On Sun, 04 Dec 2005 07:45:45 GMT, olczyk <·············@yahoo.com>
wrote:
>Title says it all.
Maybe he found some polite, non-nagging, beta testers, and decided to
do a 3-year beta test.
On Sun, 04 Dec 2005 10:53:28 -0800, alex.gman wrote:
> Do you have any reason to believe it wasn't a joke in the first place?
Aha. The default mode for posting to usenet is joking so all those
announcements we see are jokes. Do you have even one tiny tiny thing to
indicate it was a joke?
BTW: Before you suggest that post was a joke maybe you should read it.
http://groups.google.com/group/comp.lang.lisp/browse_thread/thread/f0e61f93a30da883/e8c117d8671b8c25?lnk=st&q=cmucl+shapiro+win32&rnum=5&hl=en#e8c117d8671b8c25
Does it sound anything remotely like a joke to you?
PS: If it really were a joke, I would expect CS to come along and say
something like, "It was just a joke. Sorry for getting your hopes up."
PPS: Since lemonodor carried an announcement several months before the
post that people were working on ports of both CMUCL and SBCL to Windows,
and mention CS, does that meant that the webmaster is on the joke?
olczyk wrote:
> On Sun, 04 Dec 2005 10:53:28 -0800, alex.gman wrote:
>
>
>>Do you have any reason to believe it wasn't a joke in the first place?
>
> Aha. The default mode for posting to usenet is joking so all those
> announcements we see are jokes. Do you have even one tiny tiny thing to
> indicate it was a joke?
I suggest you sue him for the return of every last dollar
you paid him to work on this.
Paul
Here's what I think really happened. CMUCL/SBCL officials (or official
web site maintainers) always claimed that porting to Win32 is trivial
for anyone competent, and the only reason it hasn't been done is
because no one is interested in that "crap OS".
"Crap OS" or not, the practical importance of a Win32 port is obvious,
and only people completely unequipped with critical thinking abilities
would of course buy this explanation. How the hell would they know how
hard it is to port to Win32 without doing it?
Carl Shapiro probably learned it the hard way. I'm not in contact with
Carl Shapiro, but something tells me he had wanted to present CMUCL for
Win32 to Lisp Conference 2005 attendees, a conference he organized. He
probably fell victim to the very common programmer fault of
underestimating the effort required to complete the project. So he had
to settle for the next best thing, i.e. PROMISING to deliver CMUCL for
Win32 real soon.
To all: I acknowledge that SBCL web site estimated that a Windows port
would take 1-12 months. It seems some people here are interested in
arguing whether that warrants calling it "trivial". I say that whether
1-12 months is "trivial" depends on the importance of the project.
In any case, they did say that the possible Windows port would probably
be "straightforward".
This concludes my participation in this useless discussion, as, unlike
yours, my time is very much non-free.
P.S. If you got nothing better to do than to argue about this subject,
instead of wasting time here, why don't you head over to
alt.suicide.holiday and try to convince some people not to kill
themselves this holiday season. Some of them are serious (CNN wrote
about the group earlier) I've done my share. I think I saved a dozen of
them or so. It's your turn now.
From: Christophe Rhodes
Subject: Re: Carl Shapiro: Where is the Win32 port of CMUCL?
Date:
Message-ID: <sqy82huely.fsf@cam.ac.uk>
·········@gmail.com writes:
> Here's what I think really happened. CMUCL/SBCL officials (or official
> web site maintainers) always claimed that porting to Win32 is trivial
> for anyone competent, and the only reason it hasn't been done is
> because no one is interested in that "crap OS".
The SBCL web site has had, since October 2002, the following text:
In increasing order of estimated difficulty, the following ports are
possible:
* Porting to another UNIX-y OS, or extending SBCL's coverage of an
existing operating system to an existing processor port
(e.g. Darwin or Solaris on x86, NetBSD on PowerPC, etc.)
* Porting to a CPU supported by CMUCL but not by SBCL (the only
remaining such architecture seems to be HPPA, where we have a
dormant Linux/HPPA port with many bugs.)
* Porting to Microsoft Windows or another non-UNIX OS.
* Extending an existing port to a 64-bit variant of that CPU
(e.g. 64-bit PowerPC, MIPS, SPARC, Alpha, or x86).
* Porting to a new CPU from scratch.
All of these except the first are probably between a man-month and a
man-year for someone who already knows what he's doing. For someone
learning compilers or assembler or Lisp implementation fundamentals
or SBCL implementation idiosyncrasies or whatever as he goes, it
might take longer.
I don't think that this is a claim that a port of SBCL to Microsoft
Windows is a trivial affair for anyone competent; certainly I don't
consider between a month and a year of my time as trivial.
As an additional data point, SBCL 0.8.9 was released on the 25th March
2004; Alastair Bridgewater's basically functional binary release of
SBCL 0.8.9 for Windows was announced on the 2nd of July 2004. This
would seem to indicate that the estimate of the difficulty of porting
SBCL to Windows on the SBCL website was broadly correct.
Would you like to reassess what you think really happened?
Christophe
Christophe Rhodes wrote:
> certainly I don't
> consider between a month and a year of my time as trivial.
There is no general numerical value for what constitutes "trivial".
Making restaurant reservations is non-trivial if it requires staying on
hold for an hour.
> Would you like to reassess what you think really happened?
No. If you read my post, you'll notice that the whole topic of what
CMUCL/SBCL people claimed is an aside (And SBCL web site did use the
word "straightforward" to describe the possible Win32 port. Sifting
through forums, you'll see the lack of interest quoted as the major
reason)
What "really happened" is that Carl underestimated the difficulty of
completing the project. And this seems like a very safe assessment,
unless something bad happened to him following the conference.
From: Christophe Rhodes
Subject: Re: Carl Shapiro: Where is the Win32 port of CMUCL?
Date:
Message-ID: <sq4q54g79y.fsf@cam.ac.uk>
·········@gmail.com writes:
> Christophe Rhodes wrote:
>> certainly I don't
>> consider between a month and a year of my time as trivial.
>
> There is no general numerical value for what constitutes "trivial".
> Making restaurant reservations is non-trivial if it requires staying on
> hold for an hour.
Well, that's nice, but I don't consider between a month and a year of
my time to be trivial, whatever I'm doing with that time.
>> Would you like to reassess what you think really happened?
>
> No. If you read my post, you'll notice that the whole topic of what
> CMUCL/SBCL people claimed is an aside
I don't care if your misinformation is an aside or not; it needs to be
corrected.
> (And SBCL web site did use the word "straightforward" to describe
> the possible Win32 port.
To the best of my knowledge, this is false. Until May of 2001,
Windows was not mentioned on the SBCL website at all; the word
"straightforward" was used to describe ports of SBCL to
architecture/OS combinations that CMUCL supported at the time. May
2001 saw the introduction of the phrase
Ports to completely new CPUs or OSes (Windows?) are also possible,
but a new CPU is known to be a lot of work, and despite much
speculation no one really knows how hard a Windows port would be.
which doesn't sound like an implication of straightforwardness. This
phrase was then replaced by the excerpt I quoted in my previous mail,
which remains to this day. The wayback internet archive has 16
versions of the SBCL website; anyone who wishes to can check my
description.
> What "really happened" is that Carl underestimated the difficulty of
> completing the project. And this seems like a very safe assessment,
> unless something bad happened to him following the conference.
Well, unless Carl chooses to enlighten this forum with an account of
his actions, this is unfalsifiable, but I hope it's clear to the
general onlooker that this is not a safe assessment at all; that there
are many reasons why one might not wish to make one's work available
to everybody; and that there is an existence proof that porting a
CMUCL-like system to Windows is the matter of a few man-months.
Christophe
Christophe Rhodes wrote:
> I don't care if your misinformation is an aside or not; it needs to be
> corrected.
>
> > (And SBCL web site did use the word "straightforward" to describe
> > the possible Win32 port.
>
> To the best of my knowledge, this is false.
If you have to add "to the best of my knowledge", maybe you shouldn't
call what I wrote "misinformation"?
From: Christophe Rhodes
Subject: Re: Carl Shapiro: Where is the Win32 port of CMUCL?
Date:
Message-ID: <sqoe3ceq9r.fsf@cam.ac.uk>
·········@gmail.com writes:
> Christophe Rhodes wrote:
>
>> I don't care if your misinformation is an aside or not; it needs to be
>> corrected.
>>
>> > (And SBCL web site did use the word "straightforward" to describe
>> > the possible Win32 port.
>>
>> To the best of my knowledge, this is false.
>
> If you have to add "to the best of my knowledge", maybe you shouldn't
> call what I wrote "misinformation"?
Why? What you wrote was
| CMUCL/SBCL officials (or official web site maintainers) always
| claimed that porting to Win32 is trivial for anyone competent, and
| the only reason it hasn't been done is because no one is interested
| in that "crap OS"
and I was being generous in only calling that "misinformation".
The issue that I'm not completely sure about, and so I add "to the
best of my knowledge", is whether or not the SBCL web site ever used
the word "straightforward" to describe the possible Win32 port; that
does not change the fact that you've already lied (I'm no longer
feeling generous) once in this thread.
Christophe
trivial = straightforward, and they did call it straightforward, which
I think you know, because you quoted very selectively from the same
page, excluding the "straightforward" part. So you lied about it.
From: Christophe Rhodes
Subject: Re: Carl Shapiro: Where is the Win32 port of CMUCL?
Date:
Message-ID: <sq1x08enqg.fsf@cam.ac.uk>
·········@gmail.com writes:
> trivial = straightforward, and they did call it straightforward, which
> I think you know, because you quoted very selectively from the same
> page, excluding the "straightforward" part.
For the sake of doubt avoidance: which page?
My guess is that alex.gman is referring to this statement, currently
found on <http://www.sbcl.org/porting.html>:
Also note that cross-compilation is straightforward, [...]
which is taken from a different -- "Bootstrapping" -- section of the
page, and helpfully links to a description of how to cross-compile the
lisp files of an already-ported SBCL using an suitably ANSIish Lisp
running on an arbitrary CPU/OS platform; it also directs the reader:
(See the platform support page for information on which existing
platforms SBCL runs on.)
I'll happily stand by that statement, whether or not it's the one that
alex.gman is fantasizing about; cross-compiling SBCL using any of the
host Lisps listed is straightforward, irrespective of the CPU/OS
combination the host Lisp is running on.
> So you lied about it.
No.
Christophe
ObLisp: <http://www.dridus.com/~nyef/sbcl-win32-0.9.6-take2.patch.gz>
+ ·········@gmail.com:
| trivial = straightforward, and they did call it straightforward, which
| I think you know, because you quoted very selectively from the same
| page, excluding the "straightforward" part. So you lied about it.
Are we talking about <http://www.sbcl.org/porting.html> here, or the
seemingly identical <http://sbcl.sourceforge.net/porting.html>, in
their present versions? Because if we are, the only occurence of the
word "straightforward" on the page is in the sentence
Also note that cross-compilation is straightforward, so if you have
access to any CPU/OS combination which runs one of the "works"
choices above it should be possible to bootstrap an SBCL for your
target.
And that is about something entirely different. So unless you can
come up with a much more precise pointer to the statement you're
talking about, you look like the more likely liar here.
--
* 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>:
| [...] So unless you can come up with [...]
Whoops ...
;;; too late to cancel, given the turnaround time on this group ...
(push *new-year-resolutions* "Do not feed the trolls.")
--
* 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
From: Edi Weitz
Subject: Re: Carl Shapiro: Where is the Win32 port of CMUCL?
Date:
Message-ID: <uwti1nfh5.fsf@agharta.de>
On 19 Dec 2005 09:40:31 -0800, ·········@gmail.com wrote:
> CMUCL/SBCL officials (or official web site maintainers) always
> claimed that porting to Win32 is trivial for anyone competent
I don't think I've ever read or heard that. Would you care to provide
evidence for this claim or are you just trolling around?
--
Lisp is not dead, it just smells funny.
Real email: (replace (subseq ·········@agharta.de" 5) "edi")
·········@gmail.com ha escrito:
> Scratch "trivial", replace with "straightforward". Happy now?
According to SBCL web page, it may take "between a man-month and a
man-year for someone who already knows what he's doing". I assume that
CMUCL will take a similar time.
That's a lot of time when people are having other things to do with
their lifes apart from programming, for *free*, a Lisp implementation
for a SO which they don't ussualy are interested for themselves.
So, if Carl has said that he is going to port it for win32, you should
thank him, not bothering him.
javuchi wrote:
> ·········@gmail.com ha escrito:
>
> > Scratch "trivial", replace with "straightforward". Happy now?
>
> According to SBCL web page, it may take "between a man-month and a
> man-year for someone who already knows what he's doing". I assume that
> CMUCL will take a similar time.
> That's a lot of time when people are having other things to do with
> their lifes apart from programming, for *free*, a Lisp implementation
> for a SO which they don't ussualy are interested for themselves.
> So, if Carl has said that he is going to port it for win32, you should
> thank him, not bothering him.
On that note, people interested in the nature of trolls might want to
look over this Trolling HOWTO:
http://linux.nullcode.org/troll.txt
Tayssir
--
"Let's talk about the question of why people are wealthy. There is a
myth that it's a function of enormous personal attributes... the
individual wealth which is generated in this economy is, in my
judgment, and I doubt that there is much that anyone could disagree
with about this, is a function of the innovative businesses which are
created as a result of federal research. But you understand that the
people who benefit from that research get it free... It starts from
this incredible research activity which is going on with federal
money."
-- Bill Gates Sr., 2003
From: Edi Weitz
Subject: Re: Carl Shapiro: Where is the Win32 port of CMUCL?
Date:
Message-ID: <uk6e0or75.fsf@agharta.de>
On 19 Dec 2005 10:21:06 -0800, ·········@gmail.com wrote:
> Scratch "trivial", replace with "straightforward". Happy now?
No. I didn't ask you to modify your claim but to provide evidence
which you haven't done yet. Where did you read the statement that
porting CMUCL/SBCL to Windows is "straightforward?"
FWIW, my impression from what I've heard and read from the developers
always was that vital parts of CMUCL/SBCL are based on the assumption
that the underlying OS is "Unix-y" and that porting to a platform for
which this assumption doesn't hold is neither "trivial" nor
"straightforward."
--
Lisp is not dead, it just smells funny.
Real email: (replace (subseq ·········@agharta.de" 5) "edi")
Edi Weitz wrote:
> On 19 Dec 2005 10:21:06 -0800, ·········@gmail.com wrote:
>
> > Scratch "trivial", replace with "straightforward". Happy now?
>
> No. I didn't ask you to modify your claim but to provide evidence
> which you haven't done yet. Where did you read the statement that
> porting CMUCL/SBCL to Windows is "straightforward?"
If you think you can get links, references and thank-you's from people
you try to publicly insult, you need better social skills.
From: Edi Weitz
Subject: Re: Carl Shapiro: Where is the Win32 port of CMUCL?
Date:
Message-ID: <ubqzcoqce.fsf@agharta.de>
On 19 Dec 2005 11:02:40 -0800, ·········@gmail.com wrote:
> If you think you can get links, references and thank-you's from
> people you try to publicly insult, you need better social skills.
I never thought I would get links or references because it was pretty
clear from the beginning that you were making this up. I just didn't
want to let that slip through. Thanks for confirming that and end of
discussion for me.
--
Lisp is not dead, it just smells funny.
Real email: (replace (subseq ·········@agharta.de" 5) "edi")
Edi Weitz wrote:
> On 19 Dec 2005 11:02:40 -0800, ·········@gmail.com wrote:
>
> > If you think you can get links, references and thank-you's from
> > people you try to publicly insult, you need better social skills.
>
> I never thought I would get links or references because it was pretty
> clear from the beginning that you were making this up. I just didn't
> want to let that slip through. Thanks for confirming that
Confirming what? That I don't respond with links and references to rude
accusations of trolling? You need to get out more, if your social
skills are so bad.
> and end of
> discussion for me.
Just to let people know, Carl showed me CMUCL under Windows on my
laptop over a year ago. It was, of course, not usable because of many
bugs. There are a lot of low-level differences between Unix and
Windows.
The main show-stopper now is memory management.
The problem is that Carl works on this when he's not working on
projects that actually bring income to SRI; those obviously have
priority.
--
Fred Gilham ······@csl.sri.com
Communism is a murderous failure.
Socialism is communism with movie stars.
From: Eric Marsden
Subject: Re: Carl Shapiro: Where is the Win32 port of CMUCL?
Date:
Message-ID: <87oe3c4yl2.fsf@free.fr>
>>>>> "ag" == alex gman <·········@gmail.com> writes:
ag> Here's what I think really happened. CMUCL/SBCL officials (or official
ag> web site maintainers) always claimed that porting to Win32 is trivial
ag> for anyone competent, and the only reason it hasn't been done is
ag> because no one is interested in that "crap OS".
I maintain the CMUCL web site, and am subscribed to both of the main
CMUCL mailing lists. I do not remember seeing in those locations a
statement regarding either the "trivial" nature of a port to win32,
or even a statement from a CMUCL developer using terms like "crap OS".
Please provide references backing up your accusations, or stop
spreading crap over comp.lang.lisp.
--
Eric Marsden
Eric Marsden wrote:
> Please provide references backing up your accusations, or stop
> spreading crap over comp.lang.lisp.
Oh, my Lord, he shat in the church!
Actually, when I wrote CMUCL/SBCL, I didn't remember which. Now, I
looked it up, and it's SBCL. CMUCL did in fact say it was going to be
hard to port to Windows.
olczyk <·············@yahoo.com> writes:
> Title says it all.
So why, exactly, do you think that harassing someone in a public forum
is a good way to get them to give you something they don't owe you?
Unlike yourself, Mr. Shapiro posts to usenet using an obviously legit
e-mail address -- why don't you inquire politely as to the status of
the presumably unfinished port? Oh, right, becuase you're a troll and
all but a stalker.
If I were in Shapiro's shoes, I would feel *very* uninspired to
continue or release that work. Can you imagine how this guy would
file bug reports?
--
/|_ .-----------------------.
,' .\ / | Free Mumia Abu-Jamal! |
,--' _,' | Abolish the racist |
/ / | death penalty! |
( -. | `-----------------------'
| ) |
(`-. '--.)
`. )----'
From: Giorgos Keramidas
Subject: Re: Carl Shapiro: Where is the Win32 port of CMUCL?
Date:
Message-ID: <868xujv2p6.fsf@flame.pc>
On 05 Dec 2005 02:25:13 -0800,
···@conquest.OCF.Berkeley.EDU (Thomas F. Burdick) wrote:
> olczyk <·············@yahoo.com> writes:
>> Title says it all.
>
> So why, exactly, do you think that harassing someone in a public forum
> is a good way to get them to give you something they don't owe you?
[...]
> If I were in Shapiro's shoes, I would feel *very* uninspired to
> continue or release that work. Can you imagine how this guy would
> file bug reports?
Can you imagine how fast this guy's email address would be blocked in my
bug report system? :-)
From: Giorgos Keramidas
Subject: Re: Carl Shapiro: Where is the Win32 port of CMUCL?
Date:
Message-ID: <86aceytsu2.fsf@flame.pc>
On 17 Dec 2005 15:06:53 -0800, "javuchi" <·······@gmail.com> wrote:
> What is win32? ;-)
32-bit Microsoft Windows. I'm not sure if there's a 64-bit Windows port
and, frankly, I don't care at all to find out :)
Giorgos Keramidas ha escrito:
> On 17 Dec 2005 15:06:53 -0800, "javuchi" <·······@gmail.com> wrote:
> > What is win32? ;-)
>
> 32-bit Microsoft Windows. I'm not sure if there's a 64-bit Windows port
> and, frankly, I don't care at all to find out :)
It was just a joke... ;-) I don't use Windows since several years ago,
and asked myself it it still exists... ;-)
Which brings me up to my mind this question: why a Lisp hacker would
want to port CMCL or SBCL to windows, and have the responsability of
manteining it?
> Which brings me up to my mind this question: why a Lisp hacker would
> want to port CMCL or SBCL to windows, and have the responsability of
> manteining it?
An even better question would be: Why would a lisp hacker want to use a UNIX
like operating system? UNIX is so ugly, so different from the philosophy of
a good Lisp system it makes me want to puke.
Which brings me up to my mind this question: do joe lisp developers have any
knowledge of operating system design?
"Mlop" <···@ml.com> writes:
> > Which brings me up to my mind this question: why a Lisp hacker would
> > want to port CMCL or SBCL to windows, and have the responsability of
> > manteining it?
>
> An even better question would be: Why would a lisp hacker want to use a UNIX
> like operating system? UNIX is so ugly, so different from the philosophy of
> a good Lisp system it makes me want to puke.
Most likely because, even though (L|U)*nix sucks big time, it is
better than any other currently "viable" alternatives. Pretty sad,
but there it is.
> Which brings me up to my mind this question: do joe lisp developers
> have any knowledge of operating system design?
Oh yes. Again, most likely, as a group, they are far more versed in
this knowledge than joe/jane <most-any-other-language> developers.
/Jon
--
'j' - a n t h o n y at romeo/charley/november com
From: Ulrich Hobelmann
Subject: Re: Carl Shapiro: Where is the Win32 port of CMUCL?
Date:
Message-ID: <40lop6F1b3ihrU1@individual.net>
jayessay wrote:
> "Mlop" <···@ml.com> writes:
>
>>> Which brings me up to my mind this question: why a Lisp hacker would
>>> want to port CMCL or SBCL to windows, and have the responsability of
>>> manteining it?
>> An even better question would be: Why would a lisp hacker want to use a UNIX
>> like operating system? UNIX is so ugly, so different from the philosophy of
>> a good Lisp system it makes me want to puke.
>
> Most likely because, even though (L|U)*nix sucks big time, it is
> better than any other currently "viable" alternatives. Pretty sad,
> but there it is.
To me it seems like quite many Lisp developers are Mac-users (even
pre-OS X), compared to other computer-users or programmers.
Of course now that Mac OS is Unix-based, it's usable even for the rest
of us (Unix-heads).
I agree that in itself Unix is quite cheap, but as a de-facto standard
platform it's still the best we have (and I still use ssh and X11 to
Unix systems more or less regularly).
--
If you have to ask what jazz is, you'll never know.
Louis Armstrong
Mlop ha escrito:
> > Which brings me up to my mind this question: why a Lisp hacker would
> > want to port CMCL or SBCL to windows, and have the responsability of
> > manteining it?
>
> An even better question would be: Why would a lisp hacker want to use a UNIX
> like operating system? UNIX is so ugly, so different from the philosophy of
> a good Lisp system it makes me want to puke.
>
> Which brings me up to my mind this question: do joe lisp developers have any
> knowledge of operating system design?
Unix is made by the summ of small utilities, and is the result of many,
many years of improvements.
As far as I know, Lisp is also the summ of small utilities, and has
been improved for many years.
But just take the alternative... Windows is a block, monolitic,
fragile, inestable in comparation.
And the other alternative from Apple... has become Unix, of course. OSs
are moving to Unix in the same way that computer languages are moving
to Lisp.
> Unix is made by the summ of small utilities, and is the result of many,
> many years of improvements.
The problem with these tools is that they can't communicate with the outside
world without a translator(ie. text parsing and pipping). They're autistic.
You can hire a Chinese mathematician and an Italian biologist that are great
at solving little tasks but when you want them to work together you have to
hire another person to make them understand each other. There's a loss of
productivity and flexibility right there.
> As far as I know, Lisp is also the summ of small utilities, and has
> been improved for many years.
No, Emacs is not the sum of small utilities. It is a large integrated system
of many people who speak the same language. There is no translation needed.
No text parsing, no pipping, no productivity loss.
> But just take the alternative... Windows is a block, monolitic,
> fragile, inestable in comparation.
Monolithic? You must be refering to Linux.
> And the other alternative from Apple... has become Unix, of course. OSs
> are moving to Unix in the same way that computer languages are moving
> to Lisp.
Unix is the idiot savant. The autistic child of operating systems. You try
to talk to it but there's no useful feedback.
Mlop ha escrito:
> > Unix is made by the summ of small utilities, and is the result of many,
> > many years of improvements.
>
> The problem with these tools is that they can't communicate with the outside
> world without a translator(ie. text parsing and pipping). They're autistic.
> You can hire a Chinese mathematician and an Italian biologist that are great
> at solving little tasks but when you want them to work together you have to
> hire another person to make them understand each other. There's a loss of
> productivity and flexibility right there.
Windows is even worse, as it started without pipes.
Elsewhere, I can imagine lot of ways for unix tools to intercomunicate,
not only text parsing nor pipping. For example: databases, sockects
(which I remind you: Unix is the mother of Internet, don't forget it),
dynamic libraries (they were avaiable in Unix much time before even
Windows was born).
Can you tell me a way in which windows "utilities" (if they exists at
all) can comunicates and Unix do not?
> > As far as I know, Lisp is also the summ of small utilities, and has
> > been improved for many years.
>
> No, Emacs is not the sum of small utilities.
That is what you think, but not the real thing. Just think about Emacs
standalone and you do not have much more than a simple editor. But sum
to it GNUS, SLIME, Info, planner, and so on, and you get a completly
better thing.
> It is a large integrated system
> of many people who speak the same language. There is no translation needed.
> No text parsing, no pipping, no productivity loss.
I don't think so. You can substitute SLIME for anything different (for
example ILisp). You can add or remove add-ons. I call this modularity.
> > But just take the alternative... Windows is a block, monolitic,
> > fragile, inestable in comparation.
>
> Monolithic? You must be refering to Linux.
If you're talking about the kernel, yes, you're right. But this is
radically not aplicable to the rest of the system (the important thing
at the end), which is an order of magnitude much more modularized than
Windows. Just look at where all Windows configuration is stored: in a
file called the register, isn't it monolitic? And, how much kernels of
Windows do you have? We have a lot of Unix ones to choose. Do you have
something similar to grep, find, and shell programming in Windows by
default? No, you have to install Unix-like utilities which do not have
the same power at the end.
Even Windows NT (or whatever they call now) does have utilities which
are hard (or even imposible) to be extended by normal people. They are
nice, yes, but you can have the power and the nice thing with linux
too. Look at Webmin, you have no excuse, as it allows you to configure
Unix without having to enter the shell. Elsewhere, the shell is there
when more power is needed.
> > And the other alternative from Apple... has become Unix, of course. OSs
> > are moving to Unix in the same way that computer languages are moving
> > to Lisp.
>
> Unix is the idiot savant. The autistic child of operating systems. You try
> to talk to it but there's no useful feedback.
Just your opinion.
Elsewhere, I stop this flamewar here. Unix programmers (I dont like
very much the term "hacker") know very well why they choose Unix, the
same as Lisp programmers do. And I'm not going to repeat my words again
in this thread, so bye.
(If you want to convert this to something personal, don't bother
others, this is my mail: ·······@gmail.com)
From: Ulrich Hobelmann
Subject: Re: Carl Shapiro: Where is the Win32 port of CMUCL?
Date:
Message-ID: <40n6d1F1bd3vcU1@individual.net>
javuchi wrote:
> Windows is even worse, as it started without pipes.
> Elsewhere, I can imagine lot of ways for unix tools to intercomunicate,
> not only text parsing nor pipping. For example: databases, sockects
> (which I remind you: Unix is the mother of Internet, don't forget it),
> dynamic libraries (they were avaiable in Unix much time before even
> Windows was born).
> Can you tell me a way in which windows "utilities" (if they exists at
> all) can comunicates and Unix do not?
Windows program fragments can be dynamically loaded, and can cooperate
just like other processes on other OSes: through messaging in general.
Ok, there may have been no pipes, but pipes are from a time when there
was no dynamic loading, and thus every piece of code had to reside in a
single process. Pipes were basically a shell feature, to allow process
pipes to be constructed. They're deeply related with the fork() model
of process creation (instead of something like pthread_create(function)).
Process boundaries, or not, every decent OS today has messaging, and
honestly I emulate messages even on Unix over sockets (which are
probably the lowest common denominator as regards cross-system messaging).
>>> As far as I know, Lisp is also the summ of small utilities, and has
>>> been improved for many years.
>> No, Emacs is not the sum of small utilities.
>
> That is what you think, but not the real thing. Just think about Emacs
> standalone and you do not have much more than a simple editor. But sum
> to it GNUS, SLIME, Info, planner, and so on, and you get a completly
> better thing.
It's an environment for programs to extend and to work with, somewhat
like Eclipse, less structured, but arguably easier to program for.
>> It is a large integrated system
>> of many people who speak the same language. There is no translation needed.
>> No text parsing, no pipping, no productivity loss.
>
> I don't think so. You can substitute SLIME for anything different (for
> example ILisp). You can add or remove add-ons. I call this modularity.
I think the no parsing, no piping comes from Emacs being single-process.
As data doesn't have to be de/serialized anywhere, you can *work* with
the data instead of worrying about reformatting and re-parsing it all
the time (as you would have to with pipes).
It's also modular.
--
If you have to ask what jazz is, you'll never know.
Louis Armstrong
On Mon, 19 Dec 2005 08:39:12 +0100, Ulrich Hobelmann <···········@web.de>
wrote:
> javuchi wrote:
>> Windows is even worse, as it started without pipes.
>> Elsewhere, I can imagine lot of ways for unix tools to intercomunicate,
>> not only text parsing nor pipping. For example: databases, sockects
>> (which I remind you: Unix is the mother of Internet, don't forget it),
>> dynamic libraries (they were avaiable in Unix much time before even
>> Windows was born).
>> Can you tell me a way in which windows "utilities" (if they exists at
>> all) can comunicates and Unix do not?
>
> Windows program fragments can be dynamically loaded, and can cooperate
> just like other processes on other OSes: through messaging in general.
> Ok, there may have been no pipes, but pipes are from a time when there
> was no dynamic loading, and thus every piece of code had to reside in a
> single process. Pipes were basically a shell feature, to allow process
> pipes to be constructed. They're deeply related with the fork() model
> of process creation (instead of something like pthread_create(function)).
Windows has always had pipes of a sort.. That comes from the MS-DOS era.
Of course they were just a command utility that wrote to a file
and the printed the file to the stdin of the receiving process.
Of course at the time there was no incentive to compete with UNIX
which would have been way to big to run on the PC's of the time
anyhow. From MS-DOS 3.3 on up it has supported networking. Thus the
need to communicate with other systems made convergence toward
UNIX networking protocols and utilities desirable. Certainly
modern windows has system pipes. UNIX pipes are not just a command
shell utility by the way. In the old day they were the only way to
distribute computation. Multiple processes would share data using pipes.
You are correct that the fork()/exec() model in UNIX makes it cheaper
to create and run separate processes. Under Windows you would use
threads for most of the things you would use processes for.
Like when they ported Apache to Windows it only gains satisfactory
performance when they started spawning a thread for each request rather
than a process. You do seem to be forgetting remote procedure calls (RPC)
however which provide a message based communication.
Under windows you might prefer COM+ style however.
--
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/
From: Ulrich Hobelmann
Subject: Re: Carl Shapiro: Where is the Win32 port of CMUCL?
Date:
Message-ID: <40oc62F1ailr6U1@individual.net>
John Thingstad wrote:
> UNIX pipes are not just a command
> shell utility by the way. In the old day they were the only way to
> distribute computation. Multiple processes would share data using pipes.
Yes.
> You are correct that the fork()/exec() model in UNIX makes it cheaper
> to create and run separate processes. Under Windows you would use
I didn't even say that. To me the cost of process creation or
context-switching is a matter of implementation quality. I can't
imagine why other OSes took *several* times as long as even old,
unoptimized Unix, to do that.
I think something like spawn(function or program) is just as good as
fork+exec, in fact it's a single syscall instead of two. The reason
Unix had fork(), was that it followed KISS for a while, and didn't want
to distinguish processes and threads. Of course now Unix is more
complex than anything, so that doesn't really apply anymore.
> threads for most of the things you would use processes for.
For a long time I bought into the "processes are elegant, threads are
evil" thing. By now I'm on the thread side. I don't want to discuss
here, why, or even that, one side is better than the other. Matter of
taste.
> Like when they ported Apache to Windows it only gains satisfactory
> performance when they started spawning a thread for each request rather
> than a process. You do seem to be forgetting remote procedure calls (RPC)
> however which provide a message based communication.
> Under windows you might prefer COM+ style however.
I've read that Windows is still expensive with its thread- and
process-creation and context switching, so that's probably why the world
has great inventions like Thread Pools that reduce load by maybe 1% on
heavily loaded web servers (because I don't believe that efficient
thread creation, say on Linux, adds more overhead than that to a web
request).
RPC is usually based on UDP, AFAIK, which is interesting, but unreliable
(unless you reimplement half of TCP). COM+, CORBA and friends are
probably different.
--
If you have to ask what jazz is, you'll never know.
Louis Armstrong
From: Marcin 'Qrczak' Kowalczyk
Subject: Re: Carl Shapiro: Where is the Win32 port of CMUCL?
Date:
Message-ID: <87u0d4oqu4.fsf@qrnik.zagroda>
Ulrich Hobelmann <···········@web.de> writes:
> I think something like spawn(function or program) is just as good as
> fork+exec, in fact it's a single syscall instead of two. The reason
> Unix had fork(), was that it followed KISS for a while, and didn't
> want to distinguish processes and threads.
I think the primary advantage of fork+exec vs. spawn is that you don't
have to encode environmental differences between the child and the
parent process as parameters: I/O redirection, current directory,
environment variables, ignored signals, memory limits, scheduling
priority etc. They can be applied as function calls between the fork
and the exec.
--
__("< Marcin Kowalczyk
\__/ ······@knm.org.pl
^^ http://qrnik.knm.org.pl/~qrczak/
On Mon, 19 Dec 2005 19:59:47 +0100, Marcin 'Qrczak' Kowalczyk
<······@knm.org.pl> wrote:
> Ulrich Hobelmann <···········@web.de> writes:
>
>> I think something like spawn(function or program) is just as good as
>> fork+exec, in fact it's a single syscall instead of two. The reason
>> Unix had fork(), was that it followed KISS for a while, and didn't
>> want to distinguish processes and threads.
>
> I think the primary advantage of fork+exec vs. spawn is that you don't
> have to encode environmental differences between the child and the
> parent process as parameters: I/O redirection, current directory,
> environment variables, ignored signals, memory limits, scheduling
> priority etc. They can be applied as function calls between the fork
> and the exec.
>
Seemed to remember I had discussed fork before..
Cameron MacKinnon seems to have a pretty descent explanation.
http://groups.google.com/group/comp.lang.lisp/browse_thread/thread/252c1cebdbde94c0/793fbc9779c9a3f9?lnk=st&q=fork+John+Thingstad+group%3Acomp.lang.lisp&rnum=1&hl=en#793fbc9779c9a3f9
--
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/
From: Ulrich Hobelmann
Subject: Re: Carl Shapiro: Where is the Win32 port of CMUCL?
Date:
Message-ID: <40os70F1agvcvU1@individual.net>
John Thingstad wrote:
>> I think the primary advantage of fork+exec vs. spawn is that you don't
>> have to encode environmental differences between the child and the
>> parent process as parameters: I/O redirection, current directory,
>> environment variables, ignored signals, memory limits, scheduling
>> priority etc. They can be applied as function calls between the fork
>> and the exec.
Yes, but I don't like process-wide information like the environment or
the current directory, anyway. Those should just be local variables (or
special variables, in a decent language).
> Seemed to remember I had discussed fork before..
> Cameron MacKinnon seems to have a pretty descent explanation.
>
> http://groups.google.com/group/comp.lang.lisp/browse_thread/thread/252c1cebdbde94c0/793fbc9779c9a3f9?lnk=st&q=fork+John+Thingstad+group%3Acomp.lang.lisp&rnum=1&hl=en#793fbc9779c9a3f9
Cameron describes how processes share memory segments. But threads do
so, too. But forking a process to implement multithreading has the
problem that if process1 frees some chunk of heap memory, because it's
done with it, the memory lives on in process2, even though that *thread*
might not really need any of those heap contents. Threading allows one
thread to free stuff and then it's freed. Similarly for other
process-global resources, like open files. You don't have to keep track
of them globally, you simply let each thread clean up its own garbage.
Plus, on Unix, you can have an extra thread to synchronously handle
signals, so you don't have to worry about async signals all the time.
Of course if you exec() and set close-on-exec flags, most of the above
isn't a problem, but as I said it's cheaper because it's only a single
system call instead of two, and you can send whole data structures
between threads for free (no un/marshalling required). Having a single
page table makes context switching cheaper, too.
I think both approaches have their advantages, and certainly processes
have very legitimate uses, but I wouldn't use processes in situations
that just ask for a thread, now that most OSes allow the choice between
both of them.
--
If you have to ask what jazz is, you'll never know.
Louis Armstrong
From: rydis (Martin Rydstr|m) @CD.Chalmers.SE
Subject: Re: Carl Shapiro: Where is the Win32 port of CMUCL?
Date:
Message-ID: <w4c7j9za557.fsf@boris.cd.chalmers.se>
"javuchi" <·······@gmail.com> writes:
> (which I remind you: Unix is the mother of Internet, don't forget it),
Go claim that over in alt.sys.pdp10 or alt.folklore.computers and see
if it'll fly.
',mr
--
[Emacs] is written in Lisp, which is the only computer language that is
beautiful. -- Neal Stephenson, _In the Beginning was the Command Line_
From: Ulrich Hobelmann
Subject: Re: Carl Shapiro: Where is the Win32 port of CMUCL?
Date:
Message-ID: <40ri8vF1c74e7U1@individual.net>
Martin Rydstr|m wrote:
> "javuchi" <·······@gmail.com> writes:
>> (which I remind you: Unix is the mother of Internet, don't forget it),
>
> Go claim that over in alt.sys.pdp10 or alt.folklore.computers and see
> if it'll fly.
Yeah. Everybody there knows Al Gore invented the internet.
--
I wake up each morning determined to change the world...
and also to have one hell of a good time.
Sometimes that makes planning the day a little difficult.
E.B. White
Mlop wrote:
> > Unix is made by the summ of small utilities, and is the result of many,
> > many years of improvements.
>
> The problem with these tools is that they can't communicate with the outside
> world without a translator(ie. text parsing and pipping). They're autistic.
> You can hire a Chinese mathematician and an Italian biologist that are great
> at solving little tasks but when you want them to work together you have to
> hire another person to make them understand each other. There's a loss of
> productivity and flexibility right there.
Some of the console programs that come with a Unix operating system do
not work very well with each other, that's certainly true. Also, some
of them do.
The problem is that almost no other operating system comes with similar
tools.
So, although 'find' may not take a similar set of arguments to awk or
egrep, at least these programs exist. No alternatives exist for most
of them on Windows, or on other systems. If you don't use them you're
in the same state you start in with most operating systems, you've lost
nothing.
> > As far as I know, Lisp is also the summ of small utilities, and has
> > been improved for many years.
>
> No, Emacs is not the sum of small utilities. It is a large integrated system
> of many people who speak the same language. There is no translation needed.
> No text parsing, no pipping, no productivity loss.
>
> > But just take the alternative... Windows is a block, monolitic,
> > fragile, inestable in comparation.
>
> Monolithic? You must be refering to Linux.
Both are monolithic in their own way. Linux has the kernel and all the
drivers built together in a monolithic block. Windows has the kernel
and the GUI built in a monolithic block. Neither are particularly good
design.
> > And the other alternative from Apple... has become Unix, of course. OSs
> > are moving to Unix in the same way that computer languages are moving
> > to Lisp.
>
> Unix is the idiot savant. The autistic child of operating systems. You try
> to talk to it but there's no useful feedback.
Are you referring to Unix the system or the kernel itself?
On 18 Dec 2005 06:27:40 -0800, "javuchi" <·······@gmail.com> tried to
confuse everyone with this message:
>
>Giorgos Keramidas ha escrito:
>
>> On 17 Dec 2005 15:06:53 -0800, "javuchi" <·······@gmail.com> wrote:
>> > What is win32? ;-)
>>
>> 32-bit Microsoft Windows. I'm not sure if there's a 64-bit Windows port
>> and, frankly, I don't care at all to find out :)
>
>It was just a joke... ;-) I don't use Windows since several years ago,
>and asked myself it it still exists... ;-)
>Which brings me up to my mind this question: why a Lisp hacker would
>want to port CMCL or SBCL to windows, and have the responsability of
>manteining it?
>
To achieve worldwide fame?
--
|a\o/r|,-------------.,---------- Timofei Shatrov aka Grue ------------.
| m"a ||FC AMKAR PERM|| mail: grue at mail.ru http://grue3.tripod.com |
| k || PWNZ J00 || Kingdom of Loathing: Grue3 lvl 18 Seal Clubber |
`-----'`-------------'`-------------------------------------------[4*72]