm also constantly amazed when I see
instructions from vendors like Oracle who suggest that if your having
problems geting their installer to run under Unix/Linux because your
runing it as root, but logged into X as a normal user, to use xhost to
get around the "Cannot connect to X" error, yet xhost has been a known
security problem for years. I guess they are just too lazy to try and
explain how to merge xauth cookies.
>
> > Security is always a play off between user friendliness/functionality
> > and safety. MS have emphasized the user friendly aspect at the cost of
> > security. Many windows users are often frustrated when they first
> > start to use a Linux or BSD OS because there are certain things they
> > cannot do unless they login as root - which can become even more
> > complex if the program they want to run as root requires access to X
> > windows.
>
> Funny you should mention that...
really? why?
> > However, all of this is manageable and configurable, so ultimately,
> > under Linux/BSD you can choose which you want. Under windows, your
> > usually stuck with what MS has decided is the best for you.
>
> What, the users' brains are configurable?
> You must be running a different distro than me!
Huh?
--
Tim Cross
The e-mail address on this message is FALSE (obviously!). My real e-mail is
to a company in Australia called rapttech and my login is tcross - if you
really need to send mail, you should be able to work it out!
From: Christopher C. Stacy
Subject: Re: I've thought better of Linux
Date:
Message-ID: <uslzi2g5b.fsf@news.dtpq.com>
Greg Menke <············@toadmail.com> writes:
> a fresh XP install can quite easily take hours to get right
> if you have to track down drivers and work your way through
> all the reboots.
I believe the claim is not absolute, but merely that XP
is _much_ better than Linux in this regard. Based on my
own experience, which includes setting up two XP laptops
(HP Pavillions) this week, I believe it. I did have to
reboot about 4 times and keep clicking on Windows Update.
But that's all I had to do. (The rest was installing
applications such as Firefox, Nero, etc, which was only
a matter of popping CDs in the machine. The hardest
thing to set up was XEmacs, which wasn't hard at all.)
Greg Menke wrote:
> Show me an OS that "just works" and I'll show you the deed to the
> Brooklyn Bridge that I have.
Preinstalled Win XP just works, except for the usual weird Windows
problems that pop up once in a while. Mac OS X on a Mac just
works. Most "normal" hardware does, too. Don't expect to run
exotic, proprietary stuff with it (or ooold stuff).
--
Don't let school interfere with your education. -- Mark Twain
Ulrich Hobelmann wrote:
> Greg Menke wrote:
> > Show me an OS that "just works" and I'll show you the deed to the
> > Brooklyn Bridge that I have.
>
> Preinstalled Win XP just works, except for the usual weird Windows
> problems that pop up once in a while. Mac OS X on a Mac just
> works. Most "normal" hardware does, too. Don't expect to run
> exotic, proprietary stuff with it (or ooold stuff).
>
> --
> Don't let school interfere with your education. -- Mark Twain
I installed Debian Sarge on my machine and it "just work" I can play
movie, DVD, play game, listen to OGG, use CMUCL, Emacs, Evolution,
GnuCash, OpenOffice. At 0$ dollar. Only problem I have is with using my
Thai language on some application. But that's normal, if the program
wasn't supporting unicode then it won't work even though it's Windows
program.
MacOS only run on one architecture, a Mac. MacOS, come preinstalled on
its vendor's hardware know what hardward they are using, it will be so
stupid if they are not the one getting it work right the first time.
Windows or Linux doesn't get to choose what hardware it was installed
in. So yes MacOS could win the "Just Work" test.
It's like "A car comes in any color you want as long as it's black"
"MacOS just work pre-installed as long as it's installed on Mac machine
it was designed for"
joesb <··········@gmail.com> wrote:
>"MacOS just work pre-installed as long as it's installed on Mac machine
>it was designed for"
So, is that a bad thing? I fondly remember when departments were
buying workstations (DEC, Sun, HP, IBM) where the (well engineered)
hardware and OS came from one vendor. Those things "just worked"
aswell. Today, you have el-cheapo PCs with marginal components and
Linux and all kinds of shithouse problems. Not the technological
advantage as I would like it. But then, of course it's a lot cheaper,
both in price, as in quality but especially in price, and that's
what seems to count.
mkb.
"joesb" <··········@gmail.com> writes:
> Windows or Linux doesn't get to choose what hardware it was installed
> in. So yes MacOS could win the "Just Work" test.
Well, not even then. When my sweetie switched to a mac laptop, one of
the annoyances was that her HP Scanner didn't work and there was no plan
to release drivers for OS X. Whoops, another trip to the Office supply
store.
--
Kirk Job-Sluder
"The square-jawed homunculi of Tommy Hilfinger ads make every day an
existential holocaust." --Scary Go Round
From: Christopher C. Stacy
Subject: Re: I've thought better of Linux
Date:
Message-ID: <uwtou2gez.fsf@news.dtpq.com>
"joesb" <··········@gmail.com> writes:
> Ulrich Hobelmann wrote:
> > Greg Menke wrote:
> > > Show me an OS that "just works" and I'll show you the deed to the
> > > Brooklyn Bridge that I have.
> >
> > Preinstalled Win XP just works, except for the usual weird Windows
> > problems that pop up once in a while. Mac OS X on a Mac just
> > works. Most "normal" hardware does, too. Don't expect to run
> > exotic, proprietary stuff with it (or ooold stuff).
I have been running Debian Sarge on my totally generic
machine for the last year. Here's a short list of
complaints off the top of my head.
1. It just doesn't work with my generic (HP) scroll mouse.
I can't figure out why. (Actually, it starts out working,
for a few minutes, then it totally stops working!)
2. Cut and paste do not work in the applications that I use
(eg. paste a URL or text between Emacs and Firefox).
3. Sound doesn't work right.
Audio applications mess up and even crash all over the place.
4. I had to manually add and configure the totally
generic network device driver.
5. I had to manually configure X for the totally generic
video card. And I fucked something up: when Gnome launches,
there's an extra mouse cursor ("X") sitting in the middle
of the screen. It's a turd. I can make it go away
by launching a shell and clicking on its titlebar.
Then I close that shell, because I do everything from Emacs.
(Clicking on the Emacs titlebar doesn't clear the turd.)
5. I tried using KDE, but when I tried to customize the
Start menu, the system got all confused, lost apps, etc.
6. I decided Gnome was the lesser of two evils, and switched.
The start menu subsystem here is even worse, but at least
I'm not tempted to try and reconfigure it (no such feature).
Gnome runs the weather (temp/sky/radar) widget that I like,
which doesn't run under KDE. Other than that, I think I
like KDE better. But KDE has too many other tempting
things to click on, and after a week I decided they have
too many bugs, so I stay away (although I do have KDE
available from the login screen).
I know there are even more desktop/window environments.
Why can't Linux just pick ONE and make that really WORK?
At least as well as Windows?
I suppose I should be thankful that the entire destkop
doesn't crash all the time, like it did a few years ago.
7. The "security update notification" widget doesn't work.
It beeps and complains that security upgrades are ready.
Then it launches Synaptic, which doesn't agree.
I am left wondering what critical security patches I am missing.
By running apt manually, I suspect the problem is
related either to qpopper or wine. I think these problems
first manifested the week that Sarge became Stable.
I suppose I ought to look into them.
8. I spent two hours yesterday trying to install a working imapd.
Tried several different packages. No success so far.
9. Maybe it's an xinetd problem.
Nothing like having to manually configure a seperate inetd.
It's supposed to be a replacement, but it's not actually
integrated into the system! Applications install as inetd,
then I run some perl script over inetd.conf and manually
edit the xinetd. Someday I think some package will need
to configure and start a network service before completing
installation. I just can't wait to watch that nightmare.
10. Tried to get my friend to run Debian, since he wanted
to use Linux for some program development.
But Sarge doesn't work with his generic video card on
the generic $400 AMD64 system he bought the other day
at the computer store. If he wants to run Sarge, he will
have to hire a consultant to manually configure X Windows.
I haven't been able to help him over the phone, because
he understandably lost patience trying to debug this.
I am sure there would have been many more problems,
but not having a display screen was the first show stopper.
RedHat (Fedora) and Knoppix do come up on his machine.
So he's running Knoppix, but with all the problems,
he's abandoned Linux for now, and the machine just
sits there. He uses Windows XP, instead.
He's been a system programmer for 25 years, too,
but he has practically no experience with Unix.
At work (doing real-time embedded systems),
he uses RedHat, which someone else set up and maintains.
Windows 2000 and XP "just works" perfectly out of the box,
in all those respects. If you're not sure: that means you
merely apply AC power to the system, and you're done.
Let us not even speak about trying to do "advanced" things
like set up local network file sharing on Linux.
Yes, of course, I have samba working on my network.
But on Windows, you just do this by clicking on menu items
named "Share" and "Permissions"; nothing else needed.
I'm a system programmer and have administered probably
every kind of Unix-based system over the last 25 years
for both tiny and huge organizations of various kinds.
If my perception is that Linux doesn't work "out of the box",
it's a safe bet that newbies who just want to turn on their
desktop and get to work will be in for a very rough ride.
I wouldn't try to kid them.
You're just going to wind up souring them on Linux.
I had to reinstall my Windows 2000 desktop last week.
That horrible nightmare happens about twice a year.
But it's far less time than I spend trying to make Linux work.
But I also have extensive experience with Windows 2000
in a server environment, and I would NOT recommend that!
Some flavor of Unix is going to be much better.
Lately, I use BSD and Debian for that, and Debian has
been much better in that role than BSD, to my surprise.
I once tried to install Gentoo, and that was a bad joke.
Maybe Gentoo will get better, though.
In the recent past I've had very good luck with Solaris.
For a desktop, I would always recommend Macintosh.
(I wish I was using MacOS X, myself...)
> I know there are even more desktop/window environments.
> Why can't Linux just pick ONE and make that really WORK?
> At least as well as Windows?
Some distros try to do this. Ubuntu is essentially Debian with Gnome -
but things have a tendency to work much better out of the box. You
mentioned Knoppix, which is Debian based but KDE-centric, and once
again, there's a lot less hassle to get things to work. Debian
attempts to be all things to all people on all architectures, and, not
surprisingly, doesn't do certain things as well as more specialized
distributions.
And there are good reasons to have multiple window managers. I
administer three machines with three different WMs - I prefer fluxbox
and use it on mine, but I would never inflict it on my wife, so I
installed KDE on the machine she uses. I installed IceWM for my
mother, because she needs something easy to use, but her computer is
far too slow to run the pigs known as KDE and Gnome. All of them have
there place.
From: Robert Uhl
Subject: Re: I've thought better of Linux
Date:
Message-ID: <m364wej7e6.fsf@4dv.net>
······@news.dtpq.com (Christopher C. Stacy) writes:
>
> I have been running Debian Sarge on my totally generic machine for the
> last year. Here's a short list of complaints off the top of my head.
I'm running Fedora 3...
> 1. It just doesn't work with my generic (HP) scroll mouse.
The Dell scroll mouse here at work does just fine, although personally I
prefer a middle mouse button.
> 2. Cut and paste do not work in the applications that I use
> (eg. paste a URL or text between Emacs and Firefox).
Works for me. Only cut & paste problem I have is Unicode & emacs.
> 3. Sound doesn't work right.
> Audio applications mess up and even crash all over the place.
I've had rhythmbox crash multiple times in an older version, but sound
itself is perfect--I never have any issues with xmms, mpg321, mpg123,
ogg123, xine or mplayer.
> 4. I had to manually add and configure the totally
> generic network device driver.
Well, that's Debian for you. Fedora found my cards just fine.
> 5. I had to manually configure X for the totally generic
> video card.
That's Debian alright.
> 5. I tried using KDE, but when I tried to customize the
> Start menu, the system got all confused, lost apps, etc.
That's KDE.
> 6. I decided Gnome was the lesser of two evils, and switched.
> The start menu subsystem here is even worse, but at least
> I'm not tempted to try and reconfigure it (no such feature).
There is (or was) a GNOME menu editor, but I've not used it in ages.
> I know there are even more desktop/window environments.
> Why can't Linux just pick ONE and make that really WORK?
> At least as well as Windows?
Because there is no 'Linux' to choose. Some folks like to work on KDE
and write apps for KDE; other folks like GNOME and like to write apps
for GNOME. Some folks like Windows and write apps for Windows.
What would you propose--that every nation in the world outlaw working on
any desktop environment other than FooDesk?
> I suppose I should be thankful that the entire destkop
> doesn't crash all the time, like it did a few years ago.
Again, I never had that, despite using GNOME for many years (and fvwm2
before that). Used to have issues with Enlightenment, though that was
6-7 years ago.
Hey--use what you like. But it seems to me that your problems are
related to using Debian, which is not at all known as a user-friendly
distro.
--
Robert Uhl <http://public.xdi.org/=ruhl>
We will print no letters to the editor. We will give no space to
opposing points of view. They are wrong. The Underground Grammarian is
at war and will give the enemy nothing but battle.
--The Underground Grammarian
From: Christopher C. Stacy
Subject: Re: I've thought better of Linux
Date:
Message-ID: <uaclq2ain.fsf@news.dtpq.com>
Robert Uhl <·········@NOSPAMgmail.com> writes:
> What would you propose--that every nation in the world outlaw
> working on any desktop environment other than FooDesk?
It's just interesting that, so far, the collective
efforts of the entire open software community have
been unable to get even one of them to actually work.
From: Robert Uhl
Subject: Re: I've thought better of Linux
Date:
Message-ID: <m3wtotj4dd.fsf@4dv.net>
······@news.dtpq.com (Christopher C. Stacy) writes:
>
> > What would you propose--that every nation in the world outlaw
> > working on any desktop environment other than FooDesk?
>
> It's just interesting that, so far, the collective efforts of the
> entire open software community have been unable to get even one of
> them to actually work.
That's an odd position to take, given that they demonstrably _do_ work
for a very large number of people.
In my own case, I do not own a Windows box. And yet I am able to be
happily productive on my computer: I run a blog and associated website;
I retouch my photos; I manage my accounts; I program; I chat with my
friends and so on. From where I'm sitting, it seems to me that my
desktop works pretty damn well.
--
Robert Uhl <http://public.xdi.org/=ruhl>
The average size of text documents doesn't double every 18 months
except for MS Word users. --David Schultz
······@news.dtpq.com (Christopher C. Stacy) writes:
>Robert Uhl <·········@NOSPAMgmail.com> writes:
>> What would you propose--that every nation in the world outlaw
>> working on any desktop environment other than FooDesk?
>
>It's just interesting that, so far, the collective
>efforts of the entire open software community have
>been unable to get even one of them to actually work.
That's because until recently (i.e., a couple years ago, with
KDE/Gnome), noone has actually tried. I would think that most people
(myself included) do not use fully equipped desktop environments, but
more simplistic window-manager-only solutions, that just manage the
dozen or so applications they have been using for years. And that
works pretty well, imho. In fact, it works so well, that I've always
preferred that to Windows.
When the "open source community" (for whatever nebulous value thereof)
wanted to introduce a "desktop environment", they made the mistake of
trying to retrofit the look&feel of other operating systems on top of
Linux (Unix). In the case of KDE, despite the name, it was a mix of
Windows95 and OS/2. In the case of Gnome, it is currently some kind of
studied-by-screenshots Macintosh.
Both are now bloated monstrosities that fail to provide the look&feel
and functionality of the originals they strived to copy. Everytime a
new feature appears in Windows, or on the Macintosh, it is (badly)
copied and put, with little or no thinking, into the desktop
environment. The net result is, that neither long-time X11 users, nor
migrators from Windows or Mac can feel at home. Had they created
something unique from start, something that didn't try and pretend
that it itself is the operating system and no Unix is under the hood,
I'm sure it would've worked out a lot better.
And then there's the bugs. I've used Gnome for a while, mostly for
evalutation if I can put it on someone else's desktop (I opted for
KDE, then) and followed its development thereafter. Gnome 1.4 wasn't
quite as bad but every release of Gnome2 has been horribly buggy. I'm
not talking about a Gnome game crashing every now and then (although
that happens, of course). The thing is designed for end users. How
should an end user cope with Gnome crapping into the gconf registry,
because nautilus crashes in some wicked way, and that then disables
the ability to log back in. Or you can log in again, and all the icons
(or nautilus itself) is gone. You'll have to run a shell (how would
the end user do that, when the desktop is more or less frozen?) and
delete the respective Gnome directories from his homedir manually,
taking care that no important data is lost. I've seen this several
times, and other rather disillusionizing fuckups, like nautilus
blocking the entire Gnome part of the desktop when it's waiting on
some kind of network timeout (hello, is this Win 3.1?) To deal with
these kind of problems, potentially debug and workaround them, is out
of scope for the target audience, namely unsophisticated end users.
In this light, it's little surprising that the only halfway
well-working desktop environments I've seen on X11, namely Openlook
and CDE, were commercial offers. And yes, they could be used by
unsophisticated end users, without crashing and locking up all the
time.
mkb.
Matthias Buelow wrote:
> Had they created
> something unique from start, something that didn't try and pretend
> that it itself is the operating system and no Unix is under the hood,
> I'm sure it would've worked out a lot better.
Why? Why are you blaming these design problems on "it's not
sufficiently like Windows or Mac?" Why aren't they just UI design
problems with not enough open source effort being put into them, period?
Thought and effort are either applied or they aren't. Why is a clean
slate some kind of magic bullet for producing an outstanding desktop UI?
What magic principle would UI designers follow, other than the
principles the mainstream desktop OSes have been using?
I haven't followed Mac much, but every release of Windows has been a bit
of a hodgepodge, shifting the look and feel of stuff, relocating all the
gadgets. The corporate dummies always have to go up a retraining
learning curve when they switch Windows versions. Consequently,
corporations are less interested in gratuitously switching Windows
versions, because it incurs training and productivity costs for them.
At least they need to see evidence that they gain more than they lose by
upgrading.
The only UI principles I have seen in Windows evolution are (1) make it
look sexier, (2) try to rationalize the extant hodgepodge somehow, (3)
add new features. Windows is commercially successful as a desktop, for
many reasons including monopoly power, but clearly this ad-hoc approach
to desktop UI design is not a dealbreaker for tons of consumers.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
"We live in a world of very bright people building
crappy software with total shit for tools and process."
- Ed McKenzie
"Brandon J. Van Every" <·····················@mycompanyname.com> writes:
>Why is
>a clean slate some kind of magic bullet for producing an outstanding
>desktop UI? What magic principle would UI designers follow, other than
>the principles the mainstream desktop OSes have been using?
Simply because of the fact that you (starting simple analogy here)
cannot screw the pedals and steering wheel of a VW Beetle (or a Rolls
Royce) onto a 20 ton truck and expect it to work well. Unix is that
20t truck. It is a very powerful and quite elaborate operating
environment already without a GUI, unlike Windows or (the original)
MacOS. The GUI should integrate into and augment that environment, and
not try to replace it.
As a second best thing, it should just do a few things (like manage
windows and start a few programs), and otherwise stay out of the
user's way. That's what ordinary window managers do. Little value
added but then again, no value taken away.
mkb.
Matthias Buelow <···@incubus.de> writes:
> "Brandon J. Van Every" <·····················@mycompanyname.com> writes:
>
> >Why is
> >a clean slate some kind of magic bullet for producing an outstanding
> >desktop UI? What magic principle would UI designers follow, other than
> >the principles the mainstream desktop OSes have been using?
>
> Simply because of the fact that you (starting simple analogy here)
> cannot screw the pedals and steering wheel of a VW Beetle (or a Rolls
> Royce) onto a 20 ton truck and expect it to work well. Unix is that
> 20t truck. It is a very powerful and quite elaborate operating
> environment already without a GUI, unlike Windows or (the original)
> MacOS. The GUI should integrate into and augment that environment, and
> not try to replace it.
>
> As a second best thing, it should just do a few things (like manage
> windows and start a few programs), and otherwise stay out of the
> user's way. That's what ordinary window managers do. Little value
> added but then again, no value taken away.
>
Its funny, but the aspects of my interface that I like the most were
available under Unix back in 89, when MS was just starting to consider
a "windows" environment. I've always expected MS would add this
feature, but to date I don't think they have and that feature is
virtual desktops. I love being able to just hit a key and ahve another
blank desktop vailable to open other apps on and then, hit another key
and I'm bakc to my previous desktop. I'm constantly frustrated by
desktop clutter under Windows.
I have often thought about a clean slate wrt GUIs. It would be really
interesting to see something really different, but I fear we have all
become so indoctrinated with the existing paradigm, it will take
someone who has never used a computer to come up with something really
new. Possibly the developments in speech recognition will have a major
impact?
Tim
--
Tim Cross
The e-mail address on this message is FALSE (obviously!). My real e-mail is
to a company in Australia called rapttech and my login is tcross - if you
really need to send mail, you should be able to work it out!
On 17 Jun 2005 18:57:43 +1000, Tim X <····@spamto.devnul.com> wrote:
> Its funny, but the aspects of my interface that I like the most were
> available under Unix back in 89, when MS was just starting to
> consider a "windows" environment. I've always expected MS would add
> this feature, but to date I don't think they have and that feature
> is virtual desktops.
<http://www.microsoft.com/windowsxp/downloads/powertoys/xppowertoys.mspx>
--
Lisp is not dead, it just smells funny.
Real email: (replace (subseq ·········@agharta.de" 5) "edi")
Edi Weitz <········@agharta.de> writes:
> On 17 Jun 2005 18:57:43 +1000, Tim X <····@spamto.devnul.com> wrote:
>
> > Its funny, but the aspects of my interface that I like the most were
> > available under Unix back in 89, when MS was just starting to
> > consider a "windows" environment. I've always expected MS would add
> > this feature, but to date I don't think they have and that feature
> > is virtual desktops.
>
> <http://www.microsoft.com/windowsxp/downloads/powertoys/xppowertoys.mspx>
But it doesn't work very nicely last time I checked. Windows is
fundamentally conceived around a single desktop gui.
Gregm
Brandon J. Van Every wrote:
> Matthias Buelow wrote:
>
>
> The only UI principles I have seen in Windows evolution are (1) make it
> look sexier, (2) try to rationalize the extant hodgepodge somehow, (3)
> add new features. Windows is commercially successful as a desktop, for
> many reasons including monopoly power, but clearly this ad-hoc approach
> to desktop UI design is not a dealbreaker for tons of consumers.
>
>
Years ago, someone noticed that Microsoft products are made with the
"preservation of user dissatisfaction" in mind. And so it is.
Microsoft products must always look very capable, but in fact just
barely capable of doing the work, and must always have enough annoying
features to make the customer unhappy but with false hope that the
new version of the software would come with improvements.
The new versions do, but with other annoying features built in, to
preserve the dissatisfaction. And so it goes.
Linux is rubbish, if you insist, but the price is unbeatable.
+ Matthias Buelow <···@incubus.de>:
| I would think that most people (myself included) do not use fully
| equipped desktop environments, but more simplistic
| window-manager-only solutions, that just manage the dozen or so
| applications they have been using for years. And that works pretty
| well, imho. In fact, it works so well, that I've always preferred
| that to Windows.
Me too: I currently run the sawfish WM, and am pretty happy with
that.
ObLisp: I can configure sawfish using a mixture of the supplied GUI
and lisp, and I like that. My biggest problem is understanding the
module system of the underlying lisp. Now modules are very different
critters than packages. From earlier discussions here I understand
that these are somewhat orthogonal concepts. Maybe I haven't been
patient enough, or spent enough time on it, but I still don't quite
get the description of the module system in the librep docs. Can
anyone point me to a different description? Not necessarily of that
particular module system, but of something like it, or a more generic
discussion of module systems in various lisps?
--
* Harald Hanche-Olsen <URL:http://www.math.ntnu.no/~hanche/>
- Debating gives most of us much more psychological satisfaction
than thinking does: but it deprives us of whatever chance there is
of getting closer to the truth. -- C.P. Snow
Harald Hanche-Olsen <······@math.ntnu.no> writes:
> ObLisp: I can configure sawfish using a mixture of the supplied GUI
> and lisp, and I like that. My biggest problem is understanding the
> module system of the underlying lisp. Now modules are very different
> critters than packages. From earlier discussions here I understand
> that these are somewhat orthogonal concepts. Maybe I haven't been
> patient enough, or spent enough time on it, but I still don't quite
> get the description of the module system in the librep docs. Can
> anyone point me to a different description? Not necessarily of that
> particular module system, but of something like it, or a more generic
> discussion of module systems in various lisps?
I have no idea what librep does or anything about Sawfish.
When peple speak of "modules" in other languages,
they may be referring to a way of isolating programs
from each other. Modules are boundaries, possibly
representing compilation units. There are generally
ramifications for program linkage. Some languages may
also have features for supporting defined communication
paths between modules working across processors or networks.
Common Lisp doesn't have any "module system", and the
word "module" isn't really generally used in Lisp.
Common Lisp does have a trivial feature called modules.
It refers to a library claiming to having been loaded.
A module name (any random string) can appear in the
list stored in the global variable *MODULES*.
The function REQUIRE can take a module name,
and through some implementation manner, cause
the module to be loaded. The loader concludes
by calling the function PROVIDE, which does nothing
more than stick the given module name in *MODULES*.
I think REQUIRE/PROVIDE may be deprecated in the ANSI standard.
Let's get packages out of the way, and come back to modules later.
Common Lisp has "packages", which are nothing more than namespaces.
It is not a library; it is not code. It's a bag of symbol names.
Every symbol has a name, like CONS or FOO, and every symbol
also has a "home package". These form the fully qualified name
of the symbol. One of the most important things about a symbol
is that it has a unique name (with respect to its home package).
Every symbol's name really looks like PACKAGE:NAME.
For example, TCP:PACKET or BREAKFAST:PACKET.
When you are working with your code, you don't need to say
the name of the package. At any given moment, a current
package object is bound (as *PACKAGE*) for READ.
As you type in a symbol name, READ "interns" the symbol
in the current package. (That is, READ looks it up by name,
or else creates it.) You just say SHUTTLE rather than
having to spell out SPACE:SHUTTLE.
Packages may export symbols, thus forming an API.
Packages may import symbols, individually or wholesale
from another package. Suppose that you are writing
some source code to be handed to the compiler,
or are typing at Lisp. Let's say the current package
is "CL-USER" (which is the default scratch workspace).
To reference the exported symbol LAUNCH from the SPACE
package, you can say SPACE:SHIP. But if you had only
written LAUNCH, you would have been referencing your
own distinct CL-USER:SHIP symbol. Let's suppose that
you would like to write LAUNCH and mean SPACE:SHIP.
You can "import" SPACE:SHIP into CL-USER.
SHIP now refers to SPACE:SHIP.
If you want to reach into some foreign package to reference
a symbol that has not been exported, you may freely perform
this violation. Just use the double-colon syntax.
So you can reference SPACE::GYRO-INTERNAL-CALIBRATION,
even though it hasn't been exported.
You are now a bad person.
But if you really needed to do that,
Lisp isn't going to get in your way.
The reserved package containing the Common Lisp language
symbols is called COMMON-LISP. Most packages are used for
symbols representing Lisp code, so they "use" COMMON-LISP;
all the exported COMMON-LISP symbols are imported.
You can just write CONS instead of COMMON-LISP:CONS.
Package names do not in themselves mean anything,
and there is no standard naming convention.
Packages can also have nicknames (aliases).
I like to name mine like "COM.DTPQ.WHATEVER".
Packages are not related to files in any way.
At any point in any source file, or when typing
at Lisp, the form IN-PACKAGE is used to establish
the current package (for the remainder of the
input stream, or until another IN-PACKAGE).
Packages are not related to modules or anything, either.
Packages are only about symbol names.
Now let's talk about "Systems".
Common Lisp doesn't have "systems", either,
but the Lisp Machine had a form called DEFSYSTEM.
This is for solving the problem that you have a
bunch of files that go together, and they need
to be compiled and loaded in some order with
some particular dependancies. DEFSYSTEM is
used to describe all of that. You may then call
the function MAKE-SYSTEM, which will analyze your
system dependancies, look at the files and timestamps,
and call the compiler appropriately. There is also
LOAD-SYSTEM, which loads up your system as specified.
The Lisp Machine DEFSYSTEM was very fancy an integrated with
the system version control and release and patching features.
Most Lisp implementations have copied this idea and provided
at least some simple DEFSYSTEM extension. There's a portable
one called MK-DEFSYSTEM. Some programs include their own.
The latest popular portable DEFSYSTEM critter is called ASDF.
(I guess this stands for "Another System Definition Facility").
Systems and Packages interact in the sense that a package has to
be defined before it is used, which is the sort of compile/load
dependancy that is described by the DEFSYSTEM utility.
Now, back to "modules".
The REQUIRE/PROVIDE interface was intended as a least
common denominator, poor man's kind of LOAD-SYSTEM function.
Modules may or may not interact in any way with whatever
nice "systems" (eg. DEFSYSTEM) feature you may have.
Speaking of "features", there's another global variable
called *FEATURES*, which is just a list of symbols that
represent features that have somehow become available
to you. These are used with the #+ and #- reader macros.
This is the equivalent of #ifdef in other languages,
although you can also query *FEATURES* at runtime.
Yes, it's a hack. No, there's no standard for how
features should be named or how to prevent feature
name collisions.
Modules are a lot like features, except they are
relatively useless. There are no "module" objects.
A module is just a name (presumably of something or another),
that somehow got stuck onto the list in *MODULES*.
Sometimes a Lisp image is called a "world", by the way.
(But there are no interplanetary features implied.)
On the Lisp Machine, the entire operating system and
all of the applications were loaded into one gigantic
image (a single address space execution environment),
which was dumped out and booted from. Therefore,
some people still refer to a "Lisp image" as a "world".
So, in the world, you have some packages, which make up the
namespaces in which your symbols "live". You may also have
some modules (which are supposed to represent that you have
loaded some particular program or something). Modules are
a vaccuous abstraction, and do not represent any language
boundary; that's the job of Packages and is only accomplished
lexically. Nothing strictly prevents anything in the Lisp world
from interacting with anything else in the same Lisp world.
(And there's only ever the one world.)
Along comes Linux, and especially Debian (and Gentoo).
They have this wonderful operating system feature called
the "package manager", which is used to automatically
load and manage packages of software on your machine.
You can just type a quick command to tell the system that
you want "Emacs", and presto -- Emacs is magically
installed on your machine and ready for use.
Wouldn't it be nice if you could do something like
that with your Lisp programs, from inside Lisp?
So some nice people created a program called the
Common Lisp Controller, currently ported to
Debian and Gentoo Linux.
Debian's notion of "package" has nothing to do with Lisp's
concept of a "package". Common Lisp Controller sort of
conceptually translates a Debian "package" into a Lisp
"module" or "system".
We'll need something that can register Lisp programs
and their dependancies. It will need to be able to
find the files, see what needs compiling, and load files.
It may need to understand that some systems depend
on first loading other systems.
This is a job for some DEFSYSTEM!
Specifically, they choose ASDF.
Next, some conventions are established for how to package
up arbitrary Lisp libraries for the Debian distribution:
having them use ASDF, and putting their "system definition"
(your program's files that contain the calls to ASDF)
in a particular place.
Finally, the Lisp implementations that want to fully play along
can come pre-built with a standard interface to the ASDF-based
distribution package scheme.
And the interface that was chosen was REQUIRE.
After all, we're not using "modules" for anything else!
So now you can log in to Debian, and use its package
manager to magically fetch some software you want.
At the shell, you type: # apt-get "cl-base64"
and it gets installed on your machine.
Then, in Lisps such as SBCL, you can just fire
it up and say: (require "cl-base64")
and the "cl-base64" package is loaded by ASDF into your
Lisp world, all ready for you to use.
So, they have partially recycled the REQUIRE function
as a handy loading interface for these systems defined
by ASDF. By the way, there are no particularly interesing
standard semantics to the *MODULES* variable; it doesn't
get updated by ASDF when a system is loaded. ASDF decides
if something needs loading or reloading (or if a file needs
recompiling and loading) by looking at other information.
I suppose they could have also made REQUIRE go all the
way by bringing up a shell and doing the "apt-get" command
for you, but they didn't. Perhaps because you have to
be "root" to do that.
So I guess that's what "modules" are, now.
Not very related to "modules" in languages like MODULA.
Aside from extensions like ASDF and Common Lisp Controller,
Lisp doesn't really know or care very much about files.
It basically just has LOAD and COMPILE-FILE, and they
don't care what's in a file, nor relate to packages
or modules or anything else.
Lisp's packages are the language element that provides
logical seperation of programs from one another,
through naming. It works out extremely well.
For more information, please see:
http://www.lispworks.com/documentation/HyperSpec/Body/v_module.htm
http://www.lispworks.com/documentation/HyperSpec/Body/f_provid.htm
http://www.lispworks.com/documentation/HyperSpec/Body/v_featur.htm#STfeaturesST
http://www.lispworks.com/documentation/HyperSpec/Body/11_.htm
http://www.cliki.net/Debian
http://www.cliki.net/common-lisp-controller
http://distrocenter.linux.com/article.pl?sid=02/07/18/0523242&tid=23&tid=29&tid=104
Rep, the Lisp implementation used by Sawfish, is a kind of Scheme.
The language is a lexically scoped "Lisp-1". Every closure is
associated with a "module" object, which provides a namespace
for resolving unbound variable names. Modules may reference other
modules, importing and exporting references. Modules are also
associated with files and dynamic loading and linking.
I don't know anything about rep.
See also:
http://en.wikipedia.org/wiki/Modula-2
http://www.gwydiondylan.org/books/dpg/db_190.html#marker-9-610
http://www.schemers.org/Documents/FAQ/#id2515329
http://pauillac.inria.fr/cdrom/www/bigloo/manual/bigloo-3.html
http://en.wikipedia.org/wiki/Erlang_programming_language
http://librep.sourceforge.net/librep-manual.html
Chris
+ ······@news.dtpq.com (Christopher C. Stacy):
| Harald Hanche-Olsen <······@math.ntnu.no> writes:
|
| > ObLisp: I can configure sawfish using a mixture of the supplied GUI
| > and lisp, and I like that. My biggest problem is understanding the
| > module system of the underlying lisp. Now modules are very different
| > critters than packages. From earlier discussions here I understand
| > that these are somewhat orthogonal concepts.
I should have elaborated a bit on this. I somehow forgot how easy it
is to find things with google. Here is the reference I had in mind.
Of course it's from Kent:
http://www.google.no/groups?selm=sfwogunvny8.fsf%40world.std.com
| I have no idea what librep does or anything about Sawfish.
Well, you certainly provided a very long answer for someone who
doesn't know anything. 8-)
Librep is just rep in a library to be used by applications that need
rep as a scripting language. (And as you noted (obviously you know
/something/, after all) rep is a lexically scoped Lisp-1 which uses
modules of a kind as an information hiding mechanism.) Sawfish used
to be called sawmill, but changed its name because of some name
collision with another product. Much of it is implemented in rep, in
the same way that much of emacs is implemented in elisp.
http://sawmill.sourceforge.net/
| Common Lisp does have a trivial feature called modules.
Yup, and that is not the rep kind of module.
| [...] Not very related to "modules" in languages like MODULA.
MODULA is a good example:
I think rep's modules are a bit like MODULA's modules.
| Lisp's packages are the language element that provides
| logical seperation of programs from one another,
| through naming. It works out extremely well.
Indeed. I certainly find them easier to use and understand than rep's
modules, but that could be because I've had access to tutorials and
books explaining packages, while for rep I only have some rather terse
documentation.
Oh, well. Yeah, before somebody points it out, I know that this is
de facto a CL newsgroup, so this discussion is somewhat off topic
here. Though perhaps less off topic than the thread I branched off
of. ...
| Rep, the Lisp implementation used by Sawfish, is a kind of Scheme.
| The language is a lexically scoped "Lisp-1". Every closure is
| associated with a "module" object, which provides a namespace
| for resolving unbound variable names.
Hmm. So while packages affect the symbol-name -> symbol mapping,
modules affect the symbol -> global binding mapping? I guess that
makes sense. As Kent implies (or seems to ...) in the above-mentioned
reference, there is no reason why you cannot do both in the same
language, though it seems to me it could easily be confusing.
--
* Harald Hanche-Olsen <URL:http://www.math.ntnu.no/~hanche/>
- Debating gives most of us much more psychological satisfaction
than thinking does: but it deprives us of whatever chance there is
of getting closer to the truth. -- C.P. Snow
Harald Hanche-Olsen <······@math.ntnu.no> writes:
> Hmm. So while packages affect the symbol-name -> symbol mapping,
> modules affect the symbol -> global binding mapping? I guess that
> makes sense. As Kent implies (or seems to ...) in the above-mentioned
> reference, there is no reason why you cannot do both in the same
> language, though it seems to me it could easily be confusing.
Indeed, you could. I just don't see any real urgent call for modules
other than from a community of people who is used to solving this
problem another way.
Whereas, within the CL community, there were a lot of programs like
MACSYMA, KEE, MYCIN, ELIZA, and others that traditionally relied
HEAVILY on having complete control of symbols in the Lisp and could
not be co-loaded in large-address implementations like Lisp Machines
where people wanted to have more than one such application loaded at
once. Packages allowed these programs not to trample each other, and
modules would have helped not an iota.
Also, even in the ISO ISLISP design, where we had no formal commitment
to compatibility with the past (and merely a "preference" for
compatibility if there was no reason to avoid it), there was strong
desire for modules with several of the member countries, but we
eventually did not add a modules proposal because no nation managed to
come forward with a working modules proposal. That's just a matter of
history, and not a proof it couldn't be done, but the point is that
traditionally people have whined as they got used to things, but once
settled in haven't really complained, since in the end, it just ins't
what stands in the way of serious programmers getting serious work done.
······@news.dtpq.com (Christopher C. Stacy) writes:
> Robert Uhl <·········@NOSPAMgmail.com> writes:
> > What would you propose--that every nation in the world outlaw
> > working on any desktop environment other than FooDesk?
>
> It's just interesting that, so far, the collective
> efforts of the entire open software community have
> been unable to get even one of them to actually work.
No, there are plenty of users out there who have their desktop working
just fine. I think the problem seems to be limited to your situation,
which should tell you something.
Tim
--
Tim Cross
The e-mail address on this message is FALSE (obviously!). My real e-mail is
to a company in Australia called rapttech and my login is tcross - if you
really need to send mail, you should be able to work it out!
From: Christopher C. Stacy
Subject: Re: I've thought better of Linux
Date:
Message-ID: <uhdfx736r.fsf@news.dtpq.com>
Tim X <····@spamto.devnul.com> writes:
> ······@news.dtpq.com (Christopher C. Stacy) writes:
>
> > Robert Uhl <·········@NOSPAMgmail.com> writes:
> > > What would you propose--that every nation in the world outlaw
> > > working on any desktop environment other than FooDesk?
> >
> > It's just interesting that, so far, the collective
> > efforts of the entire open software community have
> > been unable to get even one of them to actually work.
>
> No, there are plenty of users out there who have their desktop working
> just fine. I think the problem seems to be limited to your situation,
> which should tell you something.
Oh, yeah, sure --- it's just me.
Linux desktop works perfectly for everyone else.
May I please have some of what you are smoking?
Christopher C. Stacy wrote:
> Tim X <····@spamto.devnul.com> writes:
>>>It's just interesting that, so far, the collective
>>>efforts of the entire open software community have
>>>been unable to get even one of them to actually work.
>>
>>No, there are plenty of users out there who have their desktop working
>>just fine. I think the problem seems to be limited to your situation,
>>which should tell you something.
>
>
> Oh, yeah, sure --- it's just me.
> Linux desktop works perfectly for everyone else.
>
> May I please have some of what you are smoking?
No one said "perfectly". No one said "everyone else".
These discussions are quite unproductive as long as participants push
every viewpoint to the extreme. When one says "It doesn't work",
obviously it's going to provoke responses from those for whom it does
work. As for Linux desktops, both KDE and Fluxbox "work" for me. I
prefer them to Windows. But they are far from "perfect". We have
different views, and no one is necessarily "smoking" anything. Get over it.
--
Morten
From: Christopher C. Stacy
Subject: Re: I've thought better of Linux
Date:
Message-ID: <uu0jx5ixi.fsf@news.dtpq.com>
Morten Alver <······@invalid.no> writes:
> Christopher C. Stacy wrote:
> > Tim X <····@spamto.devnul.com> writes:
> >>>It's just interesting that, so far, the collective
> >>>efforts of the entire open software community have
> >>>been unable to get even one of them to actually work.
> >>
> >>No, there are plenty of users out there who have their desktop working
> >>just fine. I think the problem seems to be limited to your situation,
> >>which should tell you something.
> >
> >
> > Oh, yeah, sure --- it's just me.
> > Linux desktop works perfectly for everyone else.
> >
> > May I please have some of what you are smoking?
>
> No one said "perfectly".
> No one said "everyone else".
What, then, is the opposite of "limited to your situation"?
And what "should that tell me"?
On Thu, 16 Jun 2005 20:30:29 +0000, Christopher C. Stacy wrote:
>
> I know there are even more desktop/window environments.
> Why can't Linux just pick ONE and make that really WORK?
> At least as well as Windows?
Because Linux isn't an entity that makes decisions. Linux isn't a person,
and it's not a company, or a standards body. I know you know this - I
just thought it worth reiterating since you asked the question above.
For what it's worth, I've used Linux since 1998, and Gentoo for the past
year. I use it on my server, my desktop and both my laptops. Everything
I buy is "generic" hardware - never the cheapest (since that's usually
crap), but I never bother to check Linux compatibility any more, and I
don't remember the last time I was burned with hardware issues that Google
couldn't find me the answer to in an hour at the most. I use XFCE as my
desktop and I honestly can't recall ever having had a problem with it.
Some of my colleagues use Windows, and they seem to spend their lives
trying to get the OS to leave them alone so that they can get some work
done. I've never used a Mac, so I can't comment. Linux just works for me.
Best wishes,
Bill.
······@news.dtpq.com (Christopher C. Stacy) writes:
> I have been running Debian Sarge on my totally generic
> machine [...]
> RedHat (Fedora) and Knoppix do come up on his machine.
> So he's running Knoppix, [...]
There are distributions for which some hand assembly is required, and
there are distributions designed to be booted on certified hardware
plug-and-play. RedHat, Knoppix, SuSE, etc, are delivered with big
kernels containing modules for all the device drivers, and with
hardware auto-detection and auto-configuration tools. Of course, the
first step is to check that your hardware is certified to work with
these distributions, otherwise you're back to case 1, and here happily
you have the freedom to do the hand assembly (while with proprietary
OSes, you don't).
> For a desktop, I would always recommend Macintosh.
> (I wish I was using MacOS X, myself...)
I wish I could use MacOSX on my AMD CPU. I guess no hand assembly
allowed here will prevent me to do that...
Anyway, you can choose between:
Linux ($2) + generic hardware ($300) + hand assembly (priceless) = $302
MS-Windows-XP ($189.99) + Dell hardware (3000-XP=$247.01) = $462
MacOSX ($129) + Apple hardware (eMac-MacOSX=$670) = $799
Of couse, the Linux option is more difficult to price, either you
consider the hand assembly is worth you something, because you're
learning something, or because you've got the freedom to configure it
like you want, or because of the pleasure you get from twiddling bits,
in which case the priceless value is negative and a linux system comes
out even cheaper, or you don't and you rather hire a Linux technician
to install the system on your computer, and you get an invoice of
between $50 and $500, depending on the compatibility of the hardware
with Linux drivers and technician competence and price, so it may or
may not be cost effective.
There's also another consideration if you want to be precise: the
priceless hand assembly, if you do it yourself or have a friend do it
without invoicing won't involve taxes. In the other cases the state
takes the occasion to rob you. So still more freedom by using Linux!
--
__Pascal Bourguignon__ http://www.informatimago.com/
In a World without Walls and Fences,
who needs Windows and Gates?
Pascal Bourguignon wrote:
> Anyway, you can choose between:
>
> Linux ($2) + generic hardware ($300) + hand assembly (priceless) = $302
> MS-Windows-XP ($189.99) + Dell hardware (3000-XP=$247.01) = $462
> MacOSX ($129) + Apple hardware (eMac-MacOSX=$670) = $799
+ how much time to look up what components ("generic hardware")
work with Linux, to set Linux up (unless it works out of the box
then). And does this apply to Laptops, too, which can't be
hand-assembled from components (well, usually aren't)?
--
Don't let school interfere with your education. -- Mark Twain
Ulrich Hobelmann <···········@web.de> writes:
> Pascal Bourguignon wrote:
>> Anyway, you can choose between:
>> Linux ($2) + generic hardware ($300) + hand assembly (priceless) =
>> $302
>> MS-Windows-XP ($189.99) + Dell hardware (3000-XP=$247.01) = $462
>> MacOSX ($129) + Apple hardware (eMac-MacOSX=$670) = $799
>
> + how much time to look up what components ("generic hardware") work
> with Linux, to set Linux up (unless it works out of the box then).
> And does this apply to Laptops, too, which can't be hand-assembled
> from components (well, usually aren't)?
Not too much, you just keep you hardware compatibility Howto close to
your eyes when selecting the hardware. Same job when selecting PC
component to build a NeXTSTEP system, or a Solaris system. Or even a
MS-Windows system, for MS-Windows while running on a wider range of
hardware (and happily for them, on clearly labelled as compatible
hardware), does not run on everything either.
--
__Pascal Bourguignon__ http://www.informatimago.com/
You never feed me.
Perhaps I'll sleep on your face.
That will sure show you.
> Ulrich Hobelmann <···········@web.de> writes:
>
>
>>Pascal Bourguignon wrote:
>>
>>>Anyway, you can choose between:
>>> Linux ($2) + generic hardware ($300) + hand assembly (priceless) =
>>>$302
>>> MS-Windows-XP ($189.99) + Dell hardware (3000-XP=$247.01) = $462
>>> MacOSX ($129) + Apple hardware (eMac-MacOSX=$670) = $799
>>
>>+ how much time to look up what components ("generic hardware") work
>>with Linux, to set Linux up (unless it works out of the box then).
>>And does this apply to Laptops, too, which can't be hand-assembled
>>from components (well, usually aren't)?
I put 1.5 weeks into trying to get Linux working on my HW. I'd say that
was at least 40 hours of work. If I paid myself $8/hour that's $320.
So I'd say the time cost is at least comparable to going the XP+Dell
route. It does depend on whether you've got more time than money though.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
I won't spend more than 1 day configuring 1 thing.
On Fri, 17 Jun 2005 09:42:28 -0700, Brandon J. Van Every wrote:
>
> I put 1.5 weeks into trying to get Linux working on my HW. I'd say that
> was at least 40 hours of work. If I paid myself $8/hour that's $320.
> So I'd say the time cost is at least comparable to going the XP+Dell
> route. It does depend on whether you've got more time than money though.
These days I have very little spare time at all. And yet, last week I
managed to put together a nice Gentoo server from generic components I
bought at Frys. Total time spent, about three hours including the time
taken to buy the stuff. Everything just works, and the whole thing was
really quite cheap (< $400 I would estimate, but I haven't taken the
time to add it all up). And I didn't bother to check hardware
compatibility before I bought any of it - I never bother anymore.
Best wishes,
Bill.
William Bland wrote:
> And I didn't bother to check hardware
> compatibility before I bought any of it - I never bother anymore.
Probably because your server doesn't need a dial-up modem, on-board graphics
or a printer.
On Fri, 17 Jun 2005 10:18:26 -0700, alex goldman wrote:
> William Bland wrote:
>
>> And I didn't bother to check hardware
>> compatibility before I bought any of it - I never bother anymore.
>
> Probably because your server doesn't need a dial-up modem, on-board graphics
> or a printer.
Fair point. I was burned a couple of years back by a Lexmark printer that
advertised support for Linux. Turned out it was for one specific version
of RedHat, the drivers were closed, there was never an updated release,
and getting them to work on anything else was pretty much impossible.
Having said that, I still don't check hardware compatibility, even when
buying printers - I just buy HP now.
Cheers,
Bill.
From: Robert Uhl
Subject: Re: I've thought better of Linux
Date:
Message-ID: <m3is0cllh4.fsf@4dv.net>
William Bland <·······@abstractnonsense.com> writes:
>
> Fair point. I was burned a couple of years back by a Lexmark printer that
> advertised support for Linux. Turned out it was for one specific version
> of RedHat, the drivers were closed, there was never an updated release,
> and getting them to work on anything else was pretty much impossible.
Hey--I had that printer too!
--
Robert Uhl <http://public.xdi.org/=ruhl>
There are two days when a woman is a pleasure: the day one marries her
and the day one buries her. --Hipponax, c. 570-520 BC
Robert Uhl <·········@NOSPAMgmail.com> writes:
> William Bland <·······@abstractnonsense.com> writes:
> >
> > Fair point. I was burned a couple of years back by a Lexmark printer that
> > advertised support for Linux. Turned out it was for one specific version
> > of RedHat, the drivers were closed, there was never an updated release,
> > and getting them to work on anything else was pretty much impossible.
>
> Hey--I had that printer too!
>
I read all this thread and continue to be amazed at how obviously
lucky I've been! I was given a Lexmark 403910+ printer a few years
back and its worked flawlessly under Linux form the day I got it. I've
used it with lpr and now CUPS and have been very happy with it.
Tim
--
Tim Cross
The e-mail address on this message is FALSE (obviously!). My real e-mail is
to a company in Australia called rapttech and my login is tcross - if you
really need to send mail, you should be able to work it out!
William Bland <·······@abstractnonsense.com> wrote:
> On Fri, 17 Jun 2005 10:18:26 -0700, alex goldman wrote:
>> William Bland wrote:
>>> And I didn't bother to check hardware
>>> compatibility before I bought any of it - I never bother anymore.
>>
>> Probably because your server doesn't need a dial-up modem, on-board graphics
>> or a printer.
>
> Fair point. I was burned a couple of years back by a Lexmark printer that
> advertised support for Linux. Turned out it was for one specific version
> of RedHat, the drivers were closed, there was never an updated release,
> and getting them to work on anything else was pretty much impossible.
> Having said that, I still don't check hardware compatibility, even when
> buying printers - I just buy HP now.
I'd say that what you had there was a "Lexmark lied to you" problem, not
a "Linux" problem.
But I'll warn you that HP (or their resellers) have also been known to
lie, too. We bought an HP large-format printer (DesignJet 500ps) that
was advertised by the reseller as having PostScript support ... as it
turned out, the "PostScript support" was a software component of the
Windows driver.
HP shipped no Mac OS X, Linux, or other Unix drivers for this printer.
The printer's proprietary language was not supported by any available
open-source drivers ... a third-party rasterizer for Solaris and SGI was
available, for $500 per host!
Today, there is *supposedly* a way to get it to work with free drivers,
but the utter absence of debugging information from the printer (even
the uninformative "MIO ERROR" that some HPs will give you to acknowledge
that they saw your job but didn't like it) means I haven't been able to
get it going.
It was enough to remind me of the, ahem, special role that proprietary
printer software played in the founding of the Free Software Movement.
--
Karl A. Krueger <········@example.edu> { s/example/whoi/ }
"Karl A. Krueger" <········@example.edu> writes:
>lie, too. We bought an HP large-format printer (DesignJet 500ps) that
>was advertised by the reseller as having PostScript support ... as it
>turned out, the "PostScript support" was a software component of the
>Windows driver.
Why didn't you return it when you found out it didn't have the
advertised features?
mkb.
Matthias Buelow <···@incubus.de> wrote:
> "Karl A. Krueger" <········@example.edu> writes:
>>lie, too. We bought an HP large-format printer (DesignJet 500ps) that
>>was advertised by the reseller as having PostScript support ... as it
>>turned out, the "PostScript support" was a software component of the
>>Windows driver.
>
> Why didn't you return it when you found out it didn't have the
> advertised features?
Wasn't my decision.
--
Karl A. Krueger <········@example.edu> { s/example/whoi/ }
······@news.dtpq.com (Christopher C. Stacy) writes:
> "joesb" <··········@gmail.com> writes:
> I have been running Debian Sarge on my totally generic
> machine for the last year. Here's a short list of
> complaints off the top of my head.
>
> 1. It just doesn't work with my generic (HP) scroll mouse.
> I can't figure out why. (Actually, it starts out working,
> for a few minutes, then it totally stops working!)
With some mice, you can get complications arising between gpm (the
mouse driver for the text consoles) and X. When I had some problems
with an optical mouse, I found the solution was very simple. All I had
to do was follow advice from the gpm man pages on setting up gpm to
use 'raw' mode. Check the gpm HOWTOs and google and I'm sure you find
a straight forward solution. Alternatively, try disabling/turning off
gpm before starting X.
> 2. Cut and paste do not work in the applications that I use
> (eg. paste a URL or text between Emacs and Firefox).
I suspect this is also due to a misconfigured mouse. As mentioned,
either its interference between gpm and the X mouse driver or you have
not configured the mouse driver correctly in XF86Config. Is the scroll
wheel working? If not, its very likely you have the mouse configured
for a very basic 3 button mouse rather than a 5 button scroll wheel
mouse.
>
> 3. Sound doesn't work right.
> Audio applications mess up and even crash all over the place.
Which sound subsystem are you running - OSS or Alsa? Apparently Alsa
is the direction Linux is going with its sound subsystem. However, I'm
still using OSS based subsystem and it works really well for me. I do
a lot of sound related work, including a lot of Text-to-speech type
applications and they work really well.
>
> 4. I had to manually add and configure the totally
> generic network device driver.
Did you add the kernel module for the network driver to the
/etc/modules file and run update-modules?
>
> 5. I had to manually configure X for the totally generic
> video card. And I fucked something up: when Gnome launches,
> there's an extra mouse cursor ("X") sitting in the middle
> of the screen. It's a turd. I can make it go away
> by launching a shell and clicking on its titlebar.
> Then I close that shell, because I do everything from Emacs.
> (Clicking on the Emacs titlebar doesn't clear the turd.)
Again, almost certainly a misconfigured mouse driver for X. Get that
sorted out and many of your problems will go away. Doubt it is
anything to do with the vdeo settings. Note that recent versions of
the X server have a switch which you can use that will get the server
to return what it thinks is your hardware and how it should be
configured. This can be very useful when trying to diagnose
problems. Also note that using the frame buffer device is a useful way
to get X to work on video hardware it doesn't automatically support. I
have only needed this once, when putting Linux on a Dell laptop about
6 years ago. At that time, Linux didn't support the video chipset -
using the virtual frame buffer device overcame this limitation and
Linux worked without issues running X on the laptop.
>
> 5. I tried using KDE, but when I tried to customize the
> Start menu, the system got all confused, lost apps, etc.
Personally, I hate overly complex window managers. I've got rid of
Gnome, KDE etc and now just run sawfish, which I really like as its
very simple, but being written in rep, is nice an lispish and very
easy to extend. My whole X session consists of
#!/bin/bash --login
xsetroot -solid black &
xscreensaver -no-splash &
xload -bg black -fg green -hl red -update 5 -geometry 170x69+4-4 &
xmailbox -geometry 64x38+607-1 &
pload -bg black -fg green -logscale -horiz -incolor cyan -outcolor red -device eth0 -geometry 241x69+181-5 &
xmem -bg black -fg green -geometry 177x71+427-4 &
swisswatch -geometry 100x100-3--11 &
ssh-add &
if [ -x /usr/bin/sawfish ]; then
exec /usr/bin/sawfish
else
echo "Error: /usr/bin/sawfish could not be executed"
exec xterm -geometry 80x24+0+0
fi
this is running on a Debian Testing distro.
> I know there are even more desktop/window environments.
> Why can't Linux just pick ONE and make that really WORK?
> At least as well as Windows?
Firstly, the comment "Why can't Linux just pick ONE and make that
really WORK?" reveals what I think is a basic conceptual
flaw. Firstly, it implies Linux is something akin to a company like
Microsoft or Apple. In reality, Linux is just the kernel and a few
utils. All the rest is really just packages collected by the Distro
(Debian, Red Hat, SuSe, etc). Therefore, there is no central body to
just "pick one and make it work". I'd also argue many of the window
managers do work as well and even better than Windows. How many of us
have defined a colour scheme under windows only to find some apps
refuse to follow it, or even worse, follow only part of it and you end
up with dialog boxes or windows using the same colour for the
foreground and background?
> I suppose I should be thankful that the entire destkop
> doesn't crash all the time, like it did a few years ago.
I sometimes wonder why it is I hear about people having all these
problems, yet I've never really experienced them. Its true that back
when I first started using Linux, I did have some hardware
compatibility problems, but I've not had these for a long time
now. I've NEVER in over 10 years of Linux had my desktop crash
once. I've never had my system do anything equivelent to a blue screen
of death or lock up on me. I have had apps crash now and again and
even lock up, but always been able to rercover without a system
reboot. I have had to spend some time geting unusual hardware working,
but I've had that sort of problem with every OS I've ever used (I
still can't get XP to work with my old HP scanner - luckily, it worked
with Linux first go).
>
> 7. The "security update notification" widget doesn't work.
> It beeps and complains that security upgrades are ready.
> Then it launches Synaptic, which doesn't agree.
> I am left wondering what critical security patches I am missing.
What have you got in your /etc/apt/sources.list file? have you got the
site for the security updates as well as the normal source sites? I
rarely using anything other than apt-get since thats pretty much what
all the other tools use under the hood anyway. sometimes, I may use
dselect, but thats pretty rare. As I tend to use the "testing" distro
of debian, I don't worry too much about the security updates as they
just end up in the normal testing archives anyway (You only need the
security updates url in the sources.list file when your running the
stable distro as security updates are not placed in the normal archive
area for the distro, but rather in the specific security archive
area).
> By running apt manually, I suspect the problem is
> related either to qpopper or wine. I think these problems
> first manifested the week that Sarge became Stable.
> I suppose I ought to look into them.
What does qpopper or wine have to do with apt and security updates?
Isn't one a pop server and the other a windows emulator?
>
> 8. I spent two hours yesterday trying to install a working imapd.
> Tried several different packages. No success so far.
I don't run imapd on my desktop - don't have any use for it, so can't
help with this one.
> 9. Maybe it's an xinetd problem.
> Nothing like having to manually configure a seperate inetd.
> It's supposed to be a replacement, but it's not actually
> integrated into the system! Applications install as inetd,
> then I run some perl script over inetd.conf and manually
> edit the xinetd. Someday I think some package will need
> to configure and start a network service before completing
> installation. I just can't wait to watch that nightmare.
I didn't like xinetd very much. We do run it at work on our Red Hat
based servers. For a desktop machine, inetd is fine as I use iptables
to control who has access from where and to what, so there is not real
security issue with inetd.
>
> 10. Tried to get my friend to run Debian, since he wanted
> to use Linux for some program development.
>
> But Sarge doesn't work with his generic video card on
> the generic $400 AMD64 system he bought the other day
> at the computer store. If he wants to run Sarge, he will
> have to hire a consultant to manually configure X Windows.
> I haven't been able to help him over the phone, because
> he understandably lost patience trying to debug this.
> I am sure there would have been many more problems,
> but not having a display screen was the first show stopper.
>
> RedHat (Fedora) and Knoppix do come up on his machine.
> So he's running Knoppix, but with all the problems,
> he's abandoned Linux for now, and the machine just
> sits there. He uses Windows XP, instead.
> He's been a system programmer for 25 years, too,
> but he has practically no experience with Unix.
> At work (doing real-time embedded systems),
> he uses RedHat, which someone else set up and maintains.
>
> Windows 2000 and XP "just works" perfectly out of the box,
> in all those respects. If you're not sure: that means you
> merely apply AC power to the system, and you're done.
>
> Let us not even speak about trying to do "advanced" things
> like set up local network file sharing on Linux.
> Yes, of course, I have samba working on my network.
> But on Windows, you just do this by clicking on menu items
> named "Share" and "Permissions"; nothing else needed.
>
> I'm a system programmer and have administered probably
> every kind of Unix-based system over the last 25 years
> for both tiny and huge organizations of various kinds.
> If my perception is that Linux doesn't work "out of the box",
> it's a safe bet that newbies who just want to turn on their
> desktop and get to work will be in for a very rough ride.
>
> I wouldn't try to kid them.
> You're just going to wind up souring them on Linux.
>
> I had to reinstall my Windows 2000 desktop last week.
> That horrible nightmare happens about twice a year.
> But it's far less time than I spend trying to make Linux work.
>
> But I also have extensive experience with Windows 2000
> in a server environment, and I would NOT recommend that!
> Some flavor of Unix is going to be much better.
> Lately, I use BSD and Debian for that, and Debian has
> been much better in that role than BSD, to my surprise.
> I once tried to install Gentoo, and that was a bad joke.
> Maybe Gentoo will get better, though.
> In the recent past I've had very good luck with Solaris.
>
> For a desktop, I would always recommend Macintosh.
> (I wish I was using MacOS X, myself...)
What I really find hard to understand is all the people out there who
continually make comparisons with commercial OSs like Windows and OS X
to "free" ones like Linux and BSD and then complain about how lousy
they are when they don't get the same degree of polish or "out of the
box" support. There is no such thing as a free lunch. If you don't
want to pay for a polished commercial product, be prepared to put in a
little extra brain time to figure some of it out.
For your friend who wants Linux to do some development, but doesn't
want to spend time trying to work things out and configure the system,
then pay for a commercial Linux distro like Red Hat Workstation - its
still relatively cheap and you get support (but check their support
matrix to ensure your hardware will be supported).
Tim
--
Tim Cross
The e-mail address on this message is FALSE (obviously!). My real e-mail is
to a company in Australia called rapttech and my login is tcross - if you
really need to send mail, you should be able to work it out!
From: Christopher C. Stacy
Subject: Re: I've thought better of Linux
Date:
Message-ID: <uy8995j1y.fsf@news.dtpq.com>
Tim X <····@spamto.devnul.com> writes:
> > 1. It just doesn't work with my generic (HP) scroll mouse.
> > I can't figure out why. (Actually, it starts out working,
> > for a few minutes, then it totally stops working!)
>
> With some mice, you can get complications arising between gpm (the
> mouse driver for the text consoles) and X. When I had some problems
> with an optical mouse, I found the solution was very simple. All I had
> to do was follow advice from the gpm man pages on setting up gpm to
> use 'raw' mode. Check the gpm HOWTOs and google and I'm sure you find
> a straight forward solution. Alternatively, try disabling/turning off
> gpm before starting X.
Thanks for this hint (and any other specific technical
advice you want to give me). I have never even heard of
gpm, but I suppose if I poke around, I can figure out where
it gets turned on.) But see below about whether it's a
problem elsewhere.
> > 2. Cut and paste do not work in the applications that I use
> > (eg. paste a URL or text between Emacs and Firefox).
>
> I suspect this is also due to a misconfigured mouse. As mentioned,
> either its interference between gpm and the X mouse driver or you have
> not configured the mouse driver correctly in XF86Config. Is the scroll
> wheel working? If not, its very likely you have the mouse configured
> for a very basic 3 button mouse rather than a 5 button scroll wheel mouse.
I don't think the cut/paste has anything to do with the
mouse driver. It happens on other people's machines, too.
General pointing and clicking works during the entire time.
Also, seems random. Sometimes cutting and pasting fails.
Sometimes immediately thereafter, it works. Then, maybe not.
I suspect that it's a bug in either Emacs or Firefox.
Some people tell me that cut and paste with those apps
just never work for them at all. (I was about to tell
them they were crazy and that it worked for me.
Then I tried it.... But I always thought it worked!
Maybe it used to?)
The scroll wheel works, sometimes, right after the machine
is booted, but then soon stops working and never works again.
One way to get it to stop working is to use the KVM switch.
However, it always works on all the other machines on the
switch, which are running Windows. Maybe the driver has
some bug related to being sensitive to something that
happens when I switch. (I don't even know what happens,
electrically, if anything, when I use the KVM switch.)
Here's what the black magic in my XF86Config-4 file says
about my mouse (which is a two-button + scroll mouse).
I wound up writing the whole XF86Config-4 by hand,
so it's entirely possible I messed this up.
Section "InputDevice"
Identifier "PS/2 Mouse"
Driver "mouse"
Option "Protocol" "IMPS/2"
Option "Device" "/dev/psaux"
Option "Emulate3Buttons" "no"
Option "ZAxisMapping" "4 5"
EndSection
Section "ServerLayout"
Identifier "XFree86 Configured"
Screen 0 "Screen0" 0 0
InputDevice "Keyboard0" "CoreKeyboard"
InputDevice "PS/2 Mouse" "CorePointer"
EndSection
There's also this junk, and I don't remember
what any of it is about. Maybe there's
something fishy in there:
Section "Module"
Load "dbe"
Load "dri"
Load "extmod"
Load "glx"
Load "record"
Load "xtrap"
Load "speedo"
Load "type1"
####
Load "GLcore"
Load "bitmap"
Load "ddc"
Load "freetype"
Load "int10"
Load "vbe"
EndSection
> > 5. I had to manually configure X for the totally generic
> > video card. And I fucked something up: when Gnome launches,
> > there's an extra mouse cursor ("X") sitting in the middle
> > of the screen. It's a turd. I can make it go away
> > by launching a shell and clicking on its titlebar.
> > Then I close that shell, because I do everything from Emacs.
> > (Clicking on the Emacs titlebar doesn't clear the turd.)
>
> Again, almost certainly a misconfigured mouse driver for X.
I think the extra mouse cursor may be a bug in Gnome.
The problem doesn't happen in KDE.
> Note that recent versions of the X server have a switch
> which you can use that will get the server to return
> what it thinks is your hardware and how it should be
> configured.
There was soms sort of X autoconfiguration utility
that ran during system installation, but it failed.
> Personally, I hate overly complex window managers.
> I've got rid of Gnome, KDE etc and now just run sawfish
So the best way to use Linux as a desktop machine that's
like Windows or MacOS is not to use the desktop software?
Maybe I should use the Zen window manager.
I could just go back to twm, I guess.
But that doesn't really seem like progress.
YMMV.
> Firstly, the comment "Why can't Linux just pick ONE and
> make that really WORK?" reveals what I think is a basic
> conceptual flaw. Firstly, it implies Linux is something
> akin to a company like Microsoft or Apple.
I am making a comment on the quality of the net result
of particular major project developed for Linux under
the free software theory of the universe, specifically
noting that the people involved in the effort have
been unable to come together to produce a high quality
component that they identify as a critical centerpiece
and which they prostletyze as being comparable or better
than the one on the commercial systems.
Instead, it seems that they got into a fight about
whose broken software was more politically correct.
I don't want a political mission or a hobby,
and I don't want to debug the operating system.
I just want the computer to work so I can use
it to run basic applications that I want.
If Linux is not up to that, fine.
Let's just be honest about it, that's all.
> All the rest is really just packages collected by
> the Distro (Debian, Red Hat, SuSe, etc). Therefore, there
> is no central body to just "pick one and make it work".
X Windows and KDE and Gnome are "little packages"?
> I'd also argue many of the window managers do work as
> well and even better than Windows.
Perhaps, but not the ones that are the standard choices
promoted and offered in the popular distributions.
> How many of us have defined a colour scheme under windows
> only to find some apps refuse to follow it, or even
> worse, follow only part of it and you end up with dialog
> boxes or windows using the same colour for the foreground
> and background?
I have never tried to change the colors under Windows,
or Linux for that matter. I don't care about things
like that. I can believe that it's broken in Windows.
I merely want Linux to correctly autodetect my video card
and mouse and stuff, like Windows _always_ does.
Well, and it would be nice if the font selection
features worked better. (They don't work better
in Windows, by the way; they just don't look so ugly
there that I feel compelled to mess with them.)
Other than that, and some other minor hard-to-fathom
general buggyness in the desktop environment.
I am fairly happy with Linux. I'm mostly giving
examples of things that would terribly impact
your average desktop user.
If I start using it as my desktop, though, I will want
the sound to work. Also the scanner, camera, printer,
DVD writer, etc. I haven't tried any of that stuff.
> I sometimes wonder why it is I hear about people having all
> these problems, yet I've never really experienced them.
Because in your very limited set of experiences
you happen to have been very lucky?
Because your hardware happens to work with the standard drivers?
Because you're not using the standad desktop environment?
To me, it's no wonder at all.
Do you have some more plausible theory?
> > 7. The "security update notification" widget doesn't work.
> > It beeps and complains that security upgrades are ready.
> > Then it launches Synaptic, which doesn't agree.
> > I am left wondering what critical security patches I am missing.
>
> What have you got in your /etc/apt/sources.list file? have you got the
> site for the security updates as well as the normal source sites?
I have:
deb http://debian.crosslink.net/debian/ testing main
deb-src http://debian.crosslink.net/debian/ testing main
deb cdrom:[Debian GNU/Linux testing _Sarge_ - Official Snapshot i386 Binary-2 (20041231)]/ unstable contrib main
deb cdrom:[Debian GNU/Linux testing _Sarge_ - Official Snapshot i386 Binary-1 (20041231)]/ unstable contrib main
deb http://security.debian.org/ testing/updates main contrib
deb http://wine.sourceforge.net/apt/ binary/
> What does qpopper or wine have to do with apt and security updates?
> Isn't one a pop server and the other a windows emulator?
One time recently when trying to debug the "security" problem,
I was playing around with Synaptic and got it to mention that
something was funky about those two packages. (At least I think
that's where I remember seeing the message. I sort of gave up
paying close attention to the system recently, figuring that
I'll probably have to do some kind of total re-install soon.
Luckily, I'm not using it for anything critical lately.)
My theory is that possibly something about the status of those
two random packages is confusing the buggy security update
icon widget. As I mentioned, I believe this all started on
the day that Sarge changed over to stable.
> There is no such thing as a free lunch. If you don't
> want to pay for a polished commercial product,
> be prepared to put in a little extra brain time
> to figure some of it out.
This is why I do mostly use commercial products.
> For your friend who wants Linux to do some development,
> but doesn't want to spend time trying to work things out
> and configure the system, then pay for a commercial
> Linux distro like Red Hat Workstation - its still
> relatively cheap and you get support (but check their
> support matrix to ensure your hardware will be
> supported).
This is the same as saying, "For your aged grandmother who
just wants to drive her car to the grocery store, but doesn't
want to spend time figuring out how the catalytic converter
works -- and it's going to need some tweaking --- she should
buy a commercially sold vehicle".
And I fully agree.
> What I really find hard to understand is all the people out there
> who continually make comparisons with commercial OSs like Windows
> and OS X to "free" ones like Linux and BSD and then complain about
> how lousy they are when they don't get the same degree of polish or
> "out of the box" support.
I'm not complaining. Did I start this thread?
No. I am merely responding to people who seem to be
claiming that Linux is as polished as those commercial
systems, and particularly well suited for your average
home desktop user.
I think RedHat 9 worked better in alot of these respects,
as I recall. My buddy's recent flirtation with Fedora
tends to support that recollection. But Debian was touted
as the stable Linux, and I like its package system better,
and it seems to get the most attention from the developers
of Lisp implementations such as CMUCL/SBCL.
I think Linux is a much better server than Windows.
Rabid announcements of its superiority in the desktop
domain are still somewhat premature, however.
I expect this to greatly improve in the future,
unless the Linux proponents deny the reality of the
situation and how far they have to go. In the past,
there seemed to be some weird antipathy towards improvements
in these areas. I thought that had passed,
but lately I am not so sure.
I am still hoping to be able to switch my main computer
off of MS Windows. But Linux isn't quite there yet.
From: Bulent Murtezaoglu
Subject: Re: I've thought better of Linux
Date:
Message-ID: <87k6ktxgwe.fsf@p4.internal>
>>>>> "CCS" == Christopher C Stacy <······@news.dtpq.com> writes:
[...]
CCS> I don't think the cut/paste has anything to do with the mouse
CCS> driver. It happens on other people's machines, too. General
CCS> pointing and clicking works during the entire time. Also,
CCS> seems random. Sometimes cutting and pasting fails.
CCS> Sometimes immediately thereafter, it works.
Occasionally this happens to me with Xemacs and firefox. But I never
looked into it. If I were to, I start with:
http://www.jwz.org/doc/x-cut-and-paste.html
[...]
CCS> So the best way to use Linux as a desktop machine that's like
CCS> Windows or MacOS is not to use the desktop software? Maybe I
CCS> should use the Zen window manager.
I have never figured out what the point of all the KDE and Gnome stuff
was. I do use Nautilus which brings in some Gnome things including
the gonfiguration daemon and places icons (that are always occluded)
onto the root window. That's the only thing apt-get upgrade
occasionally breaks, but even that hasn't happened in a while. I like
it when I sort out pictures or drag & drop music into xmms. Other
than that whatever the point of all those bars and whatever are, it is
lost on me. I run sawfish because I got it to work the way I like
once when I first discovered it. I suppose I could be using tvtwm and
not notice a difference. What do the desktop software you refer to
do? I don't run Linux as a Windows or MacOS replacement, I run it
because I am used to and reasonably happy with Unix+X. I tested
Windows once (NT 4.0) as my main machine and I couldn't get the
keyboard focus to follow the mouse w/o autoraise consistently and
satisfactorily (there was some half-broken package that did it). Oh
and little boxes popping up and grabbing focus drove me nuts. So that
lasted less than a day. I still run that under vmware, and seems to
work OK to fire up lispworks and produce those elusive singe exe lisp
applications.
[...]
CCS> I could just go back to twm, I guess. But that doesn't
CCS> really seem like progress. YMMV.
I think things should be considered progress when they do something
that is deemed useful by the particular user. Drag and drop is useful
to me under some circumstances, so that's progress for me. Task bars
and things that live there, or menus and scrollbars in emacs aren't
useful for me, so I consider them frivolous and perhaps misguided and
learn to turn them off. What I like is that I don't have to take some
group's pre-packaged progress.
[...]
CCS> ... I am fairly happy with Linux. I'm mostly giving
CCS> examples of things that would terribly impact your average
CCS> desktop user.
I understand.
CCS> If I start using it as my desktop, though, I will want the
CCS> sound to work. Also the scanner, camera, printer, DVD
CCS> writer, etc. I haven't tried any of that stuff.
Sound under ALSA can be problematic because some software still use
OSS. If you have OSS emulated, everything seems to work fine. Other
than that it should be OK. CD burning just works, but I have learned
to use cdrecord ages ago. Scanner, dunno, I have an epson that's been
sitting here in its box for more than a year making me wonder why I
bought the thing if I can't be bothered to even try it. For cameras
used to use a card reader, now I just plug mine in and mount it as a
drive. Postscript printers have never been a problem, but I just use
B&W and the fanciest thing I do is duplex printing.
[...]
CCS> Because in your very limited set of experiences you happen to
CCS> have been very lucky?
Perhaps. But I have been using it since 95 (that is when my GraphOn
X-terminal croaked). Perhaps more importantly I was occasionally
paid to get linux to work for various uses so whatever time I sunk
into learning it was mostly either directly or indirectly covered
(except X).
CCS> Because your hardware happens to work
CCS> with the standard drivers?
No, I select hardware that I am fairly sure will work and then I use
them for ages.
CCS> Because you're not using the
CCS> standad desktop environment?
Yes.
CCS> To me, it's no wonder at all.
CCS> Do you have some more plausible theory?
There's no magic there for me. So yes, yours is a plausible theory.
CCS> ... My theory is that possibly
CCS> something about the status of those two random packages is
CCS> confusing the buggy security update icon widget. As I
CCS> mentioned, I believe this all started on the day that Sarge
CCS> changed over to stable.
I don't use those icons and such. I just read the debian security
mails and either use apt or wget + dpkg directly. This is what I
used to do when I had servers facing the net running debian, so it
is more habit than anything else. I had no idea those notifier
widgets existed, anyway.
cheers,
BM
······@news.dtpq.com (Christopher C. Stacy) writes:
> Tim X <····@spamto.devnul.com> writes:
>
> > > 1. It just doesn't work with my generic (HP) scroll mouse.
> > > I can't figure out why. (Actually, it starts out working,
> > > for a few minutes, then it totally stops working!)
> >
> > With some mice, you can get complications arising between gpm (the
> > mouse driver for the text consoles) and X. When I had some problems
> > with an optical mouse, I found the solution was very simple. All I had
> > to do was follow advice from the gpm man pages on setting up gpm to
> > use 'raw' mode. Check the gpm HOWTOs and google and I'm sure you find
> > a straight forward solution. Alternatively, try disabling/turning off
> > gpm before starting X.
>
> Thanks for this hint (and any other specific technical
> advice you want to give me). I have never even heard of
> gpm, but I suppose if I poke around, I can figure out where
> it gets turned on.) But see below about whether it's a
> problem elsewhere.
Its pretty easy to eliminate gpm from the equation and see if that
makes things work better. try the following
1. Log out of X so that the screen returns to the gdm/xdm login window
2. Switch to a virtual console with Ctl-Alt-F1
3. login as root and type
/etc/init.d/gpm stop
4. Switch back to X with Alt-F7
5. At the X login window, hit Ctl-Alt-backspace. This will restart the
X server.
6. Login and see if the mouse behaves better. If it does, then we know
the problem is due to an interaction between gpm and the X mouse
driver. At this pint you can either remove the gpm package or
reconfigure gpm to pass through mouse events to X. If you never/rarely
use the virtual console and don't use a mouse when you do, then its
probably easiest to just remove gpm with apt-get remove gpm.
> > > 2. Cut and paste do not work in the applications that I use
> > > (eg. paste a URL or text between Emacs and Firefox).
> >
> > I suspect this is also due to a misconfigured mouse. As mentioned,
> > either its interference between gpm and the X mouse driver or you have
> > not configured the mouse driver correctly in XF86Config. Is the scroll
> > wheel working? If not, its very likely you have the mouse configured
> > for a very basic 3 button mouse rather than a 5 button scroll wheel mouse.
>
> I don't think the cut/paste has anything to do with the
> mouse driver. It happens on other people's machines, too.
> General pointing and clicking works during the entire time.
> Also, seems random. Sometimes cutting and pasting fails.
> Sometimes immediately thereafter, it works. Then, maybe not.
> I suspect that it's a bug in either Emacs or Firefox.
> Some people tell me that cut and paste with those apps
> just never work for them at all. (I was about to tell
> them they were crazy and that it worked for me.
> Then I tried it.... But I always thought it worked!
> Maybe it used to?)
It might be the way your using it? If you do the cut from emacs and
then switch to firefox and double click on the url to highlight it,
you will lose your previous selection as it will be replaced by the
double clicking to highlight the url. I've had no problems with cuting
and pasting between emacs and firefox in either direction.
> The scroll wheel works, sometimes, right after the machine
> is booted, but then soon stops working and never works again.
> One way to get it to stop working is to use the KVM switch.
> However, it always works on all the other machines on the
> switch, which are running Windows. Maybe the driver has
> some bug related to being sensitive to something that
> happens when I switch. (I don't even know what happens,
> electrically, if anything, when I use the KVM switch.)
Aha! A vital piece of information. Some KVM boxes, especially cheaper
ones do have some odd side effects with mice - I've had problems with
this and had to reconfigure gpm to get it to work reliably with
X. However, when I moved to a "blackbox" KVM, the problem vanished. I
found this especially with modern USB mice which were PS/2 compatible
and I was running them via a KVM using a USB to PS/2 converter. Don't
know why the KVM should do this as I thought it would just "pass
through" the data with no interference, but cheaper KVMs do seem to
cause problems. I've used two different models of the black box KVM
without any problems.
> Here's what the black magic in my XF86Config-4 file says
> about my mouse (which is a two-button + scroll mouse).
> I wound up writing the whole XF86Config-4 by hand,
> so it's entirely possible I messed this up.
>
> Section "InputDevice"
> Identifier "PS/2 Mouse"
> Driver "mouse"
> Option "Protocol" "IMPS/2"
> Option "Device" "/dev/psaux"
> Option "Emulate3Buttons" "no"
> Option "ZAxisMapping" "4 5"
> EndSection
Can't see anything wrong with this.
> Section "ServerLayout"
> Identifier "XFree86 Configured"
> Screen 0 "Screen0" 0 0
> InputDevice "Keyboard0" "CoreKeyboard"
> InputDevice "PS/2 Mouse" "CorePointer"
> EndSection
>
>
> There's also this junk, and I don't remember
> what any of it is about. Maybe there's
> something fishy in there:
>
> Section "Module"
> Load "dbe"
> Load "dri"
> Load "extmod"
> Load "glx"
> Load "record"
> Load "xtrap"
> Load "speedo"
> Load "type1"
> ####
> Load "GLcore"
> Load "bitmap"
> Load "ddc"
> Load "freetype"
> Load "int10"
> Load "vbe"
> EndSection
These modules are X extensions which are not part of the base X
specification and which can improve performance or add support for
things like GLX graphics etc. Some of the entries are not familiar to
me and not included in my config e.g.
Section "Module"
Load "bitmap"
Load "dbe"
Load "ddc"
Load "dri"
Load "extmod"
Load "freetype"
Load "int10"
Load "record"
Load "speedo"
Load "type1"
Load "vbe"
Load "xtt"
EndSection
> > > 5. I had to manually configure X for the totally generic
> > > video card. And I fucked something up: when Gnome launches,
> > > there's an extra mouse cursor ("X") sitting in the middle
> > > of the screen. It's a turd. I can make it go away
> > > by launching a shell and clicking on its titlebar.
> > > Then I close that shell, because I do everything from Emacs.
> > > (Clicking on the Emacs titlebar doesn't clear the turd.)
> >
> > Again, almost certainly a misconfigured mouse driver for X.
>
> I think the extra mouse cursor may be a bug in Gnome.
> The problem doesn't happen in KDE.
Possibly, but I'd consider it unlikely. Gnome has a very large user
base and bugs tend to get fixed fairly quickly. I certainly didn't see
anything like this when I was running Gnome.
> > Note that recent versions of the X server have a switch
> > which you can use that will get the server to return
> > what it thinks is your hardware and how it should be
> > configured.
>
> There was soms sort of X autoconfiguration utility
> that ran during system installation, but it failed.
>
> > Personally, I hate overly complex window managers.
> > I've got rid of Gnome, KDE etc and now just run sawfish
>
> So the best way to use Linux as a desktop machine that's
> like Windows or MacOS is not to use the desktop software?
> Maybe I should use the Zen window manager.
No, thats not what I said or implied. The "standard" desktop worked
fine for me. I got rid of it because Gnome didn't offer me any
functionality I needed or wanted. So, why have something you don't use
consuming resources? No reason I could think of, so I stopped running
it. Its one of the things I like about X and Linux, you are not locked
into using/running stuff you really don't need - you have the choice,
unlike Windows, where your stuck with what they provide and often
forced to upgrade hardware to support additional functionality you
never wanted and cannot disable.
> I could just go back to twm, I guess.
> But that doesn't really seem like progress.
> YMMV.
Well, I guess that depends on what you call progress. Does Gnome
provide anything you need that you couldn't get under twm? If not,
then how is it progress? If it does, then thats great and probably
justifies sticking with it. I know lots of users who have no problems
with Gnome and they are very happy with it. I ran it for a few years,
but decided I'd rather have the extra memory and disk space for stuff
I really wanted.
> > Firstly, the comment "Why can't Linux just pick ONE and
> > make that really WORK?" reveals what I think is a basic
> > conceptual flaw. Firstly, it implies Linux is something
> > akin to a company like Microsoft or Apple.
>
> I am making a comment on the quality of the net result
> of particular major project developed for Linux under
> the free software theory of the universe, specifically
> noting that the people involved in the effort have
> been unable to come together to produce a high quality
> component that they identify as a critical centerpiece
> and which they prostletyze as being comparable or better
> than the one on the commercial systems.
>
> Instead, it seems that they got into a fight about
> whose broken software was more politically correct.
>
> I don't want a political mission or a hobby,
> and I don't want to debug the operating system.
> I just want the computer to work so I can use
> it to run basic applications that I want.
> If Linux is not up to that, fine.
> Let's just be honest about it, that's all.
I'm not and have never said Linux was some sort of magic solution
which would solve all the worlds problems. I do challenge claims made
by you and others that it is a totally broken and non functional
environment because I've been using it both professionally and at home
for over 15 years very successfully. I also know a large number of
users who are also using it very successfully as well. However, I
would never claim it is right for everyone and have on many occasions
recommended that someone not use Linux based on factors like their
willingness to move to a different paradigm, ability and
requirements. In other posts in this thread, you have pretty much
insinuated I've lied when stating I've had no problems getting my
desktop to work in a fully functioning and reliable way, which makes
me suspect that your one of those users who immediately blame any
problem on bugs, when more often than not its something the user has
done wrong and refuse to accept it when others tell you they have
things functioning fine.
> > All the rest is really just packages collected by
> > the Distro (Debian, Red Hat, SuSe, etc). Therefore, there
> > is no central body to just "pick one and make it work".
>
> X Windows and KDE and Gnome are "little packages"?
The size of the package is irrelevant to the point. Linux is really
just a kernel and a few basic utilities. All the rest is put together
by the body responsible for the distribution and often, the
distribution will modify the packages to either put their own "stamp"
on them or to modify how they are configured or apply special patches
etc. This is why there is such a difference between the various
distributions. The point is, you cannot say things like th eLinux
desktop is broken because there is no such thing as one Linux desktop
or even one version of KDE or Gnome or whatever. You cannot even state
all desktops running under Linux are broken until you have tried them
all out.
> > I'd also argue many of the window managers do work as
> > well and even better than Windows.
>
> Perhaps, but not the ones that are the standard choices
> promoted and offered in the popular distributions.
Again, my experience and the experience of those I work with and many
of my colleagues and friends has been that they do work and work as
well as Windows does. In fact, for my own personal experience, I would
say its only with the most recent releases of Windows that I've found
my Windows desktop runs as reliably as my Linux one has been doing for
15 years. I've never had anything like a BSoD under Linux or had the
whole system grind to a halt while anti-virus software started
scanning or just have things lock up. However, I'm not going around
making claims that the Windows desktop is broken and useuless.
> > How many of us have defined a colour scheme under windows
> > only to find some apps refuse to follow it, or even
> > worse, follow only part of it and you end up with dialog
> > boxes or windows using the same colour for the foreground
> > and background?
>
> I have never tried to change the colors under Windows,
> or Linux for that matter. I don't care about things
> like that. I can believe that it's broken in Windows.
>
> I merely want Linux to correctly autodetect my video card
> and mouse and stuff, like Windows _always_ does.
Yes, windows does support a wider range of hardware - mainly because
many hardware vendors ensurer they have drivers available and because
they work with many vendors directly. Thats one of the side benefits
of representing the lions share of the market - harware vendors are
very keen to ensure their hardware is supported. Linux doesn't have
that luxury and many hardware vendors actually work against Linux by
refusing to provide specifications etc. However, unlike you, I have
had problems with Windows working correctly with some hardware and I
have been forced on occasions to download and install a vendor's
driver to get windows to work correctly with my hardware. I won't even
go into the problems regarding older but fully functional hardware and
XP.
If you are waiting until Linux has the same degree of hardware
detection and support as MS, then I'd suggest you give up now and just
stick with Windows. Linux is unlikely to ever reach that level because
vendors don't see it as a real market and are not interested in
working with individuals to get their hardware supported. There are
some notable exceptions, like nvidia, who actually produce Linux
drivers and more vendors are beginning to see that providing details
of hardware specifications etc is not going to threaten business
etc. But the basic fact remains that drivers and detection of new
hardware depends on volunteers willing to put in the time to do the
real work. While Linux has shown the potential of such an
orgnaisational structure, there will always be a greater lag time
between when hardware is made available and when it is supported under
Linux.
> Well, and it would be nice if the font selection
> features worked better. (They don't work better
> in Windows, by the way; they just don't look so ugly
> there that I feel compelled to mess with them.)
> Other than that, and some other minor hard-to-fathom
> general buggyness in the desktop environment.
> I am fairly happy with Linux. I'm mostly giving
> examples of things that would terribly impact
> your average desktop user.
I think the font issue is really a result of there being very few good
quality fonts which have a free or open source license. You can get
much better fonts for Linux and really improve how it all looks, but
you have to pay for them. I suspect you will also find differences in
what fonts are available between distributions because diffeent
distributions have different manifestos. For example, Red Hat isn't
that worried about licenses, while Debian has a very definite goal to
promote and distribute only software with an appropriate open source
license. This is to such an extent that in fact the XFree86 software
that is now being distributed with Debian is no longer the standard
X.org software because X.org changed its license and is no longer
compatible with the license requirements for inclusion in Debian. Now
Debian is maintaining its own branch which has a more GPL'ish license
than the new X.org license.
> If I start using it as my desktop, though, I will want
> the sound to work. Also the scanner, camera, printer,
> DVD writer, etc. I haven't tried any of that stuff.
With the exception of the camera, I've got all of that working with
minimal effort.
> > I sometimes wonder why it is I hear about people having all
> > these problems, yet I've never really experienced them.
>
> Because in your very limited set of experiences
> you happen to have been very lucky?
> Because your hardware happens to work with the standard drivers?
> Because you're not using the standad desktop environment?
> To me, it's no wonder at all.
> Do you have some more plausible theory?
Possibly its just simply that your in the minority of users and not
the majority position you feel you are in? Yes, I do tend to stick to
fairly standard hardware vendors, but not because of compatibility,
but rather just for reliability (I've found you do tend to get what
you pay for with hardware). However, I don't think my setup is at all
strange or very different. My home desktop was fairly highly spec'd
when I got it because at the time I was doing a lot of scanning and
OCR work and a fair amount of sound related stuff, so I got a
multi-channel sound card (SB Live), an Adaptec Ultra Wide SCSI
controller ATLAS SCSI HD and HP SCSI scanner. All of this hardware
worked just fine without any need to chase down problems, find
additional drivers or do anything else unusual. My desktop at work is
a packaged Dell system and it worked fine with no problems except for
an initial reconfig of the mouse (which I now know was due to the
cheap KVM I wa using). My work laptop is one of those tiny Dell X300
and I expected to have considerable problems getting it to
work. However, it all worked fine with the exception of the wireless
card, but I knew it wouldn't work as it is based on hardware with
closed source specs (though I expect someone will reverse engineer it
before too long). I just got a pmcia wireless card which is supported
and use that.
> > > 7. The "security update notification" widget doesn't work.
> > > It beeps and complains that security upgrades are ready.
> > > Then it launches Synaptic, which doesn't agree.
> > > I am left wondering what critical security patches I am missing.
> >
> > What have you got in your /etc/apt/sources.list file? have you got the
> > site for the security updates as well as the normal source sites?
>
> I have:
>
> deb http://debian.crosslink.net/debian/ testing main
> deb-src http://debian.crosslink.net/debian/ testing main
> deb cdrom:[Debian GNU/Linux testing _Sarge_ - Official Snapshot i386 Binary-2 (20041231)]/ unstable contrib main
> deb cdrom:[Debian GNU/Linux testing _Sarge_ - Official Snapshot i386 Binary-1 (20041231)]/ unstable contrib main
> deb http://security.debian.org/ testing/updates main contrib
> deb http://wine.sourceforge.net/apt/ binary/
>
> I'll probably have to do some kind of total re-install soon.
> Luckily, I'm not using it for anything critical lately.)
> My theory is that possibly something about the status of those
> two random packages is confusing the buggy security update
> icon widget. As I mentioned, I believe this all started on
> the day that Sarge changed over to stable.
Well, I may know what the problem could be. Your sources.list contains
references to both "sarge" and to "testing". Personally, I run testing
as I've found it to be very stable and more up-to-date than Debians
"stable" release (Debian has acknowledged their release cycle is a bit
slow for their stable versions). I would recommend you do the
following
1. Make a backup of /etc/apt/sources.list
2. Edit sources.list and remove the cdrom source entries
3. Remove all entries referring to "sarge"
4. Include just two lines, one for the normal debian distribution and
one for the non-us distribution.
5. Do the following
apt-get update
apt-get dist-upgrade
This should get your system in sync with the current testing version
of Debian (etch). The dist-upgrade will check to make sure all package
dependencies are met.
My sources.list file contains the following -
deb ftp://ftp.at.debian.org/debian-non-US testing/non-US main contrib non-free
deb ftp://ftp.at.debian.org/debian testing main contrib non-free
deb http://commonwerx.org/software/contrib/slime/deb ./
> > There is no such thing as a free lunch. If you don't
> > want to pay for a polished commercial product,
> > be prepared to put in a little extra brain time
> > to figure some of it out.
>
> This is why I do mostly use commercial products.
>
> > For your friend who wants Linux to do some development,
> > but doesn't want to spend time trying to work things out
> > and configure the system, then pay for a commercial
> > Linux distro like Red Hat Workstation - its still
> > relatively cheap and you get support (but check their
> > support matrix to ensure your hardware will be
> > supported).
>
> This is the same as saying, "For your aged grandmother who
> just wants to drive her car to the grocery store, but doesn't
> want to spend time figuring out how the catalytic converter
> works -- and it's going to need some tweaking --- she should
> buy a commercially sold vehicle".
>
> And I fully agree.
>
> > What I really find hard to understand is all the people out there
> > who continually make comparisons with commercial OSs like Windows
> > and OS X to "free" ones like Linux and BSD and then complain about
> > how lousy they are when they don't get the same degree of polish or
> > "out of the box" support.
>
> I'm not complaining. Did I start this thread?
> No. I am merely responding to people who seem to be
> claiming that Linux is as polished as those commercial
> systems, and particularly well suited for your average
> home desktop user.
I agree. I don't beleive Linux is for everyone and it depends on the
person, what they do and their aptitude and interest. However, I do
beleive that Linux can provide a stable reliable and fully functional
alternative to Windows for many people.
> I think RedHat 9 worked better in alot of these respects,
> as I recall. My buddy's recent flirtation with Fedora
> tends to support that recollection. But Debian was touted
> as the stable Linux, and I like its package system better,
> and it seems to get the most attention from the developers
> of Lisp implementations such as CMUCL/SBCL.
I use Debian because I also like its package system and find its
selection of packages better suits my own needs. I've seen a lot of
people get bitten by fedora - its really very "bleeding edge" and a
testbed for RH. In that respect, I'd argue Debian is much more
stable.
> I think Linux is a much better server than Windows.
> Rabid announcements of its superiority in the desktop
> domain are still somewhat premature, however.
> I expect this to greatly improve in the future,
> unless the Linux proponents deny the reality of the
> situation and how far they have to go. In the past,
> there seemed to be some weird antipathy towards improvements
> in these areas. I thought that had passed,
> but lately I am not so sure.
>
I am doubtful Linux will ever represent a very large slice of the
desktop market. Its strength is definately in the server world. I
think a lot of developers are likely to use Linux simply because there
is a lot of stuff out there you can use and try out which is just too
expensive to do under Windows. Personally, I've lost interest in
really following Linux that much. I use it because I find it more
stable and easier than Windows. In fact, I constantly get frustrated
at how hard it is to diagnose problems under windows - there seems to
be a lot of "hidden" knowledge - even worse than Linux and I find the
inconsistencies annoying. One thing I've always like about Unix and
Linux is that it seems far more consistent in how it does things - you
may not like directories full of text files which need to be changed
by hand and find command line programs frustrating, but personally, I
prefer this to (what seems to me to be) obscure little GUI utilities
which get info from data files in odd locations and only work for the
common situation and which log to weird and unexpected
locations. Under Linux, once you know the basic framework, it just
seems more consistent - config files tend to be in /etc somewhere,
logs tend to be in /var, documentation in /usr/share/doc, manpages
devided up into fairly standard sections, boot stuff in /etc/init.d
etc. More than likely, its just what I'm use to, but even co-workers
who adminster Windows admit there is a lot of "acquired" knowledge
which is not immediately accessible
tim
--
Tim Cross
The e-mail address on this message is FALSE (obviously!). My real e-mail is
to a company in Australia called rapttech and my login is tcross - if you
really need to send mail, you should be able to work it out!
Tim X <····@spamto.devnul.com> writes:
>With some mice, you can get complications arising between gpm (the
>mouse driver for the text consoles) and X. When I had some problems
>with an optical mouse, I found the solution was very simple. All I had
>to do was follow advice from the gpm man pages on setting up gpm to
>use 'raw' mode. Check the gpm HOWTOs and google and I'm sure you find
>a straight forward solution. Alternatively, try disabling/turning off
>gpm before starting X.
Ah. I have seen the same with a Dell mouse on Fedora 3. The same setup
works on FreeBSD (both Fedora and FreeBSD use the X.org server)
without problems. Thanks for the hint.
>> 2. Cut and paste do not work in the applications that I use
>> (eg. paste a URL or text between Emacs and Firefox).
>
>I suspect this is also due to a misconfigured mouse. As mentioned,
>either its interference between gpm and the X mouse driver or you have
>not configured the mouse driver correctly in XF86Config. Is the scroll
No, this has nothing to do with the hardware. The problem is, that
there is no standard cut&paste mechanism on X11, so there exist about
a handful or more, with most toolkits implementing more than one.
Traditional X clients (like xterm) afaik use cut buffers, some others
use the CLIPBOARD selection (I guess that's an atom named so in the X
server), and newer toolkits like GTK seem to use yet another method,
called XDND. Emacs seems to put stuff just in cut buffers, and GTK
seems to only support XDND and selections (note that this is guesswork
from observation, I'm not really interested in how this mess works),
so for example you cannot paste text from Emacs into a thunderbird
composer window (or vice versa, sometimes). Xemacs does it a bit
differently, and there cut&paste works better. Presumably it's also
storing selections in the CLIPBOARD selection, or in some other
mechanism, or reading from them. I've found that one solution is to
use the xclipboard(1) program, which comes with X11, as a kind of
indirection, when cut&paste doesn't work at all. In the case of emacs
vs. Mozilla, I tend to blame GTK, since it doesn't seem to support the
"traditional" methods and instead insists on cooking its own stuff.
There's an older "standard" (ICCCM) and a newer one (freedesktop.org)
for resolving such ("inter-client communications") issues, with the
earlier having found wide acceptance, whereas proposals including XDND
on freedesktop.org seem to be restricted to Gnome+KDE atm (but offer
more features).
mkb.
joesb wrote:
> MacOS only run on one architecture, a Mac. MacOS, come preinstalled on
> its vendor's hardware know what hardward they are using, it will be so
> stupid if they are not the one getting it work right the first time.
>
> Windows or Linux doesn't get to choose what hardware it was installed
> in. So yes MacOS could win the "Just Work" test.
> It's like "A car comes in any color you want as long as it's black"
> "MacOS just work pre-installed as long as it's installed on Mac machine
> it was designed for"
Sure, you miss out on choice, but at least I don't have to spend
hours and hours (time is money too) looking for a PC laptop that
is just halfway decent (with all that LOADS of crap out there) and
actually cheaper than a Mac, I don't have to spend time looking
for a Linux/laptop combo that works. No, it just works. If I'm
fine with the machine (I mostly am, happier than with any
PC+software I had so far), I don't need choice.
Maybe the software only runs on a Mac, but it saves me time (and
downtime), and with Cocoa it also saves programming time. I
gladly pay the price (quote from Equilibrium ;) ).
--
Don't let school interfere with your education. -- Mark Twain
"Brandon J. Van Every" <·····················@mycompanyname.com> writes:
> Kirk Job Sluder wrote:
> > But a bit too often, I see complaints about Linux that have very
> > little
> > to do with actually using Linux.
>
> One of my complaints is that Linux adherants usually have a different
> mentality for parcelling application and system functionality than the
> rest of us "mere mortals" who simply expect things to work.
Well, I don't quite agree. Some of us "mere mortals" also work with
stripped-down Microsoft Windows for a variety of purposes. While
Microsoft and Apple like to sell their OS bundled with a mess of
unrelated applications, that does not especially mean that it is wise to
confuse their attempts to construct vertical monopolies with the OS.
And while it certainly would be nice if things simply just worked,
frequently they don't. In the last year, I've run into hardware
compatibility problems with both Windows XP and OS X. On the Windows XP
end, it involved a serial touchpad and a printer for which XP drivers
were not available. And for OS X, it involved a flatbed scanner. I've
had to deal with legacy software that just didn't like Windows XP.
Thankfully, I stopped using XP just before the strongly suggested
service pack killed Endnote for about a month.) And, I've struggled
with gimpy videocard drivers that didn't work under Windows XP.
The moral of the story: Never, ever, EVER expect it to "simply expect
things to work." If you run an OS, it pays to check, and recheck
compatibility every time you buy a new device.
> I'm sure we can agree on these things being "Linux distro"
> problems. You just think this doesn't implicate Linux, and I think it
> does. Furthermore, all the lay people coming from a Windows world (for
> instance, consumers) consider these sorts of things to be "Linux
> problems." People talk about "Windows problems" even though there's
> W2K, XP, Win98SE, ME, and so forth... doesn't make a problem any less
> of a "Windows problem."
Well, that depends. The fact that a program that was not designed with
the Microsoft GUI toolkits will have a different look and feel is an
issue with that particular program. Security upgrades that break the
functionality of other software in unpredictable ways is a problem with
the OS.
> You can try to give people a massive re-education / brainwashing, that
> somehow if you put the label "Linux" on the right components and
> categories, it suddenly doesn't have problems.
Oh, I'm not saying that Linux does not have problems. I don't know
there you get that from my post.
My statement is simply this. A lot of complaints about Linux are really
about things that have not a thing to do with Linux: different UI
standards between desktop environments, spacial nautilus, emacs, YaST,
apt-get, gimp. There are lots of things to complain about there, but
these really are not about Linux.
Now, if you want to argue about the fact that too much hardware out
there requires binary drivers that are unavailable too Linux developers
and distributions. That's a valid complaint about Linux. Actually,
it's a complaint about the fact that hardware vendors are frequently
reluctant to publish details about the protocols they use.
> But I think that's
> both pissing in the wind and a lot of malarky. Consumers aren't
> interested in hearing it, they just want things to work.
I'll tell you what. If you ever find an operating system where things
"just work." Let me know. I'd be overjoyed, delighted, thrilled to an
extreme degree. But after too many years using computer systems, and
too much time pounding my head against a wall over things that should
"just work" but don't, I'm not holding my breath.
> It remains to be seen if any Linux distro will rise above its peers,
> provide things that just work, and capture mindshare as "the reliable
> Linux."
Linux actually has quite proven reliability. This is one of the reasons
I run it. (The fact that the software *I* need to do my work is gimped
on Windows XP is another issue.) The fact that you can't just run to
Best Buy, grab any device off the shelf and expect it to work doesn't
speak to reliability.
I also don't care that much about what "consumers" are interested in
hearing. I use Linux when it is the best choice for what I need to do,
and I use Windows XP when it is the best choice for what I need to do.
In both cases, I accept the fact that I'll have to deal with some
headaches now and then.
If Windows XP is the best choice for what you need to do, go forth, be
happy, and make beautiful things using it. Just don't try the,
"malarky" that Windows XP "just works."
--
Kirk Job-Sluder
"The square-jawed homunculi of Tommy Hilfinger ads make every day an
existential holocaust." --Scary Go Round
Kirk Job Sluder wrote:
> I'll tell you what. If you ever find an operating system where things
> "just work." Let me know. I'd be overjoyed, delighted, thrilled to an
> extreme degree. But after too many years using computer systems, and
> too much time pounding my head against a wall over things that should
> "just work" but don't, I'm not holding my breath.
If you mean by that "getting weird hardware to work", then, sure!,
you have to ask the hardware vendor if they support a certain
system! My definition of just works is if I can get a decent
machine (maybe have some amount of choice, too), and get an OS to
run on that, and support the machine with its full capabilities,
including video-out (just yesterday I saw someone get his SuSE
laptop to EITHER display stuff on screen OR on the VGA output to
project), sleep mode etc. So far that often doesn't include Linux.
> If Windows XP is the best choice for what you need to do, go forth, be
> happy, and make beautiful things using it. Just don't try the,
> "malarky" that Windows XP "just works."
Well, its usability sucks, compared to Mac or Unix, but with most
hardware it just works, so on a PC, unfortunately, I would have to
use Windows, not Linux, that's why I have a Mac now... (I want Unix).
--
Don't let school interfere with your education. -- Mark Twain
Ulrich Hobelmann <···········@web.de> writes:
> If you mean by that "getting weird hardware to work", then, sure!, you
> have to ask the hardware vendor if they support a certain system!
Perhaps it's just me, but I don't consider inkjet printers and touchpads
that to be especially exotic. With the inkjet printer, if you didn't
have the Windows 98 driver CD that came with that particular OEM system,
you were out of luck, and I found I was faced with an interesting choice
of either a touchpad or mouse but not both.
Again, not really exotic. Just a fairly typical practice that support
for the required drivers vanished along with the product. It wouldn't
surprise me to have to deal with similar problems with Longhorn when it
comes out.
> My
> definition of just works is if I can get a decent machine (maybe have
> some amount of choice, too), and get an OS to run on that, and support
> the machine with its full capabilities, including video-out (just
> yesterday I saw someone get his SuSE laptop to EITHER display stuff on
> screen OR on the VGA output to project), sleep mode etc. So far that
> often doesn't include Linux.
I've actually found laptop support to be a problem all over. I have an
older Dell sitting idle on my desk that has never quite worked right
with Windows XP. Never mind the fact that Windows XP runs pretty darn
slow on otherwise reasonable older hardware, but between sound card
issues, lockups, and crashes, it wasn't "just working." It runs linux
reasonably well though. Or at least well enough that I can actually get
some work done on it at the coffee shop.
I would say that it's a matter of picking the annoyances you can live
with. I can live with ordering hardware from a list of supported
devices. It only happens once every other year after all.
> --
> Don't let school interfere with your education. -- Mark Twain
--
Kirk Job-Sluder
"The square-jawed homunculi of Tommy Hilfinger ads make every day an
existential holocaust." --Scary Go Round
Brandon J. Van Every wrote:
> I've also become deeply averse to large stacks of dependent packages.
Sofware A uses library X, Y, Z, so, when we install software A, apt-get or
the similar application installs A, X, Y, Z automatically, then we install
software B, which uses X, Y and W, so apt-get installs B and W. How is that
worst than downloading a big executable file that we run with top
privileges on our operating system and installs some kind of weird glued
AXYZ and then we do the same for B which installs some kind of weird glued
BXYW ?
An example ? Dreamweaver comes with its own buggy implementation of ssh
(secure shell) used to upload web sites thru scp, which is by no way the
same used by puty, a ssh client. Resulting is duplication of software,
effort and the fact that although Puty works quite well, Dreamweaver's
implementation is buggy and it's not going to get fixed because Macromedia
people doesn't care (if you don't care, better than shipping a crapy
implementation is depending on an existing implementation done by someone
else). Of course, Windows doesn't provide anything to manage dependencies
(it's like Linux From Scratch in that sense) so software vendors are forced
to do that.
Honestly, I prefeer any package manager like Debian's apt-get, Gentoo's
portage or any of the BSD's package managers.
--
Pupeno <······@pupeno.com> (http://pupeno.com)
Reading ? Science Fiction ? http://sfreaders.com.ar
Pupeno wrote:
> An example ? Dreamweaver comes with its own buggy implementation of ssh
> (secure shell) used to upload web sites thru scp, which is by no way the
> same used by puty, a ssh client. Resulting is duplication of software,
> effort and the fact that although Puty works quite well, Dreamweaver's
> implementation is buggy and it's not going to get fixed because Macromedia
> people doesn't care (if you don't care, better than shipping a crapy
> implementation is depending on an existing implementation done by someone
> else).
I wouldn't be shocked if a Windows-centric company did a half-assed job
on Unix. Doesn't mean the same product has such problems on Windows.
> Of course, Windows doesn't provide anything to manage dependencies
> (it's like Linux From Scratch in that sense) so software vendors are forced
> to do that.
And if they make one huge package, and test the entire build themselves,
then you're shielded from many problems. Dozens of independent packages
means dozens of things that haven't necessarily been tested together and
can all represent points of failure.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
I won't spend more than 1 day configuring 1 thing.
"Brandon J. Van Every" <·····················@mycompanyname.com> writes:
> Pupeno wrote:
> > An example ? Dreamweaver comes with its own buggy implementation of ssh
> > (secure shell) used to upload web sites thru scp, which is by no way the
> > same used by puty, a ssh client. Resulting is duplication of software,
> > effort and the fact that although Puty works quite well, Dreamweaver's
> > implementation is buggy and it's not going to get fixed because Macromedia
> > people doesn't care (if you don't care, better than shipping a crapy
> > implementation is depending on an existing implementation done by someone
> > else).
>
> I wouldn't be shocked if a Windows-centric company did a half-assed
> job on Unix. Doesn't mean the same product has such problems on
> Windows.
Macromedia has not become windows-centric over the last several years?
> And if they make one huge package, and test the entire build
> themselves, then you're shielded from many problems. Dozens of
> independent packages means dozens of things that haven't necessarily
> been tested together and can all represent points of failure.
Pardon if this just does not make sense. Of course most vendors for
Windows don't make one self-contained package either. Most Windows
software depends on linking to Windows libraries for GUI widgets,
networking, user information and cryptography. While this is generally
a good thing, the Windows XP service pack ended up changing the function
of those libraries in ways that had unpredictable results on third-party
software. The other side of the story was that you couldn't *not*
install SP1 in a network environment, due to some rather nasty security
vulnerabilities.
Now in an FOSS world, you can either rebuild from source or, as is
frequently the case, someone else will have done it for you and uploaded
patched packages for you. However, I watched colleagues become
extremely frustrated wondering if updates to their much-needed software,
and sometimes very expensive software, would be released before SP1
became mandatory within our organization.
The problem of package conflicts in MS Windows land has become so
widespread that there is even jargon to describe it: "dll hell."
There is actually good money to be made fixing these problems for
users.
> --
> Cheers, www.indiegamedesign.com
> Brandon Van Every Seattle, WA
>
> I won't spend more than 1 day configuring 1 thing.
--
Kirk Job-Sluder
"The square-jawed homunculi of Tommy Hilfinger ads make every day an
existential holocaust." --Scary Go Round
Kirk Job Sluder wrote:
> "Brandon J. Van Every" <·····················@mycompanyname.com> writes:
>
>
>>Pupeno wrote:
>>
>>>An example ? Dreamweaver comes with its own buggy implementation of ssh
>>>(secure shell) used to upload web sites thru scp, which is by no way the
>>>same used by puty, a ssh client. Resulting is duplication of software,
>>>effort and the fact that although Puty works quite well, Dreamweaver's
>>>implementation is buggy and it's not going to get fixed because Macromedia
>>>people doesn't care (if you don't care, better than shipping a crapy
>>>implementation is depending on an existing implementation done by someone
>>>else).
>>
>>I wouldn't be shocked if a Windows-centric company did a half-assed
>>job on Unix. Doesn't mean the same product has such problems on
>>Windows.
>
>
> Macromedia has not become windows-centric over the last several years?
I thought with all the terms you were using above, you were describing a
Unix version of Macromedia. If not, my mistake.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
I won't spend more than 1 day configuring 1 thing.
From: Christopher C. Stacy
Subject: Re: I've thought better of Linux
Date:
Message-ID: <uk6ku2evf.fsf@news.dtpq.com>
Kirk Job Sluder <····@jobsluder.net> writes:
> the Windows XP service pack ended up changing the function
> of those libraries in ways that had unpredictable results
> on third-party software.
> The problem of package conflicts in MS Windows land has become so
> widespread that there is even jargon to describe it: "dll hell."
I believe that "dll Hell" refers to a slightly different
specific issue, which was supposed to be solved in Windows XP.
That is, many different vendors loading different versions
of a given DLL, with incompatabilities, and then having some
conflict or management problem.
The problem you are describing is different: Microsoft
just plain broke some DLLs (and the apps which used it).
In Linux, this would be as if the existing versions of some
shared libraries were declared unusable due to security problems,
and applications were forbidden to dynamically link to them,
and then the replacement ".so" files were incompatible.
> So, I'm nukeing my Linux partition and not looking back. Good riddance.
You will not regret it. Unix is harmful to Lisp. Those who don't understand
it are not real lispers or have been reading too much slashdot.
lbm wrote:
>>So, I'm nukeing my Linux partition and not looking back. Good riddance.
>
>
> You will not regret it. Unix is harmful to Lisp. Those who don't understand
> it are not real lispers or have been reading too much slashdot.
As is Cocoa. I'm already turning to the dark (OO) side again...
And seriously, the Unix way has a lot to offer ;)
--
Don't let school interfere with your education. -- Mark Twain
On Sun, 12 Jun 2005 16:43:48 +0200, Lars Rune Nøstdal wrote:
> On Sun, 12 Jun 2005 02:11:25 -0700, Brandon J. Van Every wrote:
>
>> blablabla
>
> you suck.
>
> (haha)
ups .. did i say this?
hm, well after using Linux for 4 years, i can say you are wronk - ups
wrong .. and you also suck - ups, i said it again, sorry..
my grandma also uses Linux, she does not know nor care how to adjust the
clock on the vcr .. she does not seek the "outer limits", so she will
never face "problems" and never gain any significant knowledge or
experience either (but you are different here, of course)
if i was your bitch, would you really be happy? i can be your new OS, you
can call me "Bitch OS" or "LiBitchIx" .. you can code me like this:
(let ((my-very-own-game (make-instance 'doom :version 4 :author "BJVE"))
(run my-very-own-game))
ah, all is good in la-la-la-land .. never any problems, and i'm also rich
and successfull
everything and everyone rotates around me and i can form physical matter
in thin air; i'm a god like you
i never see nor need to solve my own problems, i only need to state
problems other should fix for me.. this is usefull for me and others, this
is nice - but you suck .. you are boring .. you never bicycle uphill nor
downhill .. you are stuck, and you suck - do you hear me?
(haha)
--
mvh,
Lars Rune Nøstdal
http://lars.nostdal.org/
Lars Rune N�stdal wrote:
> my grandma also uses Linux, she does not know nor care how to adjust the
> clock on the vcr .. she does not seek the "outer limits", so she will
> never face "problems" and never gain any significant knowledge or
> experience either (but you are different here, of course)
Hell, I used Linux for maybe two years, then Free & NetBSD for
another two, now a Mac. Still, Linux sucks in a way. My main
reason for getting a Mac was that I couldn't find a laptop (or
information about one), that decently supported Linux, as all the
information sites are -- let's say -- very conservative about the
amount of information they give you. I have (somewhat bad)
experience with two laptops and Unix already, so the Mac was the
obvious choice.
Bitching about Linux often has a reason, believe it or not.
--
Don't let school interfere with your education. -- Mark Twain
> Lars Rune N�stdal wrote:
>
>> my grandma also uses Linux, she does not know nor care how to adjust the
>> clock on the vcr .. she does not seek the "outer limits", so she will
>> never face "problems" and never gain any significant knowledge or
>> experience either (but you are different here, of course)
I believe Linux can run fine as a consumer OS if it comes preinstalled
with known HW and one is exceedingly conservative about adding HW.
Converting from previously bought Windows HW is a gamble that can easily
result in a nightmare. In the sense of tech support, Linux is quite
fragile. This reality will keep Linux from being of any interest to
consumers in general. Well, excepting perhaps poor countries. I'm not
excited about the Chinese intellectual property market though. :-)
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
"The pioneer is the one with the arrows in his back."
- anonymous entrepreneur
"Brandon J. Van Every" <·····················@mycompanyname.com> writes:
> I believe Linux can run fine as a consumer OS if it comes preinstalled
> with known HW and one is exceedingly conservative about adding HW.
Which is exactly the same case with MacOSX -- Apple takes great care
to produce only a very small palette of hardware underwhich to run
MacOSX, or even MS-Windows -- the only difference being that Microsoft
assigns a lot more resources to this problem and hardware vendors help
it more.
--
__Pascal Bourguignon__ http://www.informatimago.com/
-----BEGIN GEEK CODE BLOCK-----
Version: 3.12
GCS d? s++:++ a+ C+++ UL++++ P--- L+++ E+++ W++ N+++ o-- K- w---
O- M++ V PS PE++ Y++ PGP t+ 5+ X++ R !tv b+++ DI++++ D++
G e+++ h+ r-- z?
------END GEEK CODE BLOCK------
"Brandon J. Van Every" wrote:
[...]
> I believe Linux can run fine as a consumer OS if it comes preinstalled
> with known HW and one is exceedingly conservative about adding HW.
How many hours Linux usage do you have? Have you:
a) Installed any Linux distribution
b) Googled for assistance
c) Asked a group other than c.l.l to help?
> Converting from previously bought Windows HW is a gamble that can easily
> result in a nightmare. In the sense of tech support, Linux is quite
> fragile. This reality will keep Linux from being of any interest to
> consumers in general. Well, excepting perhaps poor countries. I'm not
> excited about the Chinese intellectual property market though. :-)
Fragile? No. Other than hardware failures and kernel rebuilds (think once a
year), my Linux boxes do not fail. My Windows boxes die every other day.
Chinese IP? Well I'd be pretty excited about it if I knew anything about
the speed of economic and academic development in China (yes, yes, I know
it's a damn police state).
>
> "The pioneer is the one with the arrows in his back."
> - anonymous entrepreneur
Does he become a pioneer at the moment the arrows hit? Or before?
Damn this natural language. I've wasted *decades* of research time of this.
I'm going to learn Lisp^H^H^H^Hto keep my mouth shut.
Jamie
Jamie Border wrote:
> a) Installed any Linux distribution
> b) Googled for assistance
> c) Asked a group other than c.l.l to help?
he is asking for the best OS and programming language at least since 1996:
http://groups.google.de/group/comp.os.linux.development.system/browse_frm/thread/08b756d01869223a/fbb8703dba732859
Looks like we should heed this advice:
.:\:/:.
+--------------------+ .:\:\:/:/:.
| PLEASE DO NOT | :.:\:\:/:/:.:
| FEED THE TROLLS | :=.' - - '.=:
| | '=(\ 9 9 /)='
| Thank you, | ( (_) )
| The Management | /`-vvv-'\
+--------------------+ / \
| | @@@ / /|,,,,,|\ \
| | @@@ /_// /^\ \\_\
@·@@·@ | | |/ WW( ( ) )WW
\||||/ | | \| __\,,\ /,,/__
\||/ |jgs| | (______Y______)
/\/\/\/\/\/\/\/\//\/\\/\/\/\/\�/\/\/\/\/\\/\/\/\/\/\/\/\/\/\/�\/\/\
--
Frank Bu�, ··@frank-buss.de
http://www.frank-buss.de, http://www.it4-systems.de
Frank Buss wrote:
> he is asking for the best OS and programming language at least since 1996:
>
>
http://groups.google.de/group/comp.os.linux.development.system/browse_frm/thread/08b756d01869223a/fbb8703dba732859
>
Poor Brandon. Someone should tell him about Visual Basic.
Frank Buss wrote:
> Jamie Border wrote:
>
> > a) Installed any Linux distribution
> > b) Googled for assistance
> > c) Asked a group other than c.l.l to help?
>
> he is asking for the best OS and programming language at least since 1996:
>
> http://groups.google.de/group/comp.os.linux.development.system/browse_frm/thread/08b756d01869223a/fbb8703dba732859
>
> Looks like we should heed this advice:
I do not think he can be attributed "troll". Troll would mean something
different I guess.
Surely, I am also convinced in the meantime that he isn't after a
solution with his game business. His behavior can easily be nailed
down: the American dream doesn't work for him.
However, I have never understood why people in America want to make the
rise from "rags to the riches" when there is the strong believe in
bible words. Why is it important to get rich in America. How insane
must a person be in order to believe that he counts more in front of
God when he is a rich man. Doesn't tell you bible that being rich is
all to your disadvantage in Lord's eyes?
Call me troll for my writings above.
By the way Brandon: you should definitely invest time in learning any
of the following languages: Clean (the Clean book is really excellent),
OCaml, CommonLisp, Scheme, etc.
Even for the case you will not use any of the languages you will get a
sound understanding of some important techniques and concepts.
Schneewittchen
F�rster vom Silberwald wrote:
>
> However, I have never understood why people in America want to make the
> rise from "rags to the riches" when there is the strong believe in
> bible words. Why is it important to get rich in America. How insane
> must a person be in order to believe that he counts more in front of
> God when he is a rich man. Doesn't tell you bible that being rich is
> all to your disadvantage in Lord's eyes?
FWIW I'm an agnostic that's all but atheist. I just can't answer the
fundamental questions, like why anything exists at all, so I leave my
mind a bit open. But if there is a God, or Gods, I don't see any
evidence of them intervening in human affairs. I'm doubting that beings
of such complexity can actually relate to us. When did you last spend
quality time with an anthill? So, the ratings package of a God means
nothing to me.
Meanwhile, I see every evidence in human history of God being an
argument from authority, to coerce obedience. Many wars fought over it.
I don't think the abuse of poor man aphorisms was Jesus' fault, I
think he was genuine. But certainly, the people primarily responsible
for repeating such messages, the Popes and so forth, were the richest
people in the world for a very long time. If you have swallowed this
alpha monkey flimflam I feel sorry for you. It is very likely that
people simply live on this Earth, take what resources they can, and then
just die. Up to you to decide how much you want to take, from where,
from whom, for what.
I'm an alpha male and am not going to be someone else's underpaid lackey
for the rest of my life. My career goal is to become exceedingly
wealthy doing exactly what I want to do. Looks like I won't be reaching
it for another 5..10 years though.
> By the way Brandon: you should definitely invest time in learning any
> of the following languages: Clean (the Clean book is really excellent),
> OCaml, CommonLisp, Scheme, etc.
Do you mean learning or mastering? I've learned something about all of
them. I haven't mastered any of them, as I haven't been able to settle
on a toolbase. I'm resigned to MinGW now though. It's often a PITA,
but Linux is actually worse, and I don't want Cygwin licensing issues to
screw with my builds. The current guinea pig looks like GCL + Jabberwocky.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
"witch-hunt" - (noun) (Date: 1885)
1: a searching out for persecution of persons accused
of witchcraft
2: the searching out and deliberate harassment of
those (as political opponents) with unpopular views
- witch-hunter (noun)
- witch-hunting (noun or adjective)
Brandon J. Van Every wrote:
> I'm an alpha male and am not going to be someone else's underpaid lackey
> for the rest of my life. My career goal is to become exceedingly
> wealthy doing exactly what I want to do. Looks like I won't be reaching
> it for another 5..10 years though.
I simply mentioned the relation between bible, Americans, and being
rich out of the following reason: I have a colleague at NASA. Every
time I am going to meet him and asking him whether he is fine and how
things go he reacts seriously with the excuse that he isn't rich.
However, he concedes he is well-off (own house in America, two cars,
etc..).
I think he will never fall into the hole of underpaid jobs since he is
a upper class mathematician in his field.
But it is interesting: he comes from Russia and must feel a looser
because, though, he will always have a well-paid job due to his
excellence but he will never become exceedingly rich.
My goal: a somehow well paid job but my wish is not becoming a rich
man. Surely, I am somehow worried whether I will stay in science after
my PhD (forthcomming fall). After my Master in astrophysics I had an
interview with a job hunter. After two hours this particular job hunter
convinced me that I should go back to university and applying for a PhD
position because he is deeply convinced that I am not prepared/valuable
for business worlds since I am not greedy for money enough (I had no
career plan because such things are not important for me).
> Do you mean learning or mastering? I've learned something about all of
> them. I haven't mastered any of them, as I haven't been able to settle
> on a toolbase. I'm resigned to MinGW now though. It's often a PITA,
> but Linux is actually worse, and I don't want Cygwin licensing issues to
> screw with my builds. The current guinea pig looks like GCL + Jabberwocky.
I cannot comment on Windows. But I am deeply convinced that Windows is
not ready for the desktop. I do not understand people why they believe
that Windows makes you feel comfortable.
Surely, there is the thing that installing software often becomes a
nigthmare on Linux. And a lot of Linuxers are on crap and drugs.
Nevertheless, why not consulting a Common Lisp tutorial over the
weekend? Corman Lisp runs on your Windows - right? If I were you I
would dive into your game project as quickly as possible. Believe me
things will turn out positively for you; but you must start working on
it. I mean you are not IBM were you can put your cart before the horse:
business plan, a vague strategy, and actually your project. And if it
succeeds - okay; and if it does not succeed - bad luck; IMB has a lot
of money to waste.
You start out alone: and first must come your project and not a
detailed business plan.
Schneewittchen
F�rster vom Silberwald wrote:
> However, I have never understood why people in America want to make the
> rise from "rags to the riches" when there is the strong believe in
> bible words. Why is it important to get rich in America. How insane
It's not the USA, as far as I see. You think getting rich doesn't
count if you're in Europe?
> must a person be in order to believe that he counts more in front of
> God when he is a rich man. Doesn't tell you bible that being rich is
> all to your disadvantage in Lord's eyes?
That's right, my opinion. But a good part of US citizens aren't
Christians, thank God ;)
(and those who are are often very nice people, even though they
don't strictly follow logic)
--
Don't let school interfere with your education. -- Mark Twain
"Brandon J. Van Every" <·····················@mycompanyname.com> writes:
>> Lars Rune N�stdal wrote:
>>
>>> my grandma also uses Linux, she does not know nor care how to adjust the
>>> clock on the vcr .. she does not seek the "outer limits", so she will
>>> never face "problems" and never gain any significant knowledge or
>>> experience either (but you are different here, of course)
>
> I believe Linux can run fine as a consumer OS if it comes preinstalled
> with known HW and one is exceedingly conservative about adding
> HW. Converting from previously bought Windows HW is a gamble that can
And where the fun of installing Linux would be?
There is a big problem, and it isn't about Linux at all, but about
new users actitude (i am not saying you gur-UNIX can't have problems,
but it is of course less frecuente in this respect). They want to get
everything, and i say it, _eveything_ working out of the box in the very same
day they started using Linux for the first time in their lives.
Something, that as we can see, it isn't possible most of the time,
and the newbie ends up frustrated after a few hours of burning his/her
Ubuntu CD.
Ive been using Linux for a few years now (around 5-6) every day,
and i remenber when i first installed it, i barely could get X
and monitor to work properly in an old Slackware version,
i won't go into every detail of my journey here, but my point is,
that i couldn't get an usable system (all working nice) after
a couple of months of searching, googling, reading, tweaking stuff and asking.
It _was_ painful, yes, and i was frustrated at times, but at the end,
everytime i was able to get something working, that was a real good personal
satisfaction and suddenly i had acquired interesting knowledge that i could
use to solve a dozens of more similar problems (a bunch of basic problems are pretty
similar i realized very soon).
So definetly, you can say you can't afford to lose time in your
valuable work doing your own researching, well, i did it that way,
and i don't regret it because now i can make things better and faster
than i used to. And not only that!, i also can learn (everyday yet) new stuff
and cooperate with other people. And i don't need to take a certified program,
i just start using my linux box.
Simple moral, if you are a new user, be patient. If you aren't, then use Windows.
Regards.
Luis Araujo wrote:
>
> And where the fun of installing Linux would be?
>
> There is a big problem, and it isn't about Linux at all, but about
> new users actitude (i am not saying you gur-UNIX can't have problems,
> but it is of course less frecuente in this respect). They want to get
> everything, and i say it, _eveything_ working out of the box in the very same
> day they started using Linux for the first time in their lives.
> Something, that as we can see, it isn't possible most of the time,
> and the newbie ends up frustrated after a few hours of burning his/her
> Ubuntu CD.
For the record, I'm not remotely close to being a Linux newbie. I got
my brown belt in it, if not my black belt, back when kernel 0.99 was in
play. I'm a "recentbie." As for pleasing newbies via "it just works,"
Knoppix attempts to do this. And if one wins the gamble with one's
particular HW, it succeeds.
> Ive been using Linux for a few years now (around 5-6) every day,
> and i remenber when i first installed it, i barely could get X
> and monitor to work properly in an old Slackware version,
> i won't go into every detail of my journey here, but my point is,
> that i couldn't get an usable system (all working nice) after
> a couple of months of searching, googling, reading, tweaking stuff and asking.
> It _was_ painful, yes, and i was frustrated at times, but at the end,
> everytime i was able to get something working, that was a real good personal
> satisfaction and suddenly i had acquired interesting knowledge that i could
> use to solve a dozens of more similar problems (a bunch of basic problems are pretty
> similar i realized very soon).
The problem is, this only has inherent appeal when you're young and
don't know very much about working technology systems yet. Once you've
done this thousands of times on many platforms over a decade or more,
you don't care anymore. You just want your stuff to work, so that you
can move on to *REAL* computing problems instead of boring support
gruntwork. There is no virtue in tedium. It's just not tedious to
certain people yet. I thought Linux was fun for about 3 years. Then
all the configuring got old. One of my biggest career regrets, is that
I didn't take a part-time job getting paid to do all of that stuff, when
it was still fun to me. But hey, I was figuring out a lot more things
than just Linux.
> And not only that!, i also can learn (everyday yet) new stuff
> and cooperate with other people.
Cooperating with other people gets old too. Figured out where you are
on the ultimate software foodchain yet? If you're not on the top,
forcing other people to deal with your code, then you want to be free of it.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
Brandon's Law (after Godwin's Law):
"As a Usenet discussion grows longer, the probability of
a person being called a troll approaches one RAPIDLY."
"Brandon J. Van Every" wrote:
> Cooperating with other people gets old too. Figured out where you are
You've been plaguing c.l.l, asking for their help and cooperation in your
'evaluation' of Lisp.
> on the ultimate software foodchain yet? If you're not on the top, forcing
> other people to deal with your code, then you want to be free of it.
I would suggest you turn your hand to something that requires less human
interaction than software development then :-)
> Brandon's Law (after Godwin's Law):
> "As a Usenet discussion grows longer, the probability of
> a person being called a troll approaches one RAPIDLY."
No. The probability of the OP ranting and trolling approaches 1 as the
discussion continues to go against his personal ideology.
Brandon, why don't you chill out and start at least listening to the (good)
advice you have been given.
If you start writing stuff with Lisp, and come across problems, people here
_will_ try to help you.
I mean, I only started learning Lisp about 6 weeks ago. With the help of a
few people here, I've managed to do some good stuff (COBOL -> Lisp/C/C#/VB
translator for declarations - useful if you've ever had to send a huge bunch
or REDEFINEd data back and forth to a machine running a COBOL listener
process).
Common Lisp _is_ flawed, at least in my eyes. But then so is everything
else. For me, it's a question of finding the least evil language for the
task, and CL is less evil than most others for most tasks.
So my challenge to you would be this:
1) Set yourself a problem that would be difficult to solve in (for example)
C#.
2) Solve the problem with Lisp
3) Now try to implement the *same* program in C#.
Substitute C# for any other language you want. The point of this is to make
you think about how you think about writing software.
If you don't like Lisp after that, either you've chosen a poor problem, or
you know that it's time to move on.
Of course, you should feel free to do what many others have done - develop a
language that works exacly how you want.
Jamie
From: =?ISO-8859-15?Q?Andr=E9_Thieme?=
Subject: Re: I've thought better of Linux
Date:
Message-ID: <d8qk5q$f7k$1@ulric.tng.de>
Jamie Border schrieb:
> So my challenge to you would be this:
>
> 1) Set yourself a problem that would be difficult to solve in (for example)
> C#.
> 2) Solve the problem with Lisp
> 3) Now try to implement the *same* program in C#.
Yes, sounds like a fair challenge... (not!).
Why not think of a problem that is hard to solve in Lisp and easy in C#?
First do it in C# and then solve the same problem in Lisp.
Andr�
--
"Andr� Thieme" wrote:
>
> Yes, sounds like a fair challenge... (not!).
Wasn't supposed to be a programming contest, but a challenge in the sense of
"can *I* do this?"
> Why not think of a problem that is hard to solve in Lisp and easy in C#?
Well how would he do that? He doesn't know Lisp yet... I suppose he could
choose some involved CLR interaction. But that's not a very fair contest
either :-)
> First do it in C# and then solve the same problem in Lisp.
OK, but I was trying to show that many of the inflexibilities of C# could be
avoided with Lisp (mostly), and finding an ugly C# problem that could be
better represented (logically) in Lisp code would do that nicely, I thought.
That is the carrot that has kept me plugging away at Lisp for a while now.
I can read my programs and understand them.
Jamie
> Andr�
> --
Luis Araujo wrote:
> Ive been using Linux for a few years now (around 5-6) every day,
> and i remenber when i first installed it, i barely could get X
> and monitor to work properly in an old Slackware version,
> i won't go into every detail of my journey here, but my point is,
> that i couldn't get an usable system (all working nice) after
> a couple of months of searching, googling, reading, tweaking stuff and asking.
> It _was_ painful, yes, and i was frustrated at times, but at the end,
> everytime i was able to get something working, that was a real good personal
> satisfaction and suddenly i had acquired interesting knowledge that i could
> use to solve a dozens of more similar problems (a bunch of basic problems are pretty
> similar i realized very soon).
Well, when I first tried SuSE, it was the worst crap ever and I
had to manually put a symlink so it would find the XF86 config
file! (6.2)
Debian 2.1 worked like a charm, out of the box, and I only got my
Windows 98 PC a year before that, so I was a relative PC newbie
(but learned to loathe Win98 crashes enough already).
Later I tried RedHat (6.2 and more recently 9.0) and Mandrake,
which seemed to have been infected by the same disease. Most of
their stuff just didn't work, like connecting to the internet from
Gnome (the modem even dialed, and the GUI used the same
commandsline tool I used on Debian, but nothing happened), getting
the screen working properly (instead of putting the whole picture
the the right of the CRT's center). Debian just did what I told
it to. I simply can't understand why people say that Debian is
hard for newbies, just because it's text based, and RedHat is easy
and nice (or Mandrake, Fedora these days), because it has a GUI.
Often the GUI is just broken, and the tools don't work.
The point is that some Linuxes GIVE you success when you work
hard, but some are just utter crap -- much worse than ANY Windows
was or will ever be.
> Simple moral, if you are a new user, be patient. If you aren't, then use Windows.
No. I dropped Windows back then, because I wasn't patient enough
to operate dull, repetitive "next" dialogues when configuring
ANYthing, or to tolerate constant crashes or reboots. Today I
would get severely pissed off by trojan horses, viruses, spyware
etc. what my friends deal with.
(oh, and for Linux, be patient and get a decent distro; last time
I tried Debian testing, it didn't work anymore, something was
broken even in good ole Deb, but Gentoo seemed to be a nice one;
or even better, get some BSD, they tend not to crap too much on you)
--
Don't let school interfere with your education. -- Mark Twain
Ulrich Hobelmann <···········@web.de> writes:
> The point is that some Linuxes GIVE you success when you work hard,
> but some are just utter crap -- much worse than ANY Windows was or
> will ever be.
>
>> Simple moral, if you are a new user, be patient. If you aren't, then use Windows.
>
> No. I dropped Windows back then, because I wasn't patient enough to
> operate dull, repetitive "next" dialogues when configuring ANYthing,
> or to tolerate constant crashes or reboots. Today I would get
> severely pissed off by trojan horses, viruses, spyware etc. what my
> friends deal with.
My point was that you need to be more patient as a
GNU/Linux newbie than as a windows newbie.
And as i said in my previous posting, after you got a few months
solving problem, you find that most of the basic problems are solved
in similar ways (but it takes time for a newbie to realize of it),
so you can end up working faster and better in the long term for sure.
Installing Linux is a crapshoot. Just because a distribution works
with your particular hardware doesn't mean it'll work on mine. My
experiences with Gentoo, Debian, and FreeBSD were quite different than
yours, and I was ready to proclaim Ubuntu the silver bullet based on my
experiences, but that's the distro that the OP is complaining about.
On Sun, 12 Jun 2005 16:42:05 +0200, Ulrich Hobelmann wrote:
> Lars Rune Nøstdal wrote:
>> my grandma also uses Linux, she does not know nor care how to adjust the
>> clock on the vcr .. she does not seek the "outer limits", so she will
>> never face "problems" and never gain any significant knowledge or
>> experience either (but you are different here, of course)
>
> Hell, I used Linux for maybe two years, then Free & NetBSD for
> another two, now a Mac. Still, Linux sucks in a way. My main
> reason for getting a Mac was that I couldn't find a laptop (or
> information about one), that decently supported Linux, as all the
> information sites are -- let's say -- very conservative about the
> amount of information they give you. I have (somewhat bad)
> experience with two laptops and Unix already, so the Mac was the
> obvious choice.
>
> Bitching about Linux often has a reason, believe it or not.
believe it or not, BJVE would bitch about Mac too - trust me, you'll see
in the mean (and past) time, what has happened? has he made a difference?
what point is reasons and opinions if nothing ever happens? has he
improved something or himself, be it Linux, Mac, C++, Lisp, Windows or
whatever? he'll never stop bitching
i'm using this Linux-thing on all my servers, all my desktop-computers and
laptops; i'm happy with both the things that work and does not work ..
things just "are", and this is the same with any OS or whatever
hey, everything sucks if you look hard enough and focus only on the
bad things or problems.. freaking moron pushing the same button over and
over and whine because something bad happens, or might happen; he hasn't
even really tried
there is also a set of buttons in his head
i do not give a shit what you or anyone else uses, but i'd never say "Mac
sucks" when there are thousands that happily use it - and it's the same
with anything else; it all depends on what people are after
BJVE does not know what he is after - he thinks he wants to do something,
but only for riches and fame, not for the sake of doing it .. this does
not work (it's the button being pressed over and over)
and what about us fools having to read this shit?
--
mvh,
Lars Rune Nøstdal
http://lars.nostdal.org/
On Sun, 12 Jun 2005 19:18:17 +0200, Lars Rune Nøstdal wrote:
> BJVE does not know what he is after - he thinks he wants to do something,
> but only for riches and fame, not for the sake of doing it .. this does
> not work
at least not in this order
--
mvh,
Lars Rune Nøstdal
http://lars.nostdal.org/
Ulrich Hobelmann wrote:
> Hell, I used Linux for maybe two years, then Free & NetBSD for another
> two, now a Mac. Still, Linux sucks in a way. My main reason for
> getting a Mac was that I couldn't find a laptop (or information about
> one), that decently supported Linux, as all the information sites are --
> let's say -- very conservative about the amount of information they give
> you. I have (somewhat bad) experience with two laptops and Unix
> already, so the Mac was the obvious choice.
The comparison isn't very meaningful, since all that hardware that Linux
didn't support, isn't supported by Mac OS X either.
It makes Mac a very good choice in your case, but it definitely doesn't
mean that Linux "sucks".
--
Morten
Brandon J. Van Every wrote:
[ much text ]
Ok, I've tried it and it took me half a day to install the latest stable
Debian (Sarge), parallel to my Windows XP installation on my Laptop, the
latest CL-SDL, libsdl and NVIDIA graphics card and OpenGL drivers:
http://www.frank-buss.de/tmp/cl-sdl-opengl.png
You are right, it is difficult for the average end-user, who doesn't even
know what a operating system is and that there are more than Microsoft
Windows, because there were several problems, for example my mouse worked
only after getting the latest linux kernel (2.4.12.11) and configuring it
is a bit difficult, because I wanted to support my WiFi PCMCIA card, too,
etc. But it is a nice system for developing software and compared to the
distributions some years ago, it is much easier. I didn't need to write one
line of code, just downloading with the nice Debian package system and
calling various "make"'s and "configure"'s for the additional libraries and
drivers. Perhaps on more normal computers, without the need to develop
OpenGL software in Lisp, Linux will work out-of-the-box, but it should be
no problem for the average programmer to install and use such a system.
If you don't want to read manuals and rebuilding kernels, the alternative
is to use Windows or MacOS, but both are not free and if you find a bug,
you can't just patch the source and recompile it, as in Linux. And you
can't look at the source of the OS to see how it works. For example I've
helped to develop an embedded system with Linux, which needs some drivers
for special hardware and sometimes it was very useful to look at the kernel
sources to check how it works. If you do driver development in Windows, or
anything other below the official high-level API, you are doomed.
--
Frank Bu�, ··@frank-buss.de
http://www.frank-buss.de, http://www.it4-systems.de
"Frank Buss" <··@frank-buss.de> wrote:
> sources to check how it works. If you do driver development in Windows, or
> anything other below the official high-level API, you are doomed.
>
Nobody told the guys who wrote Internet Explorer and Office that _they_
should stick to the high-level APIs.
Clearly it was not an issue, as IE and Office are officially bug-free.
> --
> Frank Bu�, ··@frank-buss.de
Jamie
On 2005-06-12 11:11:25, Brandon J. Van Every wrote:
> This OS
> is seriously brain dead, that people have to amass so much expertise to
> make a device driver change.
Sorry, wrong newsgroup.
Brandon J. Van Every wrote:
> This is going to amuse the cat callers to no end. But reality is
> reality. Summary: I'm fleeing in terror from Linux. Which means I
> won't be using CMUCL or SBCL after all. Which might cause me to ditch
> Common Lisp, because I'm poor.
Ah, but you have to stop being poor before your opinion matters, you
see.
By the way, do you actually have a legal copy of Windows?
At least your crappy Linux copy is, dissatisfied as you might be with
it. :)
> I'm also realizing how irrelevant Linux is to shipping games. It isn't
> "an extra platform where I might pick up a few sales."
Extra? So where are the existing sales going? And why can't you afford
a proprietary Lisp system to run on your paid for and licensed Windows
installation?
Computer games bigshot, maybe you should pawn off your Rolex watch and
put a for sale sign on that Ferrari. :)
> Plus Apple is moving to x86, and in time will do all the things that the
> Linux world won't do.
But you are poor, remember? So whatever Apple rolls out, chances are
you won't be able to afford it. Thanks for dreaming on behalf of the
rest of us, though. :)
And that means hardware too; when Apple gets into X86, I don't think
that will mean you will be able to buy a shrink-wrapped box of their
operating system to run on your vanilla PC hardware. No, to ditch
Windows and become an Apple user, you will still have to buy an Apple
computer. (Prediction: there will probably be cracks to get Apple's OS
running on a vanilla PC). This is nothing new. Back in the days of
Apple II, clone machines came out, and people ran Apple operating
systems on non-Apple hardware. We might see an ``attack of the
clones'' if Apple makes something that is so similar to PC hardware
that a few mods on the hardware plus tweaks on the software can put
them together.
The Intel chip is quite irrelevant. Chips are a commodity, really. In
an ideal world, people would swap processors like dirty laundry, and
things would just recompile themselves. This is kind of done in the
embedded world. Not the magic recompiling part, of course, but the
dizzying array of different embedded processors that are out there.
Each one has its set of things that it's good at and loyal user base.
:)
In practical terms, Apple machines using Intel processors means really
only one thing: Apple users being able to run Windows apps at a decent
speed. But of course, legacy apps compiled for the POWER PC
architecture will have to be emulated.
Linux runs on PowerPC Macintoshes today. I.e. there are users of
Apple's own hardware who prefer to have Linux on it, at least some of
the time. Apple's software has not wiped out Linux on their own
hardware.
If Apple comes out with an X86 Mac, Linux will be ported to that. That
configuration will probably be able to share executables with an X86
Linux running on PC's. Great thing: compile Linux app on PC hardware,
ship on X86 Mac.
By the way, this reminds me, Sun Microsystems once moved to Intel
hardware. Back in the late 80's, they rolled out this box running
SunOS, a thing called the Sun 386i. Oh yeah, that really wiped out
every other Unix that was running on a 386, and sent MS-DOS scurrying
for cover. :)
They did this because the 386 emerged as a cheap, relatively fast
architecture with 32 bit physical addressing and demand paged virtual
memory on-chip. I.e. basically the same type of reasoning that Apple is
applying today.
Now if they had actually ported SunOS to an ordinary, stock 386 PC and
sold it, that would have been quite different.
Brandon J. Van Every wrote:
> This is going to amuse the cat callers to no end. But reality is
> reality. Summary: I'm fleeing in terror from Linux.
Interesting choice of words.
I guess it's about the level of interest you actually have in learning a
new OS, or distribution of it. There are numerous distributions more
mature than Ubuntu. You might try Xandros, which currently offers an
Open Circulation Distribution for free. The installation's improved a lot.
Each of your claims about Linux I could make about Windows, and then
some. XP doesn't have the drivers I need for certain pieces of hardware.
It runs amok with software designed for it. The default security is
appalling. I can't run decent games on it because there bloody well are
none.
> I've also become deeply averse to large stacks of dependent packages.
Windows is worse, son. Check out the reproducing DLL's sometime.
If you want to learn Linux, do so on your own, and not under the
tutelage of someone who doesn't have a clue about the variety of
distributions out there. If you don't want to learn, that's okay - but
there's hardly much solved by bitching about Linux on a LISP list.
Alpha
greyw01f wrote:
> If you don't want to learn, that's okay - but
> there's hardly much solved by bitching about Linux on a LISP list.
The group charter only prohibits bitching about Lisp.
greyw01f wrote:
> Brandon J. Van Every wrote:
>
>> This is going to amuse the cat callers to no end. But reality is
>> reality. Summary: I'm fleeing in terror from Linux.
>
>
> Interesting choice of words.
>
> I guess it's about the level of interest you actually have in learning a
> new OS,
That would be a wrong guess, as I'm experienced with Linux, if a long
time ago.
> or distribution of it. There are numerous distributions more
> mature than Ubuntu. You might try Xandros, which currently offers an
> Open Circulation Distribution for free. The installation's improved a lot.
If I had a broadband connection, I might have been more inclined to try
this, this, that, and the other distro. But as it is, I have to get
these things from friends with broadband, and it does take 2.5 hours to
download 2 CDs. Let's say I got past that; it's still a lot of futzing
to install and uninstall a bunch of distros until I find one that works
out of the box, if indeed that happens. A few days of lost time, I'd
anticipate.
Well, let's say I was willing to spend that time. Unfortunately a
number of my strategic questions have been answered: Linux will not be a
consumer desktop OS anytime soon. Developing games on Linux might be an
ok idea, but shipping games on Linux is a brain dead idea. So why
bother developing on Linux? Corman Lisp is only $200, and even in my
impoverished state, that's only 2 days' worth of signature gathering. 2
mediocre days, really.
I thought by moving to Linux, I was going to get away from all the open
source build-it configure-it-yourself loose ends, because everything
would be packaged. Well, everything is not packaged, and there's so
much stuff that I fear loose ends will inevitably arise and be a way of
life. So it doesn't appear to be a better trade than banging on MinGW.
Supporting MinGW is painful, but Linux appears to be equally painful,
not a solution.
Finally, Apple announced they're moving to x86 as I was doing all of
this. By the time I ship, I bet Linux is increasinly irrelevant as a
consumer desktop platform. It may have permanently lost its window of
opportunity.
> Each of your claims about Linux I could make about Windows, and then
> some.
No; Windows has the game market, and Linux doesn't. Consumer HW
intended to work on Windows, works on Windows. People already have an
installed base of Windows HW, because if something didn't work, they
returned it to the store and got something that did.
> XP doesn't have the drivers I need for certain pieces of hardware.
> It runs amok with software designed for it. The default security is
> appalling.
Not relevant to selling games. Not relevant to my development either;
my W2K box has been working great for 4 years now.
> I can't run decent games on it because there bloody well are
> none.
The shittiness of the game market in general is not my problem. I
intend to *ship* games, not play other people's games.
> If you want to learn Linux, do so on your own,
People are always making this mistake that I don't know Linux. I just
don't know modern Linux. I had sub-guru knowledge back in the day, and
the nature of the beast hasn't changed *that* much since 1996. People
are too into learning stuff, not enough into just using stuff. Drivers
and installing and HW and configurations get old when you've done it,
oh, 1000+ times in your career.
> and not under the
> tutelage of someone who doesn't have a clue about the variety of
> distributions out there.
My friend is quite knowledgeable. Of course he doesn't know every
detail about every distro, but he certainly knows the risk level of the
landscape. Risk management is a valid way of making decisions about
what platforms you're going to commit to. From that standpoint, I
should have dumped the turkey known as Ubuntu on Day 2. (Where "turkey"
implies "on my HW.")
> If you don't want to learn, that's okay - but
> there's hardly much solved by bitching about Linux on a LISP list.
Initially I was only trying to explain / figure out its impact upon my
Lisp vs. Scheme choice, but the thread developed a life of its own. It
seems like Corman Lisp is the most obvious escape route. CLISP now +
wait for a CMUCL port later might also work. GCL I've decided isn't
going to work, any better than Linux worked.
--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA
"We live in a world of very bright people building
crappy software with total shit for tools and process."
- Ed McKenzie