My work consists mostly of writing web front-ends for things, and I can see
how I can win here with LISP (talking to stupid, inflexible, feature-poor
legacy apps on minicomputers).
The problem is, I can't find a good way of building GUIs.
Before shouting, consider my predicament:
* I can't use commercial LISPs, as the software _must_ be royalty-free, and
preferably not reliant on closed-source systems (not my stipulation, but I
can see the thinking)
* I need something that I can use relatively unchanged on Linux/*BSD/Win32
* Ideally, I'd like something that looks and feels like the target windowing
system. I'd settle for something that looks like Java Swing.
For my first "Look! This is clever stuff (and it's LISP code)!", I actually
cheated and built a Java app that talked to CLISP over network sockets.
That's fine for a simple example, but I don't want to have to think in LISP
for the logic, and then have a headache in Java for the UI.
Incidentally, I am really quite new to LISP, and other than some
head-scratching at first, I am completely sold on it. After SICP, CLtL2 and
PCL, my brain is a little fried, but for the first time in years I am
enjoying coding again.
I _really_ don't want to have to use anything else again (not for serious
coding). However, I can't see my boss being to impressed by his pretty GUI
being replaced by a REPL. :-(
I have Googled, read lots of posts here, and asked around (not that anybody
I know even uses LISP). Nothing that seems to be of any use.
Anybody got any ideas?
Jamie
Jamie Border wrote:
> Incidentally, I am really quite new to LISP, and other than some
> head-scratching at first, I am completely sold on it. After SICP,
CLtL2 and
> PCL, my brain is a little fried, but for the first time in years I am
> enjoying coding again.
Hi Jamie,
I'm intoxicated with Lisp as well. And it seems to be a nice
community. I'm still deciding if I want my app to be gtk based or web
based. Either way will be an exercise, I'm sure. And I expect to
write a lot of C-in-Lisp code. But that's ok.
Jonathon
From: Kenny Tilton
Subject: Re: A (non-negative) 'what is LISP good for?'
Date:
Message-ID: <427FC081.8040209@nyc.rr.com>
Jamie Border wrote:
> My work consists mostly of writing web front-ends for things, and I can see
> how I can win here with LISP (talking to stupid, inflexible, feature-poor
> legacy apps on minicomputers).
>
> The problem is, I can't find a good way of building GUIs.
>
> Before shouting, consider my predicament:
>
> * I can't use commercial LISPs, as the software _must_ be royalty-free, and
> preferably not reliant on closed-source systems (not my stipulation, but I
> can see the thinking)
>
> * I need something that I can use relatively unchanged on Linux/*BSD/Win32
>
> * Ideally, I'd like something that looks and feels like the target windowing
> system.
http://common-lisp.net/project/cells-gtk/
Note the implicit more general answer: check common-lisp.net when in
search of libraries. You might like some of the other GUIs offered
there, such as lTk.
Now can I shout?
kenny
--
Cells? Cello?: http://www.common-lisp.net/project/cells/
Cells-Gtk?: http://www.common-lisp.net/project/cells-gtk/
Why Lisp? http://lisp.tech.coop/RtL%20Highlight%20Film
"Doctor, I wrestled with reality for forty years, and I am happy to
state that I finally won out over it." -- Elwood P. Dowd
"Kenny Tilton" <·······@nyc.rr.com> wrote:
>> Before shouting, consider my predicament:
>>
[...snip...]
>
> http://common-lisp.net/project/cells-gtk/
>
> Note the implicit more general answer: check common-lisp.net when in
> search of libraries. You might like some of the other GUIs offered there,
> such as lTk.
>
Yes. This is good stuff.
> Now can I shout?
>
Yes. Please use the phrase "Let him use Lisp is he wants!" and direct it at
my employer.
Thanks
> kenny
>
> --
> Cells? Cello?: http://www.common-lisp.net/project/cells/
> Cells-Gtk?: http://www.common-lisp.net/project/cells-gtk/
> Why Lisp? http://lisp.tech.coop/RtL%20Highlight%20Film
>
> "Doctor, I wrestled with reality for forty years, and I am happy to state
> that I finally won out over it." -- Elwood P. Dowd
>
Jamie Border wrote:
> "Kenny Tilton" <·······@nyc.rr.com> wrote:
>
>
>>>Before shouting, consider my predicament:
>>>
>
> [...snip...]
>
>>http://common-lisp.net/project/cells-gtk/
>>
>>Note the implicit more general answer: check common-lisp.net when in
>>search of libraries. You might like some of the other GUIs offered there,
>>such as lTk.
>>
>
> Yes. This is good stuff.
>
>
>>Now can I shout?
>>
>
> Yes. Please use the phrase "Let him use Lisp is he wants!" and direct it at
> my employer.
I think that might be the wrong target for shouting.
Plan B: on your own time do something with Cells-gtk remotely sensible
for the company. You will discover, btw, that Cells do for GUIs what
Lisp does for programming. Anyway, show the Sensible Thing to the boss,
hopefully scoring oohs and ahhhs. If you fail at this, you could be in
trouble and might want to break off the attack. Then tell them it only
took you XXX minutes (subtract all learning-curve minutes). And then
tell them you did it using this new language Lisp you just learned about
from this new book, PCL. Mention that Lisp is the Latest Thing and
Slashdot was all over it last week. If they know of Lisp they will smile
condescendingly and not feel threatened. If they do not you are home
free, because everyone loves The Latest Thing.
Don't say anything about actually using Lisp for work, just rinse and
repeat the above every two-three weeks. If your employer does not start
using Lisp soon, someone will and you will have a phat portfolio to show
for it.
kt
Kenny Tilton wrote:
> >>http://common-lisp.net/project/cells-gtk/
> Plan B: on your own time do something with Cells-gtk remotely
sensible
> for the company. You will discover, btw, that Cells do for GUIs what
> Lisp does for programming. Anyway, show the Sensible Thing to the
boss,
Okay, I'm finally giving in to your little ad campaign. I have to say
I love py-gtk for Python. And I like how it works easily with
libglade.
Can you smooth talk me some more about how great cells-gtk (tm) is? I
need a bit more encouragement. :-)
Jonathon
jonathon wrote:
> Kenny Tilton wrote:
>
>>>>http://common-lisp.net/project/cells-gtk/
>>
>>Plan B: on your own time do something with Cells-gtk remotely
>
> sensible
>
>>for the company. You will discover, btw, that Cells do for GUIs what
>>Lisp does for programming. Anyway, show the Sensible Thing to the
>
> boss,
>
> Okay, I'm finally giving in to your little ad campaign.
Ad campaign? Actually, I do not care what people use, I am just trying
to help folks out who are looking for something I know where to find.
Sh*t, I use Cello, not Cells-Gtk. Cello was supposed to answer the Lisp
newbie Question #2: Where is the GUI? I think Cells-Gtk has won that honor.
> I have to say
> I love py-gtk for Python. And I like how it works easily with
> libglade.
You have either mentioned that twice or are the second to mention it. I
do not get it. Is this not a Gtk issue? Or does py-gtk somehow use
libglade where Gtk normally does not? Or they have done extensive
bindings for libglade which you can use from Python... no, hang on, then
it is not portable -- or has libglade been ported everywhere? Other?
>
> Can you smooth talk me some more about how great cells-gtk (tm) is? I
> need a bit more encouragement. :-)
yes and no. no because I do not use the thing and can only report two
enticing comments made by others:
- vasilis himself (the author) concluded Cells made GUI building
significantly easier. cannot recall the exact words. Gtk was just the
vehicle for his experimentation. And Vasilis seems to be one hell of a
programmer, so you might want to give this one a little extra weight.
- one of the cells-gtk users felt Gtk look-feel was superior to the
other portable GUI option, CAPI.
- as for Cells and GUIs, a little software shop called Adobe is
reporting huge productivity gains from their parallel Adam/Eve hacks. I
doubt Adam is anywhere as powerful as Cells, so you can out-Adobe Adobe.
- as for constraints and GUI, talk to any happy Garnet user. Then there
is COSI, etc etc.
- Cells solves the problem (data interdependency) Brooks thought made
impossible the development of a silver bullet.
etc etc
kt
--
Cells? Cello?: http://www.common-lisp.net/project/cells/
Cells-Gtk?: http://www.common-lisp.net/project/cells-gtk/
Why Lisp? http://lisp.tech.coop/RtL%20Highlight%20Film
"Doctor, I wrestled with reality for forty years, and I am happy to
state that I finally won out over it." -- Elwood P. Dowd
Kenny Tilton wrote:
> Ad campaign? Actually, I do not care what people use, I am just
trying
> to help folks out who are looking for something I know where to find.
> Sh*t, I use Cello, not Cells-Gtk. Cello was supposed to answer the
Lisp
> newbie Question #2: Where is the GUI? I think Cells-Gtk has won that
honor.
Okay, fair enough. And what is question #1? :-)
> You have either mentioned that twice or are the second to mention it.
I
> do not get it. Is this not a Gtk issue? Or does py-gtk somehow use
> libglade where Gtk normally does not? Or they have done extensive
> bindings for libglade which you can use from Python... no, hang on,
then
> it is not portable -- or has libglade been ported everywhere? Other?
I really don't care about portability outside POSIX. I actually run
DragonFlyBSD, and CMUCL on top of that. But it's nice to build a GUI
with glade-2, save it, and then load each widget dynamically. And
that's a nice feature about libglade/py-gtk. The problem is, as great
as Python is, I've _still_ found it bloated in some respects, where
Lisp is so much more elegant.
> yes and no. no because I do not use the thing and can only report two
> enticing comments made by others:
So why do you choose to promote it?
> - vasilis himself (the author) concluded Cells made GUI building
> significantly easier. cannot recall the exact words. Gtk was just the
> vehicle for his experimentation. And Vasilis seems to be one hell of
a
> programmer, so you might want to give this one a little extra weight.
I am very much game for learning from others' experience.
> - Cells solves the problem (data interdependency) Brooks thought made
> impossible the development of a silver bullet.
My application is very spreadsheet/calculation based, so I'll have to
check it out.
Jonathon
jonathon wrote:
> Kenny Tilton wrote:
>
>>Ad campaign? Actually, I do not care what people use, I am just
>
> trying
>
>>to help folks out who are looking for something I know where to find.
>
>
>>Sh*t, I use Cello, not Cells-Gtk. Cello was supposed to answer the
>
> Lisp
>
>>newbie Question #2: Where is the GUI? I think Cells-Gtk has won that
>
> honor.
>
> Okay, fair enough. And what is question #1? :-)
"how come when I (delete 'b '(a b c))....?"
>
>
>>You have either mentioned that twice or are the second to mention it.
>
> I
>
>>do not get it. Is this not a Gtk issue? Or does py-gtk somehow use
>>libglade where Gtk normally does not? Or they have done extensive
>>bindings for libglade which you can use from Python... no, hang on,
>
> then
>
>>it is not portable -- or has libglade been ported everywhere? Other?
>
>
> I really don't care about portability outside POSIX. I actually run
> DragonFlyBSD, and CMUCL on top of that. But it's nice to build a GUI
> with glade-2, save it, and then load each widget dynamically.
I do not really follow, but I think I do. Sounds like the GUI Builder
Debate. Static widgets are of so little interest to me. As someone else
said, any decent GUI responds dynamically to the context. The whole GUI
bUilder thing is a defensive, pathetic reaction to "oh, gosh, GUIs are
so hard to build". Turns out GUI Builders are a false minimum.
> And
> that's a nice feature about libglade/py-gtk. The problem is, as great
> as Python is, I've _still_ found it bloated in some respects, where
> Lisp is so much more elegant.
Oh, stop sucking up.
>
>
>>yes and no. no because I do not use the thing and can only report two
>
>
>>enticing comments made by others:
>
>
> So why do you choose to promote it?
Uh, because you asked? What part of "I am just trying to help the
clueless" do you not understand?
> My application is very spreadsheet/calculation based, so I'll have to
> check it out.
Oh, no, you definitely want McCLIM.
kt
--
Cells? Cello?: http://www.common-lisp.net/project/cells/
Cells-Gtk?: http://www.common-lisp.net/project/cells-gtk/
Why Lisp? http://lisp.tech.coop/RtL%20Highlight%20Film
"Doctor, I wrestled with reality for forty years, and I am happy to
state that I finally won out over it." -- Elwood P. Dowd
Kenny Tilton wrote:
> I do not really follow, but I think I do. Sounds like the GUI Builder
> Debate. Static widgets are of so little interest to me. As someone
else
> said, any decent GUI responds dynamically to the context. The whole
GUI
> bUilder thing is a defensive, pathetic reaction to "oh, gosh, GUIs
are
> so hard to build". Turns out GUI Builders are a false minimum.
Well, I've found that glade is pretty good at making gtk dialogs. It's
nice to not have to worry about all the code for margins, layout, and
so on. But yes, dynamic dialogs are the way to go. Several of my
dialogs have empty vboxes or hboxes where I can add stuff I need, and
others have combo boxes and such that have lists built on the fly.
It's really a great tool. But as I said, I'm bored with Python now,
and I want a new toy to play with. :-)
From: Kenny Tilton
Subject: Road to Cells-Gtk Survey [was Re: A (non-negative) 'what is LISP good for?']
Date:
Message-ID: <uK1ge.18703$n93.1820@twister.nyc.rr.com>
jonathon wrote:
> Kenny Tilton wrote:
>
>>I do not really follow, but I think I do. Sounds like the GUI Builder
>
>
>>Debate. Static widgets are of so little interest to me. As someone
>
> else
>
>>said, any decent GUI responds dynamically to the context. The whole
>
> GUI
>
>>bUilder thing is a defensive, pathetic reaction to "oh, gosh, GUIs
>
> are
>
>>so hard to build". Turns out GUI Builders are a false minimum.
>
>
> Well, I've found that glade is pretty good at making gtk dialogs. It's
> nice to not have to worry about all the code for margins, layout, and
> so on.
Agreed. But with something like Cells driving a GUI, one does not think
much about layout when coding. I just say gimme a row, this spacing,
this surrounding margin, and then things fall where they may. if I hate
it I close the dialog, tweak the parameters in the source (with the app
still running), recompile and reopen the dialog.
> But yes, dynamic dialogs are the way to go. Several of my
> dialogs have empty vboxes or hboxes where I can add stuff I need, and
> others have combo boxes and such that have lists built on the fly.
> It's really a great tool. But as I said, I'm bored with Python now,
> and I want a new toy to play with. :-)
>
One of the things Vasilis did in creating cells-gtk was to include an
exhaustive set of demos in which he touched on all the Gtk widgets /and/
created classic Cell-ish interfaces which reacted dynamically (widget
contents as well as layout) to user selections and options. If you stare
at the code (or Celtik (+ Ltk Cells) from which Cells-Gtk was derived)
you will also see that Cells further automates a ton of the information
exchange with Gtk, so you can forget that while building your GUIs.
Hmmm. Come to think of it, it might be interesting to hear from folks
more knowledgable than me who have used both Cells-Gtk and XXX-Gtk,
ideally with lambdaGtk as a control.
Or should we have a code shoot-out? Experts in each producing the same
dialog, perhaps one of Vasilis's demos so the Cells-Gtk team does not
have to do any work. The Uber-control could be Cello, or, while we are
at it, any Lisp GUI pretender.
The rules are: one source, three screenshots (win32, Mac OS X, Linux).
kt
--
Cells? Cello?: http://www.common-lisp.net/project/cells/
Cells-Gtk?: http://www.common-lisp.net/project/cells-gtk/
Why Lisp? http://lisp.tech.coop/RtL%20Highlight%20Film
"Doctor, I wrestled with reality for forty years, and I am happy to
state that I finally won out over it." -- Elwood P. Dowd
From: David Golden
Subject: Re: Road to Cells-Gtk Survey [was Re: A (non-negative) 'what is LISP good for?']
Date:
Message-ID: <I6tge.53010$Z14.43952@news.indigo.ie>
Kenny Tilton wrote:
> One of the things Vasilis did in creating cells-gtk was to include an
> exhaustive set of demos in which he touched on all the Gtk widgets
Hm. Does that include GtkGLArea or maybe even GtkCairo [1] widgets? Now
that would get interesting.
[1] http://cvs.cairographics.org/gtkcairo/
"jonathon" <···········@bigfoot.com> writes:
> Can you smooth talk me some more about how great cells-gtk (tm) is? I
> need a bit more encouragement. :-)
A picture is worth a thousand words. Here are pictures worth 9,000
words:
http://common-lisp.net/project/cells-gtk/screenshots.html
Paolo
--
Why Lisp? http://lisp.tech.coop/RtL%20Highlight%20Film
Recommended Common Lisp libraries/tools (see also http://clrfi.alu.org):
- ASDF/ASDF-INSTALL: system building/installation
- CL-PPCRE: regular expressions
- UFFI: Foreign Function Interface
Good advice.
I am pleasantly surprised that several people from c.l.l have been friendly
and tolerant towards a relative newcomer.
Nice to see that the folks here are still software engineers and not the
bigoted fools that lurk elsewhere....
Jamie
>
> Jamie Border wrote:
>
>> "Kenny Tilton" <·······@nyc.rr.com> wrote:
>>
>>
>>>>Before shouting, consider my predicament:
>>>>
>>
>> [...snip...]
>>
>>>http://common-lisp.net/project/cells-gtk/
>>>
>>>Note the implicit more general answer: check common-lisp.net when in
>>>search of libraries. You might like some of the other GUIs offered there,
>>>such as lTk.
>>>
>>
>> Yes. This is good stuff.
>>
>>
>>>Now can I shout?
>>>
>>
>> Yes. Please use the phrase "Let him use Lisp is he wants!" and direct it
>> at my employer.
>
> I think that might be the wrong target for shouting.
>
> Plan B: on your own time do something with Cells-gtk remotely sensible for
> the company. You will discover, btw, that Cells do for GUIs what Lisp does
> for programming. Anyway, show the Sensible Thing to the boss, hopefully
> scoring oohs and ahhhs. If you fail at this, you could be in trouble and
> might want to break off the attack. Then tell them it only took you XXX
> minutes (subtract all learning-curve minutes). And then tell them you did
> it using this new language Lisp you just learned about from this new book,
> PCL. Mention that Lisp is the Latest Thing and Slashdot was all over it
> last week. If they know of Lisp they will smile condescendingly and not
> feel threatened. If they do not you are home free, because everyone loves
> The Latest Thing.
>
> Don't say anything about actually using Lisp for work, just rinse and
> repeat the above every two-three weeks. If your employer does not start
> using Lisp soon, someone will and you will have a phat portfolio to show
> for it.
>
> kt
>
From: Bulent Murtezaoglu
Subject: Re: A (non-negative) 'what is LISP good for?'
Date:
Message-ID: <87fyww5kf7.fsf@p4.internal>
>>>>> "JB" == Jamie Border <·····@jborder.com> writes:
[...]
JB> * I can't use commercial LISPs, as the software _must_ be
JB> royalty-free, and preferably not reliant on closed-source
JB> systems (not my stipulation, but I can see the thinking)
JB> * I need something that I can use relatively unchanged on
JB> Linux/*BSD/Win32
JB> * Ideally, I'd like something that looks and feels like the
JB> target windowing system. I'd settle for something that looks
JB> like Java Swing.
Lispworks+CAPI does all this, but will need the Linux compatibility
module under FreeBSD (I don't know about other BSD's. ask them).
It wouldn't hurt to talk to Franz either.
[...]
JB> I have Googled, read lots of posts here, and asked around (not
JB> that anybody I know even uses LISP). Nothing that seems to be
JB> of any use.
I hope the above helps.
cheers,
BM
(I'm climbing into my asbestos suit as I write...)
I dream of a LISP that...
1) Is open-source
2) Is cross-platform
3) Can deliver some kind of GUI (on X and Win32)
I can't find one (and I've looked).
Maybe this puts people off learning LISP. It nearly did the same to me.
It took me quite a while before I realised that learning LISP would be the
last language I would need to learn. After a couple of days reading (On
Lisp, SICP, PCL), I realised that I would rather use LISP than anything
else.
Now I can't build a GUI without using a commercial LISP.
I am currently poor, but am _still_ considering the $1100 for a non-crippled
LispWorks because I feel that strongly about the need for GUI.
Richard Stallman and his crowd have given us an excellent free C
compiler[1]. With the glut of [free] libraries that let me produce
X-Windows applications, I can give somebody a list of URLs and they can
teach themselves how to use these. Then they can write useful software.
I can download Python and build some pretty little applications.
I can download Java with a fairly good IDE and start developing straight
away.
How hard would it be to provide something like this for an open-source LISP?
Would it help people's perception of LISP?
I think so.
Jamie
Jamie Border wrote:
> (I'm climbing into my asbestos suit as I write...)
>
> I dream of a LISP that...
>
> 1) Is open-source
> 2) Is cross-platform
> 3) Can deliver some kind of GUI (on X and Win32)
>
> I can't find one (and I've looked).
>
> Maybe this puts people off learning LISP. It nearly did the same to me.
>
> It took me quite a while before I realised that learning LISP would be the
> last language I would need to learn. After a couple of days reading (On
> Lisp, SICP, PCL), I realised that I would rather use LISP than anything
> else.
>
> Now I can't build a GUI without using a commercial LISP.
Check out Foil and JFLI.
--
Marco
>I dream of a LISP that...
>
>1) Is open-source
>2) Is cross-platform
>3) Can deliver some kind of GUI (on X and Win32)
>
>I can't find one (and I've looked).
http://home.comcast.net/~bc19191/blog/040109.html
and
http://home.comcast.net/~bc19191/blog/040111.html
will give you some suggestion (and it is actually my vote)
MrMathematica wrote:
>>I dream of a LISP that...
>>
>>1) Is open-source
>>2) Is cross-platform
>>3) Can deliver some kind of GUI (on X and Win32)
>>
>>I can't find one (and I've looked).
>
>
> http://home.comcast.net/~bc19191/blog/040109.html
>
> and
>
> http://home.comcast.net/~bc19191/blog/040111.html
>
> will give you some suggestion (and it is actually my vote)
>
Nonsense. Scheme is not Common Lisp. Folks looking for Lisp are looking
for Common Lisp, because the last thing they want is a two-page spec
with everything else unspecified such that nothing works with anything.
kt
--
Cells? Cello?: http://www.common-lisp.net/project/cells/
Cells-Gtk?: http://www.common-lisp.net/project/cells-gtk/
Why Lisp? http://lisp.tech.coop/RtL%20Highlight%20Film
"Doctor, I wrestled with reality for forty years, and I am happy to
state that I finally won out over it." -- Elwood P. Dowd
Kenny Tilton wrote:
> MrMathematica wrote:
>
> >>I dream of a LISP that...
> >>
> >>1) Is open-source
> >>2) Is cross-platform
> >>3) Can deliver some kind of GUI (on X and Win32)
> >>
> >>I can't find one (and I've looked).
> >
> >
> > http://home.comcast.net/~bc19191/blog/040109.html
> >
> > and
> >
> > http://home.comcast.net/~bc19191/blog/040111.html
> >
> > will give you some suggestion (and it is actually my vote)
> >
>
> Nonsense. Scheme is not Common Lisp. Folks looking for Lisp are
looking
> for Common Lisp, because the last thing they want is a two-page spec
> with everything else unspecified such that nothing works with
anything.
>
> kt
>
Nonsense.
The Scheme spec is short and left much things unspecified. But the
implementation is NOT. A Scheme implementation can extend as big as it
want, define all those "unspecified" and provide huge libraries. So
there even can be a Scheme implementation bigger than Common Lisp.
And I have to quote from
http://home.comcast.net/~bc19191/blog/040111.html:
Although PLT Scheme is a Scheme and not a Common Lisp implementation,
it includes a lot of standard functionality that is normally found in a
Common Lisp implementation. Therefore, CL advocates who would normally
dismiss a Scheme implementation from consideration should think
seriously about what they need in an Open Source Lisp implementation
before they reject PLT Scheme.
MrMathematica wrote:
> The Scheme spec is short and left much things unspecified. But the
> implementation is NOT. A Scheme implementation can extend as big as it
> want, define all those "unspecified" and provide huge libraries. So
> there even can be a Scheme implementation bigger than Common Lisp.
If you have to switch from one CL implementation to another, it is very
likely that you can keep a considerably large part of your source code
unchanged. This even gives you some interesting deployment options, like
developing in one CL imp and delivering in another. I don't think you
have the same degree of choice in Scheme (or most other languages for
that matter).
Pascal
--
2nd European Lisp and Scheme Workshop
July 26 - Glasgow, Scotland - co-located with ECOOP 2005
http://lisp-ecoop05.bknr.net/
Pascal Costanza wrote:
> If you have to switch from one CL implementation to another, it is
very
> likely that you can keep a considerably large part of your source
code
> unchanged. This even gives you some interesting deployment options,
like
> developing in one CL imp and delivering in another. I don't think you
> have the same degree of choice in Scheme (or most other languages for
> that matter).
>
>
> Pascal
>
One usual explanation for why Lisp (include Scheme) lost while language
such as Perl or Python gain more users is that there are too many
implementations.
For this thread in particular, Jamie Border seems to need only one
implementation.
Zhu Chongkai
Lisp Implies Smart Programming
On 2005-05-10 08:28:48 -0400, "MrMathematica" <···········@citiz.net> said:
> One usual explanation for why Lisp (include Scheme) lost while language
> such as Perl or Python gain more users is that there are too many
> implementations.
Though this explanation may be common it is clearly false else C would
have been a dismal failure - there have been dozens of implementations
of C.
Lisp "failed" for the same reason that Henckels knives have "failed" to
outsell the Ginsu - higer quality tools that require greater care and
knowlege to take advantage of their greater power will always be
outsold by cheap knock offs.[1]
In this day of powerful scripting languages with garbage collection the
one thing that sets lisp apart is macros, which, not coincidentally, is
the same thing that necessitates lisp's unusual highly parenthesised
syntax. To use the full power of lisp means using macros effectively,
and using macros effectively is something that only the top 10% of
programmers will ever learn. Therefore, lisp will always be a language
for that subset of programmers with the ability and motivation to
master macros.
From the perspective that equates number of users with "success" lisp
will always be a "failure." From the perspective that equates
programming power, ease and speed of development with "success" lisp
has been a success for decades.
Finally, to return to the lisp scheme comparison, more implementation
choices are a good thing, because we can never be sure when a
particular implementation will cease to be supported or fall behind
other implementations in features, ease of use, etc. Nor can we know in
advance whether or when we may need to port some work to another
platform where our current implementation does not exist or is not as
well supported. Therfore, the greater the amount of common
functionality supported by *all* lisp implementations, the easier the
transion between implementations.
[1] Ever try to do a chiffonade with a Ginsu? The results are not pretty.
MrMathematica wrote:
> Pascal Costanza wrote:
>
>>If you have to switch from one CL implementation to another, it is
>
> very
>
>>likely that you can keep a considerably large part of your source
>
> code
>
>>unchanged. This even gives you some interesting deployment options,
>
> like
>
>>developing in one CL imp and delivering in another. I don't think you
>
>
>>have the same degree of choice in Scheme (or most other languages for
>
>
>>that matter).
>>
>>
>>Pascal
>>
>
>
> One usual explanation for why Lisp (include Scheme) lost while language
> such as Perl or Python gain more users is that there are too many
> implementations.
Partly true, apart from the fact that Common Lisp is, IMHO gaining users.
>
> For this thread in particular, Jamie Border seems to need only one
> implementation.
>
That is his problem, not mine. He can do things with one implementation
using some specialized libraries.
Cheers
--
Marco
MrMathematica wrote:
> One usual explanation for why Lisp (include Scheme) lost while language
> such as Perl or Python gain more users is that [...]
Let's discuss this again in 50 years and see which languages will still
be around and which will be not even known anymore by then... ;)
Pascal
--
2nd European Lisp and Scheme Workshop
July 26 - Glasgow, Scotland - co-located with ECOOP 2005
http://lisp-ecoop05.bknr.net/
>> One usual explanation for why Lisp (include Scheme) lost while language
>> such as Perl or Python gain more users is that [...]
> Let's discuss this again in 50 years and see which languages will
> still be around and which will be not even known anymore by then... ;)
See you on earth.comp.lang.fortran then!
joelh
MrMathematica wrote:
> Kenny Tilton wrote:
>>
>>Nonsense. Scheme is not Common Lisp. Folks looking for Lisp are
>
> looking
>
>>for Common Lisp, because the last thing they want is a two-page spec
>>with everything else unspecified such that nothing works with
>
> anything.
>
>>kt
>>
>
>
> Nonsense.
>
> The Scheme spec is short and left much things unspecified.
Exactly. The main design goal of Scheme is to keep the spec short.
Things like object system, multidimensional arrays and a godzillion
other useful things are non-standard.
But the
> implementation is NOT.
Which implementation?
A Scheme implementation can extend as big as it
> want, define all those "unspecified" and provide huge libraries. So
> there even can be a Scheme implementation bigger than Common Lisp.
But in doing so it becomes inherently incompatible with other
implementations to a much larger extent than a Common Lisp
implementation does. Besides, a CL implementation has a much wider and
firmer foundation to build upon.
> And I have to quote from
> http://home.comcast.net/~bc19191/blog/040111.html:
>
> Although PLT Scheme is a Scheme and not a Common Lisp implementation,
> it includes a lot of standard functionality that is normally found in a
> Common Lisp implementation. Therefore, CL advocates who would normally
> dismiss a Scheme implementation from consideration should think
> seriously about what they need in an Open Source Lisp implementation
> before they reject PLT Scheme.
>
You are totally missing the point. Common Lisp advocates advocate a
large language spec. You are advocating an implemetation. My code
works on CMUCL, SBCL, LW, ACL, CLISP and other implementations I did not
have the honor to test. Your code works on PLT Scheme.
Cheers
--
Marco
Marco Antoniotti wrote:
>
>
> MrMathematica wrote:
>
>> Kenny Tilton wrote:
>>
>>>
>>> Nonsense. Scheme is not Common Lisp. Folks looking for Lisp are
>>
>>
>> looking
>>
>>> for Common Lisp, because the last thing they want is a two-page spec
>>> with everything else unspecified such that nothing works with
>>
>>
>> anything.
>>
>>> kt
>>>
>>
>>
>> Nonsense.
>>
>> The Scheme spec is short and left much things unspecified.
>
>
> Exactly. The main design goal of Scheme is to keep the spec short.
> Things like object system, multidimensional arrays and a godzillion
> other useful things are non-standard.
>
> But the
>
>> implementation is NOT.
>
>
> Which implementation?
>
> A Scheme implementation can extend as big as it
>
>> want, define all those "unspecified" and provide huge libraries. So
>> there even can be a Scheme implementation bigger than Common Lisp.
>
>
> But in doing so it becomes inherently incompatible with other
> implementations to a much larger extent than a Common Lisp
> implementation does. Besides, a CL implementation has a much wider and
> firmer foundation to build upon.
>
>
>> And I have to quote from
>> http://home.comcast.net/~bc19191/blog/040111.html:
>>
>> Although PLT Scheme is a Scheme and not a Common Lisp implementation,
>> it includes a lot of standard functionality that is normally found in a
>> Common Lisp implementation. Therefore, CL advocates who would normally
>> dismiss a Scheme implementation from consideration should think
>> seriously about what they need in an Open Source Lisp implementation
>> before they reject PLT Scheme.
>>
>
> You are totally missing the point. Common Lisp advocates advocate a
> large language spec. You are advocating an implemetation. My code
> works on CMUCL, SBCL, LW, ACL, CLISP and other implementations I did not
> have the honor to test. Your code works on PLT Scheme.
No, you are missing /his/ point: PLT Scheme is now a language, one that
can best be characterized as Common Lisp without all the libraries with
all the libraries. It has the Python/Ruby/Perl advantage of being
defined by a single implementation, and the Python/Ruby/Perl
disadvantage of needing years more development to achieve the polish of
Common Lisp. The cherished itsy bitsy teeny weeny spec lives on, because
the fat implementation standardizes things without spec-ifying them.
I hope this clears everything up.
:)
kenny
--
Cells? Cello?: http://www.common-lisp.net/project/cells/
Cells-Gtk?: http://www.common-lisp.net/project/cells-gtk/
Why Lisp? http://lisp.tech.coop/RtL%20Highlight%20Film
"Doctor, I wrestled with reality for forty years, and I am happy to
state that I finally won out over it." -- Elwood P. Dowd
Kenny Tilton wrote:
> The cherished itsy bitsy teeny weeny spec lives on, because
> the fat implementation standardizes things without spec-ifying them.
Hey! That may hold for Perl/Python, but the PLT documentation is quite
thorough (which shouldn't be a surprise considering who is behind it).
--
Jens Axel Søgaard
Jens Axel S�gaard wrote:
> Kenny Tilton wrote:
>
>> The cherished itsy bitsy teeny weeny spec lives on, because the fat
>> implementation standardizes things without spec-ifying them.
>
>
> Hey! That may hold for Perl/Python, but the PLT documentation is quite
> thorough (which shouldn't be a surprise considering who is behind it).
>
I say "spec, uou say "documentation. I say "po-tay-to", you say
"cu-cum-ber"...
OK, I was wrong. The i.b.t.w. spec does /not/ live on, because
documentation of a language defined by a single implementation
constitutes a spec. Who is going to break this to MrMathematica?
:)
kt
--
Cells? Cello?: http://www.common-lisp.net/project/cells/
Cells-Gtk?: http://www.common-lisp.net/project/cells-gtk/
Why Lisp? http://lisp.tech.coop/RtL%20Highlight%20Film
"Doctor, I wrestled with reality for forty years, and I am happy to
state that I finally won out over it." -- Elwood P. Dowd
Kenny Tilton wrote:
> Nonsense. Scheme is not Common Lisp. Folks looking for Lisp are
looking
> for Common Lisp, because the last thing they want is a two-page spec
> with everything else unspecified such that nothing works with
anything.
Sorry Kenny:
But that is simply not true. You keep up claiming such unfounded
slandering.
Btw: please show me your "nothing/everything works with anything" under
CommonLisp when it comes to platform specific issues.
Granded, the Scheme language spec is tiny. However, there are some
Scheme implementations which go far, far, far beyond that particular
standard.
Dr.Scheme for example works on many platforms and features some heavy
loaded additional libraries (data bases, GUIs, etc.)
Or consider Bigloo as an example. Personally: I have installed Bigloo
on Mac OSX, Sun OS and Linux.
CommonLisp tried to find a least common denominator. But the latter is
all what can be said about CommonLisp. I haven't understood the reasons
why one would prefer CommonLisp over Scheme. What will CommonLisp
deliver me what Bigloo, or Dr. Scheme or Chicken, etc. cannot deliver
me on my platforms: Mac OSX, Linux, Unix.
Schneewittchen
F�rster vom Silberwald wrote:
> Kenny Tilton wrote:
>
>
>>Nonsense. Scheme is not Common Lisp. Folks looking for Lisp are
>
> looking
>
>>for Common Lisp, because the last thing they want is a two-page spec
>>with everything else unspecified such that nothing works with
>
> anything.
>
> Sorry Kenny:
>
> But that is simply not true. You keep up claiming such unfounded
> slandering.
>
> Btw: please show me your "nothing/everything works with anything" under
> CommonLisp when it comes to platform specific issues.
>
> Granded, the Scheme language spec is tiny. However, there are some
> Scheme implementations which go far, far, far beyond that particular
> standard.
Mostly incompatibly.
> Dr.Scheme for example works on many platforms and features some heavy
> loaded additional libraries (data bases, GUIs, etc.)
So does LW, or CMUCL, or ACL, or CLISP or any of the available CL
implementations. But the amount of code you can share among them is far
greater.
>
> Or consider Bigloo as an example. Personally: I have installed Bigloo
> on Mac OSX, Sun OS and Linux.
>
> CommonLisp tried to find a least common denominator. But the latter is
> all what can be said about CommonLisp. I haven't understood the reasons
> why one would prefer CommonLisp over Scheme. What will CommonLisp
> deliver me what Bigloo, or Dr. Scheme or Chicken, etc. cannot deliver
> me on my platforms: Mac OSX, Linux, Unix.
Here is the rub. You continue to mix the language and the
implementations. With Common Lisp you are talking about a language,
with Dr. Scheme, BigLoo, Chicken, etc etc you are talking about
implementations.
If you take any one implementation of Common Lisp, it will provide you
with all the bells and whistles you have in one of the "Scheme" (quotes
mandatory) implementations you mention. Plus, it will guarantee you
that your STANDARD object oriented code WILL run on other
implementations as well. You simply do not have this guarantee in
Scheme as OO is NOT in the spec.
Cheers
--
Marco
Marco Antoniotti wrote:
> Here is the rub. You continue to mix the language and the
> implementations. With Common Lisp you are talking about a language,
> with Dr. Scheme, BigLoo, Chicken, etc etc you are talking about
> implementations.
>
> If you take any one implementation of Common Lisp, it will provide
you
> with all the bells and whistles you have in one of the "Scheme"
(quotes
> mandatory) implementations you mention. Plus, it will guarantee you
> that your STANDARD object oriented code WILL run on other
> implementations as well. You simply do not have this guarantee in
> Scheme as OO is NOT in the spec.
>
I am aware of such a distinction. However, I find it always unfair from
Kenny that he insists that Scheme is solely based on the tiny Scheme
standard. And that is simply not true. I think an outsider not familar
with Scheme gets a distorted view of the matter.
That said: So, Scheme simply consists of a tiny unuseful standard -
right? Okay, then I for one will everyone tell in the future that
CommonLisp features as a data structure only lists; if you are after
arrays or hash-tables you should move on to C++ or Java.
Telefonhörer
F�rster vom Silberwald wrote:
> Marco Antoniotti wrote:
>
>
>>Here is the rub. You continue to mix the language and the
>>implementations. With Common Lisp you are talking about a language,
>>with Dr. Scheme, BigLoo, Chicken, etc etc you are talking about
>>implementations.
>>
>>If you take any one implementation of Common Lisp, it will provide
>
> you
>
>>with all the bells and whistles you have in one of the "Scheme"
>
> (quotes
>
>>mandatory) implementations you mention. Plus, it will guarantee you
>>that your STANDARD object oriented code WILL run on other
>>implementations as well. You simply do not have this guarantee in
>>Scheme as OO is NOT in the spec.
>
> I am aware of such a distinction. However, I find it always unfair from
> Kenny that he insists that Scheme is solely based on the tiny Scheme
> standard. And that is simply not true. I think an outsider not familar
> with Scheme gets a distorted view of the matter.
But it is true!
> That said: So, Scheme simply consists of a tiny unuseful standard -
> right? Okay, then I for one will everyone tell in the future that
> CommonLisp features as a data structure only lists; if you are after
> arrays or hash-tables you should move on to C++ or Java.
But that is wrong!
Here is an example of why Scheme may be not enough. I have recently
investigated a paper about first-class extents - see
http://library.readscheme.org/servlets/search.ss?pattern=Lee+Shin-Der -
because that's related to my current research. There is the code for
their language extension provided as an appendix, and it's written in
Scheme.
No wait, it's written in Chez Scheme and uses an implementation-specific
macro system. So instead of using one of the Scheme implementations I
have already installed on my system, I had to download Chez Scheme and
install that, figure out how to make it work with quack, etc., which
already took quite a while, only to find out that the code doesn't work.
Why? Apparently, Chez Scheme has changed the semantics of that macro
system sometime during the last 15 years. (The paper is from the
beginning of the 90's.) So I have translated the code to use
syntax-rules instead, which took me quite a while. This was _not_ what I
was interested in.
With Common Lisp, it's much more likely that code from the past just
works (tm). To me, that's important!
If that's not important to you, then be happy. But others have different
requirements. The stuff I am currently working on runs on at least 4-5
or more different CL implementations. You can find lots of libraries on
the various CL websites that are similarly highly portable.
Pascal
--
2nd European Lisp and Scheme Workshop
July 26 - Glasgow, Scotland - co-located with ECOOP 2005
http://lisp-ecoop05.bknr.net/
F�rster vom Silberwald wrote:
> Marco Antoniotti wrote:
>
>
>>Here is the rub. You continue to mix the language and the
>>implementations. With Common Lisp you are talking about a language,
>>with Dr. Scheme, BigLoo, Chicken, etc etc you are talking about
>>implementations.
>>
>>If you take any one implementation of Common Lisp, it will provide
>
> you
>
>>with all the bells and whistles you have in one of the "Scheme"
>
> (quotes
>
>>mandatory) implementations you mention. Plus, it will guarantee you
>>that your STANDARD object oriented code WILL run on other
>>implementations as well. You simply do not have this guarantee in
>>Scheme as OO is NOT in the spec.
>>
>
>
> I am aware of such a distinction. However, I find it always unfair from
> Kenny that he insists that Scheme is solely based on the tiny Scheme
> standard. And that is simply not true. I think an outsider not familar
> with Scheme gets a distorted view of the matter.
I am sorry, but I think Kenny is right on this. If your implementation
gets tons of bells and whistles, that's what it is: implementation X
with tons of bells and whistles. The bells and whistles can be darned
good, but that does not change the basic argument.
> That said: So, Scheme simply consists of a tiny unuseful standard -
> right? Okay, then I for one will everyone tell in the future that
> CommonLisp features as a data structure only lists; if you are after
> arrays or hash-tables you should move on to C++ or Java.
I may not have understood that sentence. I assume you know that any
Common Lisp implementation has multidimensional, (almost) sliceable and
extensible arrays out of the box. Otherwise it would not be a Common
Lisp implementation.
Cheers
--
Marco
F�rster vom Silberwald wrote:
> Kenny Tilton wrote:
>
>
>>Nonsense. Scheme is not Common Lisp. Folks looking for Lisp are
>
> looking
>
>>for Common Lisp, because the last thing they want is a two-page spec
>>with everything else unspecified such that nothing works with
>
> anything.
>
> Sorry Kenny:
>
> But that is simply not true. You keep up claiming such unfounded
> slandering.
Calm down. Schemers agree that the "roll-your own environment"
philosophy of Scheme makes it hard for Schemers to share code. Go argue
with them, because their assent on this question emboldens my claim.
>
> Btw: please show me your "nothing/everything works with anything" under
> CommonLisp when it comes to platform specific issues.
If we had not had to put that stick between your teeth I might be able
to understand you. I am guessing you mean CL porting is not that easy.
Exactly. And that is in a language that has a standard object system,
amongst a kazillion other things. (You guys are the ones that like to
make fun of CL's big spec, which is now kicking your ass.)
Having done a lot of x-platform and x-implementation over the past
couple of years in a big-spec language, I would not even think of
attempting that in Scheme.
>
> Granded, the Scheme language spec is tiny. However, there are some
> Scheme implementations which go far, far, far beyond that particular
> standard.
>
> Dr.Scheme for example works on many platforms and features some heavy
> loaded additional libraries (data bases, GUIs, etc.)
>
> Or consider Bigloo as an example. Personally: I have installed Bigloo
> on Mac OSX, Sun OS and Linux.
It sounds as if you Schemers have conceded defeat on the tiny-spec thing
and have fallen back to "My chosen implementation is actually a
big-spec, one-implementation, OS-portable language."
To be fair, if I am just Joe Developer looking for a way to deliver an
application on multiple platforms, I can just pick one of the
one-implementation Scheme "languages" and forget the others. No argument
there.
The bad news for Scheme the language, unless one implementation crushes
the others, is that you now need to start a Common Scheme ANSI
standardization process to undo the fragmentation of your shrinking
community. The good news is that you can simply vote to adopt the Common
Lisp spec and be done before the first round of coffee and pastries is
gone. Failing that, the bad news is that you are now twenty years behind.
kenny
ps. which Scheme implementation gets your vote?
--
Cells? Cello?: http://www.common-lisp.net/project/cells/
Cells-Gtk?: http://www.common-lisp.net/project/cells-gtk/
Why Lisp? http://lisp.tech.coop/RtL%20Highlight%20Film
"Doctor, I wrestled with reality for forty years, and I am happy to
state that I finally won out over it." -- Elwood P. Dowd
Kenny Tilton wrote:
> If we had not had to put that stick between your teeth I might be
able
> to understand you. I am guessing you mean CL porting is not that
easy.
> Exactly. And that is in a language that has a standard object system,
> amongst a kazillion other things. (You guys are the ones that like to
> make fun of CL's big spec, which is now kicking your ass.)
>
> Having done a lot of x-platform and x-implementation over the past
> couple of years in a big-spec language, I would not even think of
> attempting that in Scheme.
>
> It sounds as if you Schemers have conceded defeat on the tiny-spec
thing
> and have fallen back to "My chosen implementation is actually a
> big-spec, one-implementation, OS-portable language."
>
> To be fair, if I am just Joe Developer looking for a way to deliver
an
> application on multiple platforms, I can just pick one of the
> one-implementation Scheme "languages" and forget the others. No
argument
> there.
>
> The bad news for Scheme the language, unless one implementation
crushes
> the others, is that you now need to start a Common Scheme ANSI
> standardization process to undo the fragmentation of your shrinking
> community. The good news is that you can simply vote to adopt the
Common
> Lisp spec and be done before the first round of coffee and pastries
is
> gone. Failing that, the bad news is that you are now twenty years
behind.
The Scheme society is fighting against the fragmentation, but by SRFI
instead of adopting the Common Lisp spec. The SRFI process is so
successful that enen CL society copied it (CLRFI).
>
> kenny
>
> ps. which Scheme implementation gets your vote?
>
MrMathematica wrote:
>>
>>To be fair, if I am just Joe Developer looking for a way to deliver
>
> an
>
>>application on multiple platforms, I can just pick one of the
>>one-implementation Scheme "languages" and forget the others. No
>
> argument
>
>>there.
The argument still stands. You can choose a Common Lisp implementation
and do the same. I do it all the time.
...
> The Scheme society is fighting against the fragmentation, but by SRFI
> instead of adopting the Common Lisp spec. The SRFI process is so
> successful that enen CL society copied it (CLRFI).
>
The SRFI machinery is good; that is why the CLRFI is patterned after it.
But the goals are different. The SRFI machinery was necessary in the
first place to get implementations to agree on what is *already* in the
CL standard (hence implementing a variation of Greenspun's Tenth :) )
CLRFI is intended to provide a machinery to restart a community based
"standardization" (quotes mandatory) effort on small and big things alike.
Pathname fixing anyone? :)
Cheers
--
Marco
From: Christophe Rhodes
Subject: Re: I dream of a LISP that...
Date:
Message-ID: <sqis1n8apr.fsf@cam.ac.uk>
"MrMathematica" <···········@citiz.net> writes:
> The Scheme society is fighting against the fragmentation, but by SRFI
> instead of adopting the Common Lisp spec. The SRFI process is so
> successful that enen CL society copied it (CLRFI).
We did? I see precisely zero "CL society" takeup of the CLRFI
process. From the website:
CLRFI listing
At present, the only CLRFI exists: CLRFI: 1: FEATUREP. It's very
much a "trial" CLRFI; it's so trivial as to contain almost no
technical content; we can practice moving it around the system until
we are sure the process works.
No vendor has committed themselves to implementing a CLRFI; no user
has submitted a CLRFI. So it's certainly not a successful process in
the CL community, at least now. (There are many plausible
explanations for this, but this news message is too short to list
them.)
Christophe
Kenny Tilton <·······@nyc.rr.com> writes:
> The bad news for Scheme the language, unless one implementation
> crushes the others, is that you now need to start a Common Scheme ANSI
> standardization process to undo the fragmentation of your shrinking
> community. The good news is that you can simply vote to adopt the
> Common Lisp spec and be done before the first round of coffee and
> pastries is gone. Failing that, the bad news is that you are now
> twenty years behind.
You can also wait a week or two before agreeing on Common Lisp, to
keep eating the pastries and drink the coffee. It's worthwhile to
speed two weeks to close 20 years.
--
__Pascal Bourguignon__ http://www.informatimago.com/
Until real software engineering is developed, the next best practice
is to develop with a dynamic system that has extreme late binding in
all aspects. The first system to really do this in an important way
is Lisp. -- Alan Kay
Kenny Tilton <·······@nyc.rr.com> writes:
>
> The bad news for Scheme the language, unless one implementation crushes
> the others, is that you now need to start a Common Scheme ANSI
> standardization process to undo the fragmentation of your shrinking
Unfortunately Lisp also suffers from fragmentation. The CL specification
is not as big as it should be: there are essential features that are
left out of the standard, and that is bad.
I am new to Lisp and can't help comparing it to Java: to me, the
enormous success of Java can be attributed in a large degree to the
fact that Java tries to standardize the whole world. Everything that the
average programmer would ever need is already in the standard: Threads,
Sockets, Java Database Connectivity, Remote Method Invocation, XML
processing, Java Logging, Java Messaging Service, Java Cryptography, JFC
for graphics, JNDI, JAAS for authentication, etc.
I wish the CL specification included at least multi-processing (threads),
networking (not only sockets but also implementation of protocols like
http and ftp), database connectivity and cryptography.
Multiprocessing is especially crucial since in the future most computers
will be based on dual and multi-core architectures, and in order to
exploit their potential, all applications by necessity will have to rely
on threads. [1]
Yes I know that the Lisp implementations include their own versions of
many of these features: but this only enforces the argument that they
should be included in the standard.
1. "The Free Lunch Is Over: A Fundamental Turn Toward Concurrency in
Software", by Herb Sutter
(http://www.gotw.ca/publications/concurrency-ddj.htm)
--
vedm
On 13 May 2005 20:54:11 -0400, <··@nospam.com> wrote:
> I am new to Lisp and can't help comparing it to Java: to me, the
> enormous success of Java can be attributed in a large degree to the
> fact that Java tries to standardize the whole world. Everything that the
Someone disagrees with you.
http://www.armadilloaerospace.com/n.x/johnc/Recent%20Updates
sample: "Write-once-run-anywhere. Ha. Hahahahaha"
somehow I think he _knows_ what he is talking about.
With Sun heading down the tubes, what will you do when the 'white
knight' that takes them over decides that Java should be a cash cow?
I cannot think of a single Java program that I've been impressed by,
where should I be looking?
--
Everyman has three hearts;
one to show the world, one to show friends, and one only he knows.
From: Robert Uhl
Subject: Re: I dream of a LISP that...
Date:
Message-ID: <m3ll6hlqo6.fsf@4dv.net>
vedm <··@nospam.com> writes:
>
> I am new to Lisp and can't help comparing it to Java: to me, the
> enormous success of Java can be attributed in a large degree to the
> fact that Java tries to standardize the whole world.
ICBW, but I believe to a certain extent CL _is_ a standardisation of the
'whole world' as of the early 80s, although of course the ANSI standard
came quite a bit later. What's needed is a CL II.
> Multiprocessing is especially crucial since in the future most
> computers will be based on dual and multi-core architectures, and in
> order to exploit their potential, all applications by necessity will
> have to rely on threads.
Multiprocessing != multithreading. IMHO threads are just a hack because
process spawning is so expensive on certain platforms.
--
Robert Uhl <http://public.xdi.org/=ruhl>
Christus surrexit ex mortuis, | Christ is risen from the dead,
Mortem morte calcans, | Trampling down death by death,
Et illis in sepulchris vitam donans! | And to those in the tombs granting life!
vedm <··@nospam.com> writes:
> I am new to Lisp and can't help comparing it to Java: to me, the
> enormous success of Java can be attributed in a large degree to the
> fact that Java tries to standardize the whole world. Everything that the
[...]
> I wish the CL specification included at least multi-processing (threads),
> networking (not only sockets but also implementation of protocols like
> http and ftp), database connectivity and cryptography.
The Java design and standardization work could probably afford much
more resources than the ANSI CL committee. Are there any estimates of
how much money Sun put into Java design and standardization (not
marketing), and how much the ANSI CL committee?
Paolo
--
Why Lisp? http://lisp.tech.coop/RtL%20Highlight%20Film
Recommended Common Lisp libraries/tools (see also http://clrfi.alu.org):
- ASDF/ASDF-INSTALL: system building/installation
- CL-PPCRE: regular expressions
- UFFI: Foreign Function Interface
Paolo Amoroso <·······@mclink.it> writes:
> vedm <··@nospam.com> writes:
>
> > I am new to Lisp and can't help comparing it to Java: to me, the
> > enormous success of Java can be attributed in a large degree to the
> > fact that Java tries to standardize the whole world. Everything that the
> [...]
> > I wish the CL specification included at least multi-processing (threads),
> > networking (not only sockets but also implementation of protocols like
> > http and ftp), database connectivity and cryptography.
>
> The Java design and standardization work could probably afford much
> more resources than the ANSI CL committee. Are there any estimates of
> how much money Sun put into Java design and standardization (not
> marketing), and how much the ANSI CL committee?
That would be comparing apples and oranges.
ANSI is about competing vendors getting together.
JAVA is about SUN deciding what it alone will choose to do.
Kenny Tilton wrote:
> The bad news for Scheme the language, unless one implementation
crushes
> the others, is that you now need to start a Common Scheme ANSI
> standardization process to undo the fragmentation of your shrinking
> community. The good news is that you can simply vote to adopt the
Common
> Lisp spec and be done before the first round of coffee and pastries
is
> gone. Failing that, the bad news is that you are now twenty years
behind.
>
> kenny
>
> ps. which Scheme implementation gets your vote?
Kenny:
Please tell me - without a stick between your teeth - how often did you
actually port your code from one CommonLisp implementation to another
one. How often?
You are always arguing on basis of a CommonLisp spec. That is okay. But
heck your arguing does not tell me what CommonLisp has in favor of
Bigloo. And your arguing neglect the fact that a whole lot of Scheme
implemenetations run on many different platforms free of charge.
You guys are permanetly commiting a huge mistake: everyone should fall
in love with CommonLisp.
Oberministrant
PS: You forgot to tell me whether your CommonLisp spec will cover all
the subtle details when it comes to platform specific extensions and
additional add-on libraries.
PSS: And you also forgot to tell me whether your CommonLisp will make
me happier and more productive in my daily work than lets say Bigloo or
Dr.Scheme.
On 14 May 2005 02:22:06 -0700, <··········@hotmail.com> wrote:
^
|
Somehow I'm not surprised he favors Bigloo
> Please tell me how often did you actually port your code from
> one CommonLisp implementation to another one. How often?
How about 10 year old code, developed when the following CLs were
popular:
Texas Instruments Common Lisp version 6.1
Macintosh Common LISP (Allegro/Coral) versions 1.3.2 and 2.0b1
Lucid Common LISP version 4.0.x, 4.1
Allegro Common LISP (Franz inc.) version 4.1
which runs flawlessly (no change required) under ECL, GCL, CLISP,
Lispworks, Corman, CMUCL, SBCL. It's 2000 lines of genetic
programming functions.
--
Everyman has three hearts;
one to show the world, one to show friends, and one only he knows.
F�rster vom Silberwald wrote:
> Kenny Tilton wrote:
>
>
>>The bad news for Scheme the language, unless one implementation
>
> crushes
>
>>the others, is that you now need to start a Common Scheme ANSI
>>standardization process to undo the fragmentation of your shrinking
>>community. The good news is that you can simply vote to adopt the
>
> Common
>
>>Lisp spec and be done before the first round of coffee and pastries
>
> is
>
>>gone. Failing that, the bad news is that you are now twenty years
>
> behind.
>
>>kenny
>>
>>ps. which Scheme implementation gets your vote?
>
>
> Kenny:
>
> Please tell me - without a stick between your teeth -
hey, find your own ad homineum imagery
> how often did you
> actually port your code from one CommonLisp implementation to another
> one. How often?
AllegroCL, Lispworks, CormanCL, CLisp, CMUCL, MCL, and OpenMCL on win32,
Linux, and OS X.
>
> You are always arguing on basis of a CommonLisp spec. That is okay. But
> heck your arguing does not tell me what CommonLisp has in favor of
> Bigloo. And your arguing neglect the fact that a whole lot of Scheme
> implemenetations run on many different platforms free of charge.
Free is nice. Being locked into one implementation is not.
>
> You guys are permanetly commiting a huge mistake: everyone should fall
> in love with CommonLisp.
No, only computer programmers.
>
> Oberministrant
> PS: You forgot to tell me whether your CommonLisp spec will cover all
> the subtle details when it comes to platform specific extensions and
> additional add-on libraries.
I did, but I guess you missed it.
> PSS: And you also forgot to tell me whether your CommonLisp will make
> me happier and more productive in my daily work than lets say Bigloo or
> Dr.Scheme.
>
Irrelevant. What matters is that we agree: Bigger is better.
Just wondering: why do you keep mentioning different Schemes? Are you
using more than one?
kt
--
Cells? Cello?: http://www.common-lisp.net/project/cells/
Cells-Gtk?: http://www.common-lisp.net/project/cells-gtk/
Why Lisp? http://lisp.tech.coop/RtL%20Highlight%20Film
"Doctor, I wrestled with reality for forty years, and I am happy to
state that I finally won out over it." -- Elwood P. Dowd
Kenny Tilton wrote:
> Irrelevant. What matters is that we agree: Bigger is better.
Maybe I did not get it right. However, thanks for the pointer and
making me stronger in my arguing as a manager that we will stick to C++
till the death of hell.
> Just wondering: why do you keep mentioning different Schemes? Are you
> using more than one?
I will tell you the truth, whether you believe it or not: if my Bigloo
runs on one specific platform only I would likely start to hate
Scheme.
Although, porting software is not an issue for me, because Bigloo runs
on many platforms (okay, not as good on Windows as it should be).
What I do not understand: where do you get your distorted misconception
that a typical Scheme user is watching out for a tiny standard and
nothing more.
Schneewittchen
PS: I do not use something else than Bigloo, because I haven't found
any other Scheme implementation which has all the features of what
Bigloo actually will deliver me. But we know: different strokes for
different people. Oh, I do not have to repeat it again: I haven't found
any Lisp either which possess the features of Bigloo.
F�rster vom Silberwald wrote:
> Kenny Tilton wrote:
>
>
>>Irrelevant. What matters is that we agree: Bigger is better.
>
>
> Maybe I did not get it right. However, thanks for the pointer and
> making me stronger in my arguing as a manager that we will stick to C++
> till the death of hell.
Ohhhhh, you are just a hobbyist with Scheme. I had our enterprise use
Common Lisp. Why don't you?
kt
Kenny Tilton schrieb:
>
> I had our enterprise use Common Lisp. Why don't you?
Kenny, it is nice that you convinced your company to use CL.
I would like to do the same.
But really, it is not realistic that every Lisper can achieve
that. People and circumstances are different, so you can't
expect that Lisp will take over the world so fast.
Andr�
--
Kenny Tilton wrote:
> Ohhhhh, you are just a hobbyist with Scheme. I had our enterprise use
> Common Lisp. Why don't you?
Kenny:
Your reasoning is sometimes really strange. Surely, I am not a
professional programmer because I am employed at the university and I
write there my PhD in physics which I will have completed by
forthcomming summer.
I use Bigloo nearly every day for numerical simulations: post- and
pre-processing radiative transfer code; I wrote my own binding from
Bigloo to the high quality plotting library DISLIN (binding for Mac
OSX/Linux is downloadable from the Bigloo homepage); etc... In the past
3 to 4 years - during my PhD period - I wrote many thousand lines of
Bigloo code.
Tankwart
F�rster vom Silberwald wrote:
> Kenny Tilton wrote:
>
>
>>Ohhhhh, you are just a hobbyist with Scheme. I had our enterprise use
>
>
>>Common Lisp. Why don't you?
>
>
> Kenny:
>
> Your reasoning is sometimes really strange.
You mean the bit where I concluded fom "...making me stronger in my
arguing as a manager that we will stick to C++ till the death of hell"
that you are a manager using C++?
> Surely, I am not a
> professional programmer because I am employed at the university and I
> write there my PhD in physics which I will have completed by
> forthcomming summer.
(1) Don't call me Shirley.
(2) Aw, jeez, you are just a student? Not even a hobbyist? I would have
had some respect for a hobbyist. Well, at least you are not a professor.
Nothing could be worse than that.
:)
kenny
Kenny Tilton wrote:
> (1) Don't call me Shirley.
> (2) Aw, jeez, you are just a student? Not even a hobbyist? I would
have
> had some respect for a hobbyist. Well, at least you are not a
professor.
> Nothing could be worse than that.
>
> :)
Kenny: I think you do not have the slightest idea what it means to
write a PhD and beeing a scientist. Although, be happy that not every
academics is a high nosed than you otherwise you wouldn't have got your
CommonLisp. Or do you think your fellows are capable of writing your
spec?
By the way: I get paid for my PhD work (but you likely earn 5 times
more than what I get by means of my grant; but I am not on earth to
earn money) and could have choosen equally well Lisp over Scheme. But I
like Scheme better.
I cannot undestand why do you can't stand the fact that not everybody
loves Lisp. My boss assigned me a new Master student. I didn't teach
him Scheme; I practised the R-language with him because we are after
results. Tools don't matter in science and when you have to write
papers.
I am not here to convince you. But I have often seen that you are
deriding Scheme and its community.
Ein LaTeX-Hasser
"F�rster vom Silberwald" <··········@hotmail.com> writes:
> write a PhD and beeing a scientist. Although, be happy that not every
> academics is a high nosed than you otherwise you wouldn't have got your
> CommonLisp. Or do you think your fellows are capable of writing your
> spec?
An initial step for writing a good spec may be to get the name of the
thing being specified right. In this case, it's Common Lisp, not
CommonLisp.
Paolo
--
Why Lisp? http://lisp.tech.coop/RtL%20Highlight%20Film
Recommended Common Lisp libraries/tools (see also http://clrfi.alu.org):
- ASDF/ASDF-INSTALL: system building/installation
- CL-PPCRE: regular expressions
- UFFI: Foreign Function Interface
F�rster vom Silberwald wrote:
> Kenny Tilton wrote:
>
>
>>(1) Don't call me Shirley.
>>(2) Aw, jeez, you are just a student? Not even a hobbyist? I would
>
> have
>
>>had some respect for a hobbyist. Well, at least you are not a
>
> professor.
>
>>Nothing could be worse than that.
>>
>>:)
>
>
> Kenny: I think you do not have the slightest idea what it means to
> write a PhD and beeing a scientist.
Nope. I am blessed by limited intelligence. I have just enough to know
when my monkeys have written crap and throw it back at them. (They love
that, btw.)
> Although, be happy that not every
> academics is a high nosed than you otherwise you wouldn't have got your
> CommonLisp.
True. So the score now is One (John McCarthy) for academia, and how many
other idiots?
> I cannot undestand why do you can't stand the fact that not everybody
> loves Lisp.
No, that is fine, lots of people get things wrong. Me included. Included
being slow to pick up on Lisp. And anyone who likes Scheme is obviously
a kindred spirit. After re-education, you all will be welcome in the
Lisp camp.
I just want lurkers to know that Scheme is a huge mistake. A small spec
is an empty result. Purity is a false god. We have code to write, not CL
primitives to re-invent.
> I am not here to convince you. But I have often seen that you are
> deriding Scheme and its community.
Well it is pretty funny to see a language based on "let's see how small
we can keep the spec". McCarthy latched onto some Church math theory or
something. Math is cool and fascinating and deep. Small is...small? See
the difference?
OK, maybe this is the problem: I am fascinated by bad Prime Directives.
Not just Scheme "Small, small, and smaller", but Smalltalkian
all-objects-all-the-time. The interesting thing is that PDs can be a
pole star for greatness, but -- well, there it is. It cuts both ways. If
you pick the wrong pole star, hello rocks.
Have you Schemers been hearing less splashing and more crunching timbers
lately?
:)
kt
--
Cells? Cello?: http://www.common-lisp.net/project/cells/
Cells-Gtk?: http://www.common-lisp.net/project/cells-gtk/
Why Lisp? http://lisp.tech.coop/RtL%20Highlight%20Film
"Doctor, I wrestled with reality for forty years, and I am happy to
state that I finally won out over it." -- Elwood P. Dowd
Kenny Tilton wrote:
> True. So the score now is One (John McCarthy) for academia, and how
many
> other idiots?
>
> > I cannot undestand why do you can't stand the fact that not
everybody
> > loves Lisp.
Kenny:
I would like to offer you having the last word. But ...
However, you are putting words in my mouth: I haven't said that people
outside academia are retards.
But rest assured. Lately I have been recomending one from private
industry. The case is as follows: one of Scheme its steering commitee
members resigned from his editorial job. They posted affairs on
comp.lang.scheme and called for sending in comments to the Scheme
commitee and proposing appropiate successors.
I for one wrote to the language committe amd recommended one from
private industry who is also a very active writer on comp.lang.scheme.
The commitee itself is brimmed by academics; I believe it is a good
idea to fill the gap between academics-life and private industry or
public life. Hence my vote for that particular person.
> OK, maybe this is the problem: I am fascinated by bad Prime
Directives.
> Not just Scheme "Small, small, and smaller", but Smalltalkian
> all-objects-all-the-time. The interesting thing is that PDs can be a
> pole star for greatness, but -- well, there it is. It cuts both ways.
If
> you pick the wrong pole star, hello rocks.
It does not make sense to discuss it further because you doesn't
listen.
Maybe you are the last person on Earth who actually believes that
Schemers are after a tiny standard. Mastering lets say Dr.Scheme or
Bigloo is a huge undertaking and the Scheme language spec plays only
minority part here.
Historischer Schiffsmodellbauer mit Hilfe von Baukästen
"F�rster vom Silberwald" <··········@hotmail.com> writes:
> It does not make sense to discuss it further because you doesn't
> listen.
I have more the impression that YOU cannot cope intellectually with the
answers given to you. In general they deserve more care from your side
when you are replying.
> Maybe you are the last person on Earth who actually believes that
> Schemers are after a tiny standard.
Maybe you should watch the videos of the SICP lecture by Abelson and
Sussman. At least those two say there explicitly, that a language should
be as small as possible. (OK, that was 1986, but also today an academic
examining new language features should like a small Scheme).
Nicolas (a former Schemer).
Nicolas Neuss wrote:
>>Maybe you are the last person on Earth who actually believes that
>>Schemers are after a tiny standard.
>
>
> Maybe you should watch the videos of the SICP lecture by Abelson and
> Sussman. At least those two say there explicitly, that a language should
> be as small as possible. (OK, that was 1986, but also today an academic
> examining new language features should like a small Scheme).
I think they mean the base language (and standard). Of course
nobody has a problem with macros building most of CL on top of
Scheme (or do they? SICP is macro-free...).
Ok, the lack of for instance a standard package system is really
bad, but otherwise Scheme could have a big language on top of a
small kernel (like CL, in fact).
--
Don't let school interfere with your education. -- Mark Twain
Ulrich Hobelmann wrote:
> Nicolas Neuss wrote:
>
>>> Maybe you are the last person on Earth who actually believes that
>>> Schemers are after a tiny standard.
>>
>> Maybe you should watch the videos of the SICP lecture by Abelson and
>> Sussman. At least those two say there explicitly, that a language should
>> be as small as possible. (OK, that was 1986, but also today an academic
>> examining new language features should like a small Scheme).
>
> I think they mean the base language (and standard). Of course nobody
> has a problem with macros building most of CL on top of Scheme (or do
> they? SICP is macro-free...).
I don't know specifically about Abelson and Sussman, but I know that
some people definitely don't want a large language at all. (Some also
don't want macros, or at least think of them as superfluous.)
There is a whole chapter about that issue in Richard Gabriel's book
"Patterns of Software", called "Language Size". (Freely available at his
website.)
> Ok, the lack of for instance a standard package system is really bad,
> but otherwise Scheme could have a big language on top of a small kernel
> (like CL, in fact).
Good example: Some people think you don't need a module system when you
have closures. The pattern for creating your own "modules" with closures
is this:
(define f1 "undefined")
(define f2 "undefined")
(letrec ((f1-impl (lambda (...) ...))
(f2-impl (lambda (...) ...))
(helper1 (lambda (...) ...))
(helper2 (lambda (...) ...)))
(set! f1 f1-impl)
(set! f2 f2-impl))
Now you have "exported" f1 and f2, and helper1 and helper2 are internal
("private").
At least that's a pattern I have seen in a number of Scheme examples.
(No claims here that this is what "real" Schemers want.)
Pascal
--
2nd European Lisp and Scheme Workshop
July 26 - Glasgow, Scotland - co-located with ECOOP 2005
http://lisp-ecoop05.bknr.net/
Pascal Costanza wrote:
>> Ok, the lack of for instance a standard package system is really bad,
>> but otherwise Scheme could have a big language on top of a small
>> kernel (like CL, in fact).
>
>
> Good example: Some people think you don't need a module system when you
> have closures. The pattern for creating your own "modules" with closures
> is this:
>
> (define f1 "undefined")
> (define f2 "undefined")
> (letrec ((f1-impl (lambda (...) ...))
> (f2-impl (lambda (...) ...))
> (helper1 (lambda (...) ...))
> (helper2 (lambda (...) ...)))
> (set! f1 f1-impl)
> (set! f2 f2-impl))
>
> Now you have "exported" f1 and f2, and helper1 and helper2 are internal
> ("private").
I coded in stuff like that in Scheme for a couple of days. Ugh.
Through with that ;)
> At least that's a pattern I have seen in a number of Scheme examples.
> (No claims here that this is what "real" Schemers want.)
Hm, after a while I left Scheme for SML and now study Lisp, so
maybe I'm just not a real Schemer.
--
Don't let school interfere with your education. -- Mark Twain
Pascal Costanza wrote:
> Good example: Some people think you don't need a module system when you
> have closures. The pattern for creating your own "modules" with closures
> is this:
>
> (define f1 "undefined")
> (define f2 "undefined")
> (letrec ((f1-impl (lambda (...) ...))
> (f2-impl (lambda (...) ...))
> (helper1 (lambda (...) ...))
> (helper2 (lambda (...) ...)))
> (set! f1 f1-impl)
> (set! f2 f2-impl))
>
> Now you have "exported" f1 and f2, and helper1 and helper2 are internal
> ("private").
>
> At least that's a pattern I have seen in a number of Scheme examples.
> (No claims here that this is what "real" Schemers want.)
Yeah - if one wants portable r5rs-only-only code that's the way to do it.
Incidently it is the exact same pattern one uses for sharing state
between functions without polluting the global namespace:
<http://schemecookbook.org/view/Cookbook/IdiomSharingState>
Again, the module system solution is the prettiest.
--
Jens Axel Søgaard
Ulrich Hobelmann <···········@web.de> writes:
> Nicolas Neuss wrote:
>> Maybe you should watch the videos of the SICP lecture by Abelson and
>> Sussman. At least those two say there explicitly, that a language should
>> be as small as possible. (OK, that was 1986, but also today an academic
>> examining new language features should like a small Scheme).
>
> I think they mean the base language (and standard). Of course nobody has a
> problem with macros building most of CL on top of Scheme (or do they? SICP
> is macro-free...).
If I understand the spirit of those lectures right, A&S' ideal is that the
programmer builds the language which is needed for the problem at hand, and
if this needs writing an own interpreter or compiler - fine, no problem.
Since they want to teach how these things are done, and since this is more
basic than macros, they do not have to introduce those. (On the other
hand, I could very well imagine that their course has changed a lot in
those 20 years. Does someone know better?)
> Ok, the lack of for instance a standard package system is really bad, but
> otherwise Scheme could have a big language on top of a small kernel (like
> CL, in fact).
My impression is that Schemers don't like the CL package system. If I
remember correctly, more powerful "modules" are implemented in PLT scheme
with the help of syntax-case and maybe this will be the route Scheme will
take. I do not care too much, because the CL system works well enough for
me.
My conclusion is the following: different people have different needs. If
I want to teach computer science to students, a small language like Scheme
(according to its spec) is better suited than CL - and indeed, I used
Scheme rather successfully in precisely this way. Also, if I want to
experiment with different evaluation orders, different macro systems, etc,
the language should be as small as possible. And there are several
Schemers who need and want precisely that.
On the other hand, I for example did not feel comfortable with Scheme as an
application programmer. The reason is that I am more a user of a computer
language and not so much a designer. More precisely, the Scheme extensions
which I was forced to design were in almost every case of worse quality
than what CL gives me for free. And, as much as I saw at that time
(<2000), this was also true for many other Schemer's code and libraries.
Nicolas.
Nicolas Neuss <·······@iwr.uni-heidelberg.de> writes:
> Ulrich Hobelmann <···········@web.de> writes:
>
> My conclusion is the following: different people have different needs. If
> I want to teach computer science to students, a small language like Scheme
> (according to its spec) is better suited than CL - and indeed, I used
> Scheme rather successfully in precisely this way.
I've seen this reasoning frequently, and as a commercial programmer,
I'd like point out that this approach has unfortunate
consequences. Programmers come out of school with the impression that
Lisp treats everything as lists, has no object system, no inherent
module system (packages) etc. ... in short, that it's not suitable for
modern commercial development, even if they liked what they _did_
see. This is on my mind because I just ran into such an attitude at
work from a young programmer this past week, but I've encountered it
several times in the past few years. It's unfortunate, because often
these misconceptions are held throughout their careers, and contribute
to (or may even be a primary cause of) the general lack of interest in
Lisp for commercial development.
I beg you to consider changing this practice if you're still teaching
courses this way. Or at the very least, make sure that students are
also told emphatically about the depth and breadth available from
"that other" major Lisp dialect, Common Lisp.
···@zedat.fu-berlin.de (Stefan Ram) writes:
> Dan Muller <·········@sneakemail.com> writes:
> >Programmers come out of school with the impression that Lisp
> >treats everything as lists, has no object system, no inherent
> >module system (packages) etc. ...
>
> These seems to be properties of the classical
> language "LISP", but not of "Common Lisp".
>
> I never did like it, when the meaning of a name is changed.
> I believe, it would be better to find a new name for
> a new language.
Apparently this belief is so important to you that you use
it to rationalize your deliberately misleading students,
willfully misinforming them and depriving them of useful
knowledge in their field of study.
Do you also teach them utterly wrong things about
the other programming languages?
···@zedat.fu-berlin.de (Stefan Ram) writes:
> >Do you also teach them utterly wrong things about the other
> >programming languages?
>
> So, early LISPs had an object system and packages?
> (They did not treat "everything" as lists, indeed.)
>
> The "properties" I refered to were "no object system",
> "no inherent module system".
My understanding is that whie you yourself know better,
you are not clearly informing the students that this
"early Lisp" is hardly representative of the Lisp
language as understood from the late 1970s onward.
I further understand your explanation of this behaviour
to be rooted in your own perverse beliefs that words
should not change meaning over time.
If I have misunderstood what you're saying, please correct me.
···@zedat.fu-berlin.de (Stefan Ram) writes:
> ······@news.dtpq.com (Christopher C. Stacy) writes:
> >My understanding is that whie you yourself know better,
> >you are not clearly informing the students that this
> >"early Lisp" is hardly representative of the Lisp
> >language as understood from the late 1970s onward.
>
> Classes I gave recently covered specific programming
> languages, but not Lisp - so I did not mention Lisp at all.
> If I will ever teach Common Lisp or a topic that includes
> Common Lisp, I will take care not to misrepresent the
> language.
Well, that sounds good!
Stefan Ram wrote:
> ······@news.dtpq.com (Christopher C. Stacy) writes:
>
>>>>Programmers come out of school with the impression that Lisp
>>>>treats everything as lists, has no object system, no inherent
>>>>module system (packages) etc. ...
>>>
>>>These seems to be properties of the classical
>>>language "LISP", but not of "Common Lisp".
>>
>>rationalize your deliberately misleading students
>
>
> Possibly you meant to say
>
> "rationalize your deliberately misleading /of/ students"?
>
> Otherwise, I can not understand this fragment.
No, that would have to be "deliberate misleading of students", with the
adjective deliberate qualifying the noun misleading qualified by of
students.
The OP was fine: the adverb deliberately qualifying the verb misleading
with the object students.
kt
Stefan Ram wrote:
> Kenny Tilton <·······@nyc.rr.com> writes:
>
>>>>rationalize your deliberately misleading students
>>
>>No, that would have to be "deliberate misleading of students", with the
>>adjective deliberate qualifying the noun misleading qualified by of
>>students.
>
>
> I agree, after submitting my posting I have thought so myself,
> too.
>
>
>>The OP was fine: the adverb deliberately qualifying the verb misleading
>>with the object students.
>
>
> I am not a native speaker of English, but the problem I have
> is bringing this together with the adjective "your". I could
> somehow parse "you", i.e., "rationalize you [as someone who is
> ]deliberately misleading students", but the adjective "your"
> seems to expect a following noun, such as "your deliberate
> misleading of students".
>
>
I agree you works better with the deliberately version, but as someone
who groks implied connectives such as [as someone who is] you can
probably imagine also a connective that saves your along with a verb
clause. But, yeah, you just works (and your English is fine). :)
kenny
--
Cells? Cello?: http://www.common-lisp.net/project/cells/
Cells-Gtk?: http://www.common-lisp.net/project/cells-gtk/
Why Lisp? http://lisp.tech.coop/RtL%20Highlight%20Film
"Doctor, I wrestled with reality for forty years, and I am happy to
state that I finally won out over it." -- Elwood P. Dowd
From: Robert Uhl
Subject: Re: I dream of a LISP that...
Date:
Message-ID: <m34qcthgu4.fsf@4dv.net>
···@zedat.fu-berlin.de (Stefan Ram) writes:
>
> > > > rationalize your deliberately misleading students
> >
> > No, that would have to be "deliberate misleading of students", with
> > the adjective deliberate qualifying the noun misleading qualified by
> > of students.
[snip]
> I am not a native speaker of English, but the problem I have is
> bringing this together with the adjective "your". I could somehow
> parse "you", i.e., "rationalize you [as someone who is] deliberately
> misleading students", but the adjective "your" seems to expect a
> following noun, such as "your deliberate misleading of students".
Well, I _am_ a native English speaker, and I'm fairly certain that while
ambiguous the original was proper English: 'your' and 'deliberately'
modify the gerund 'misleading,' which has the direct object 'students.'
It wasn't the most elegant of sentences, and I can certainly see how a
non-native would be confused.
Stylistically it's poor English, but I believe that it's syntactically
correct.
--
Robert Uhl <http://public.xdi.org/=ruhl>
Cristo Resucitado! En Verdad Resucitado!
Robert Uhl <······@SPAM4dv.net> writes:
> It wasn't the most elegant of sentences, and I can certainly see how a
> non-native would be confused.
Perhaps so, but the non-native speaker in this case did not
in fact suffer any confusion.
Dan Muller <·········@sneakemail.com> writes:
> Nicolas Neuss <·······@iwr.uni-heidelberg.de> writes:
>> My conclusion is the following: different people have different needs. If
>> I want to teach computer science to students, a small language like Scheme
>> (according to its spec) is better suited than CL - and indeed, I used
>> Scheme rather successfully in precisely this way.
>
> ...
> I beg you to consider changing this practice if you're still teaching
> courses this way. Or at the very least, make sure that students are
> also told emphatically about the depth and breadth available from
> "that other" major Lisp dialect, Common Lisp.
I did exactly that in this course, for one week at the end. And I think
this was quite right. I am rather contemplating to say more about other
languages (some Assembler, C, Smalltalk, Prolog).
BTW, is there some teaching program for assembly language, where you can
also see how registers and memory change, etc.?
Nicolas.
Nicolas Neuss wrote:
> Dan Muller <·········@sneakemail.com> writes:
> > Nicolas Neuss <·······@iwr.uni-heidelberg.de> writes:
> >> My conclusion is the following: different people have different
> >> needs. If I want to teach computer science to students, a
> >> small language like Scheme (according to its spec) is better
> >> suited than CL - and indeed, I used Scheme rather successfully
> >> in precisely this way.
> >
> > ...
> > I beg you to consider changing this practice if you're still
> > teaching courses this way. Or at the very least, make sure that
> > students are also told emphatically about the depth and breadth
> > available from "that other" major Lisp dialect, Common Lisp.
>
> I did exactly that in this course, for one week at the end. And I
> think this was quite right. I am rather contemplating to say more
> about other languages (some Assembler, C, Smalltalk, Prolog).
>
> BTW, is there some teaching program for assembly language, where
> you can also see how registers and memory change, etc.?
I believe MIT has something called BSim, which IIRC uses Java AWT (or
maybe Swing?) to show changes to registers and whatnot.
http://ocw.mit.edu/OcwWeb/Electrical-Engineering-and-Computer-Science/6-004Computation-StructuresFall2002/RelatedResources/
I don't really remember it, but I remember thinking it was nice. I
wonder if their sourcecode can be found somewhere. I asked them for it,
but received no response.
SPIM was also available as a MIPS simulator, but I could never make
sense of its Windows version.
http://www.cs.wisc.edu/~larus/spim.html
On Knuth's pages, you can probably find something for MIX/MMIX. Nothing
that's interesting to me, but I've given one or two a test spin.
Nicolas Neuss <·······@iwr.uni-heidelberg.de> writes:
> BTW, is there some teaching program for assembly language, where you can
> also see how registers and memory change, etc.?
WINDLX
I don't know if it has some official page but you can get here,
among other places:
http://cs.uns.edu.ar/~jechaiz/arquitectura/windlx/windlx.html
--
Luis Oliveira
luismbo (@) gmail (.) com
Equipa Portuguesa do Translation Project
http://www2.iro.umontreal.ca/~pinard/po/registry.cgi?team=pt
Kenny Tilton wrote:
> > Although, be happy that not every
> > academics is a high nosed than you otherwise you wouldn't have got
your
> > CommonLisp.
>
> True. So the score now is One (John McCarthy) for academia, and how
many
> other idiots?
Who says that? Please see
http://www8.informatik.uni-erlangen.de/html/lisp/mcc91.html
In the final Conclusion, "As things stand, he (John McCarthy) must
prefer SCHEME to CommonLISP -- a clear, understandable small diamond,
to a messy, incomprehensible clump." An the article give evidence of
that.
>
> "Doctor, I wrestled with reality for forty years, and I am happy to
> state that I finally won out over it." -- Elwood P. Dowd
MrMathematica wrote:
> Kenny Tilton wrote:
>
>>>Although, be happy that not every
>>>academics is a high nosed than you otherwise you wouldn't have got
>
> your
>
>>>CommonLisp.
>>
>>True. So the score now is One (John McCarthy) for academia, and how
>
> many
>
>>other idiots?
>
>
> Who says that? Please see
> http://www8.informatik.uni-erlangen.de/html/lisp/mcc91.html
> In the final Conclusion, "As things stand, he (John McCarthy) must
> prefer SCHEME to CommonLISP -- a clear, understandable small diamond,
> to a messy, incomprehensible clump." An the article give evidence of
> that.
So you are saying the score is zero for academia? I can live with that.
kenny
Kenny Tilton wrote:
> MrMathematica wrote:
> > Kenny Tilton wrote:
> >
> >>>Although, be happy that not every
> >>>academics is a high nosed than you otherwise you wouldn't have got
> >
> > your
> >
> >>>CommonLisp.
> >>
> >>True. So the score now is One (John McCarthy) for academia, and how
> >
> > many
> >
> >>other idiots?
> >
> >
> > Who says that? Please see
> > http://www8.informatik.uni-erlangen.de/html/lisp/mcc91.html
> > In the final Conclusion, "As things stand, he (John McCarthy) must
> > prefer SCHEME to CommonLISP -- a clear, understandable small
diamond,
> > to a messy, incomprehensible clump." An the article give evidence
of
> > that.
>
> So you are saying the score is zero for academia? I can live with
that.
>
> kenny
No. I am just say that you are wrong.
And the sentence "a clear, understandable small diamond, to a messy,
incomprehensible clump" exactly explain why I prefer Scheme (with a
decent implementation).
I didn't port my code between different Scheme implementation, because
my vote, DrScheme, can run on nearly all platform. If I need some
feature but it lacks, I can implement it myself because DrScheme is
open-source.
-
Programs should be written for people to read, and only incidentally
for machines to execute. -- Abelson & Sussman
MrMathematica wrote:
> Kenny Tilton wrote:
>
>>MrMathematica wrote:
>>
>>>Kenny Tilton wrote:
>>>
>>>
>>>>>Although, be happy that not every
>>>>>academics is a high nosed than you otherwise you wouldn't have got
>>>
>>>your
>>>
>>>
>>>>>CommonLisp.
>>>>
>>>>True. So the score now is One (John McCarthy) for academia, and how
>>>
>>>many
>>>
>>>
>>>>other idiots?
>>>
>>>
>>>Who says that? Please see
>>>http://www8.informatik.uni-erlangen.de/html/lisp/mcc91.html
>>>In the final Conclusion, "As things stand, he (John McCarthy) must
>>>prefer SCHEME to CommonLISP -- a clear, understandable small
>
> diamond,
>
>>>to a messy, incomprehensible clump." An the article give evidence
>
> of
>
>>>that.
>>
>>So you are saying the score is zero for academia? I can live with
>
> that.
>
>>kenny
>
>
> No. I am just say that you are wrong.
>
> And the sentence "a clear, understandable small diamond, to a messy,
> incomprehensible clump" exactly explain why I prefer Scheme (with a
> decent implementation).
>
> I didn't port my code between different Scheme implementation, because
> my vote, DrScheme, can run on nearly all platform. If I need some
> feature but it lacks, I can implement it myself because DrScheme is
> open-source.
Open Source works when folks have a standard compiler. C has GCC, Scheme
has -- oh, sorry, there is none. Common Lisp has <gasp> an ANSI spec!
You know, it is funny. Lisp started out minimalist. Actual use led diff
implementations to come up with different, more high-powered constructs,
and compete among implementations on those. The ensuing fragmentation
became an objection to the adoption of Lisp, so Common Lisp was born.
Scheme seems to have gotten over its obsession with spec-size to create
standards and specs in the form of disparate implementations. That puts
them twenty years behind, because Common Lisp is where PLT and Bigloo
and DrScheme are going.
kt
Kenny Tilton wrote:
>
> Open Source works when folks have a standard compiler. C has GCC,
Scheme
> has -- oh, sorry, there is none. Common Lisp has <gasp> an ANSI spec!
>
> You know, it is funny. Lisp started out minimalist. Actual use led
diff
> implementations to come up with different, more high-powered
constructs,
> and compete among implementations on those. The ensuing fragmentation
> became an objection to the adoption of Lisp, so Common Lisp was born.
>
> Scheme seems to have gotten over its obsession with spec-size to
create
> standards and specs in the form of disparate implementations. That
puts
> them twenty years behind, because Common Lisp is where PLT and Bigloo
> and DrScheme are going.
>
> kt
Scheme (both the spec and implementations) will progress, and it is
progressing, for example, the ongoing R6RS, the v300 of PLT, etc. But
it will never became Common Lisp. Common Lisp is premature / over
optimized and got stuck in a local maximum, while Scheme, although not
mature yet, aims the global maximum.
MrMathematica wrote:
> Kenny Tilton wrote:
>
>>MrMathematica wrote:
>>
>>>Kenny Tilton wrote:
>>>
>>>
>>>>>Although, be happy that not every
>>>>>academics is a high nosed than you otherwise you wouldn't have got
>>>
>>>your
>>>
>>>
>>>>>CommonLisp.
>>>>
>>>>True. So the score now is One (John McCarthy) for academia, and how
>>>
>>>many
>>>
>>>
>>>>other idiots?
>>>
>>>
>>>Who says that? Please see
>>>http://www8.informatik.uni-erlangen.de/html/lisp/mcc91.html
>>>In the final Conclusion, "As things stand, he (John McCarthy) must
>>>prefer SCHEME to CommonLISP -- a clear, understandable small
>
> diamond,
>
>>>to a messy, incomprehensible clump." An the article give evidence
>
> of
>
>>>that.
>>
>>So you are saying the score is zero for academia? I can live with
>
> that.
>
>>kenny
>
>
> No. I am just say that you are wrong.
>
> And the sentence "a clear, understandable small diamond, to a messy,
> incomprehensible clump" exactly explain why I prefer Scheme (with a
> decent implementation).
... which gives you back a messy incomprehensible *and incompatible*
clump. :)
Therefore we are talking about clumps no matter what. And the
conclusions is that Common Lisp clumps are still better than
Scheme-implementation-X clumps :)
Cheers
--
Marco
MrMathematica schrieb:
> Who says that? Please see
> http://www8.informatik.uni-erlangen.de/html/lisp/mcc91.html
> In the final Conclusion, "As things stand, he (John McCarthy) must
> prefer SCHEME to CommonLISP -- a clear, understandable small diamond,
> to a messy, incomprehensible clump." An the article give evidence of
> that.
Theauthor is talking alot about CommonLISP. Maybe in his next article
he will also talk about Common Lisp.
Andr�
--
"F�rster vom Silberwald" <··········@hotmail.com> writes:
> Kenny: I think you do not have the slightest idea what it means to
> write a PhD and beeing a scientist.
Well, surely, I do, having "been there and done that".
When I was a scientist, I prided myself on being
"quite a good programmer, really". Certainly, a lot better
than almost all my fellow scientists. This turned out handy
when I switched career to software.
That's when I discovered that I didn't know how to program at all. :-(
So, of course, there _are_ exceptions (heck, one well known PhD in
Physics did write Mathematica(TM)), but they're rare, and, what Kenny
is saying is that, as a student, you (probably) don't know much about
what programming "in the real world" is all about.
And he's probably right. No offense---the odds are simply in his
favour.
--ap
p.s. And congrats on using _any_ kind of lisp to do your physics degree!
That's way cool!]
p.p.s. I would be slow to take offense from Kenny. His tongue is usually
planted _very_ firmly in his cheek.
Alain Picard wrote:
> p.s. And congrats on using _any_ kind of lisp to do your physics
degree!
> That's way cool!]
>
> p.p.s. I would be slow to take offense from Kenny. His tongue is
usually
> planted _very_ firmly in his cheek.
Hey, I do not feel offended from Kenny since he a very enthuastic
Common Lisp user (I know his writings from comp.lang.lisp). Somehow
symphatic even. Lispers are the way they are. Compared to retard Eric
N. Kenny turns out being a genuine gentleman.
It happens that I sometimes deride Python and its users.
However, people can tell me what they want and can go into debating
over details and semantics. Yes I know it is imprinted into stone that
one of the first motivations for Scheme languages was a standard as
tiny as possible. However, any Scheme implementation goes far beyond
that. I do not see where the problem lies as long as people are aware
of that fact. Bigloo does not come with hidden extensions; quite to the
contrary.
Peace
Schneewittchen
From: Robert Uhl
Subject: Re: I dream of a LISP that...
Date:
Message-ID: <m37jhzc9rf.fsf@4dv.net>
"F�rster vom Silberwald" <··········@hotmail.com> writes:
>
> Yes I know it is imprinted into stone that one of the first
> motivations for Scheme languages was a standard as tiny as
> possible. However, any Scheme implementation goes far beyond that. I
> do not see where the problem lies as long as people are aware of that
> fact. Bigloo does not come with hidden extensions; quite to the
> contrary.
The issue is not whether they are hidden or not; it's whether one's code
is portable or not. E.g. Bigloo is certainly an excellent Scheme
implementation, and its extensions are surely very good as well. But
what if one should wish to use one's code (even a portion) elsewhere?
The nice thing about Common Lisp is that its standard library is very
large, and that most of one's code is portable. The core is, anyway:
such things as networking and the like haven't yet been standardised.
And should be.
There's a difference between language and standard library which gets
overlooked in discussions about specifications. A small language is a
Good Thing; a small standard library rather less so.
Yes, you're able to do lots of very cool things with Bigloo. But what
about when you'd like to run your current code on a platform it doesn't
support? Or what if you'd like to take advantage of a feature--in your
current code--that another implementation offers but Bigloo doesn't?
Then you face a rather large amount of porting.
Not that this doesn't happen with Common Lisp; it just happens rather
less often, because the standard library is so much larger.
Common Lisp isn't perfect, but it does have certain advantages over
Scheme.
--
Robert Uhl <http://public.xdi.org/=ruhl>
Ha Mashiyach qam! Ken hu qam!
On 16 May 2005 10:29:05 -0700, <··········@hotmail.com> wrote:
I vote that Chain-Lube, AKA "Ranger of the Silver Forest" [1] gets
troll-of-the-month award. I wouldn't regard any of the claims he's
made as accurately representing reality, since a Google search will
provide a different set of claims.
I believe he has played the blockhead perfectly, muddying his
responses and drawing others into the thread. It will be interesting
to see if this post actually survives. He surely has outdistanced the
recent competition.
> It happens that I sometimes deride Python and its users.
Seems like they would be more fun for you.
> However, people can tell me what they want and can go into debating
> over details and semantics. Yes I know it is imprinted into stone that
> one of the first motivations for Scheme languages was a standard as
> tiny as possible. However, any Scheme implementation goes far beyond
> that. I do not see where the problem lies as long as people are aware
Scheme is very old, please point out any Scheme implementation equally
as old. Or to make it easier on you, point out any surviving from
1984. OK, how about a scheme that is only 10 years old with full
backwards compatibility.
> Schneewittchen
I think your best post was asking a German to speak to another German,
since he was a native speaker. That was a bit out of character eh?
[1] a famous German movie from the time before Lisp. A few other
things, but "vom" doesn't seem to appear in a persons name...
--
With sufficient thrust, pigs fly fine.
GP lisper wrote:
>
> Scheme is very old, please point out any Scheme implementation
equally
> as old. Or to make it easier on you, point out any surviving from
> 1984. OK, how about a scheme that is only 10 years old with full
> backwards compatibility.
>
How old is PLT-Scheme? How old is v300 of PLT-Scheme?
GP lisper wrote:
OK, how about a scheme that is only 10 years old with full
> backwards compatibility.
Gähn. Nobody knows what you are talking about. Your writing is not
well focused and a bit confusing. Maybe you should read Kant once in
your life; so you will get better comprehension abilities; I hope
your writing is not based on any categorical-imperative.
Bigloo is around for 10 years and longer I assume.
However, where is the problem? If it were around for 5 years - so what?
I would still be using it.
Thanks for you useful contribution. Maybe you should read the posts
from Pascal K. and you will see what it means delivering useful
answers; even to a perceived troll.
That is my last word in the affair. I can live without Common Lisp and
Scheme/Bigloo - but can you? I do not think so.
Desserta Berghof Creme Joghurt
PS: Thanks for your convincing arguments; I will quickly trade in my
Bigloo for Common Lisp. Oh wait, there are others in comp.lang.lisp who
had better arguments and maybe it will happen that I am diving deeper
into Common Lisp.
·············@hotmail.com wrote:
> Thanks for you useful contribution. Maybe you should read the posts
> from Pascal K. and you will see what it means delivering useful
> answers; even to a perceived troll.
Sorry I meant Pascal Constanza. I always like his writings on
comp.lang.lisp. He is a dedicated Common Lisp user and researcher in
the field but he is not a fanatical pig-head.
Sonnenbrille
A long time ago, in a galaxy far, far away, "F�rster vom Silberwald" <··········@hotmail.com> wrote:
> It happens that I sometimes deride Python and its users.
Since they start with a tradition that involves a rich layering of
(hopefully comedic!) insults, I doubt they are _too_ thin-skinned
about that.
> However, people can tell me what they want and can go into debating
> over details and semantics. Yes I know it is imprinted into stone
> that one of the first motivations for Scheme languages was a
> standard as tiny as possible. However, any Scheme implementation
> goes far beyond that. I do not see where the problem lies as long as
> people are aware of that fact. Bigloo does not come with hidden
> extensions; quite to the contrary.
The fundamental problem with Scheme extensions is that there are so
many approaches to choose from.
Where Common Lisp has a much larger 'library' of functions to start
from, the "official" Scheme definition is really sparse.
That becomes a problem when different implementors have different sets
of extensions that force people to choose between them.
If I pick Bigloo, I get its set of extensions that are NOT the same as
those for MzScheme. MIT Scheme is a very different alternative. SCM
has aspects that are interesting, but differences too. If I want a
wildly-powerfully-optimizing compiler, Stalin is the choice.
As soon as I touch files, I have to pick one "party" or another.
In contrast, I can write Common Lisp applications of some degree of
"interestingness" and not need to pick unportable bits for accessing
files. Running external Unix processes isn't too hard either. And I
don't have to be "partisan" for one CL implementation or another...
--
(reverse (concatenate 'string "moc.liamg" ·@" "enworbbc"))
http://linuxdatabases.info/info/slony.html
"If you use pico in a professional setting, first they will talk about
you. Then they will laugh at you. Then they will fire you, and then
you lose." -- crackmonkey.org
"Förster vom Silberwald" <··········@hotmail.com> writes:
> Hey, I do not feel offended from Kenny since he a very enthuastic
> Common Lisp user (I know his writings from comp.lang.lisp). Somehow
> symphatic even. Lispers are the way they are. Compared to retard Eric
> N. Kenny turns out being a genuine gentleman.
>
> It happens that I sometimes deride Python and its users.
>
Funny thread. Quite enjoyed it guys, thanks.
I do think Kenny's point has been missed (or I've misunderstood
it). It seems to me what he is saying is that the problem with scheme
is that because it has such a small spec that all the additional
utility functions/modules/packages or whatever are implemented in
whatever way the maintainers of the version you use feel like. This
means that if yyou want to change scheme implementations, you are
going to face a much larger porting problem than with a language like
lisp where much of the utility is included in the spec. the tiny
size of scheme is great for teaching - in fact, I think its important
when learning to try and implement common utility functions because it
helps develop understanding and its much easier to swallow a small
language in a semester or two of courses. I'd even go a step further
and ban graphical debuggers which allow stepping/tracing of execution
on line at a time with variable watch windows which let you see the
value of specific variables at different times. Tools like this should
only be used once you understand how you should debug code and realise
the best debugger on the planet is totally free and between your ears!
Then again, I'm known to be an arrogant bastard full of
pomposity. However, I first learnt to program when 'graphical'
displays meant extended ascii line drawing characters and the only
successful and readily available debugging tool was a good
understanding of your algorithm, its logic and the occasional print
statement.
When you start work in the real world, productivity is what you want
and if possible you want to avoid dependency on a single
implementation so that your market is not artificially constrained by
the supported platforms of the implementation and avoid being tied to
an implementation which may take a totally different direction in the
future to the one you were expecting or hoping for.
While its true using scheme implementations lib bigloo or dr scheme
means you get a lot of utility from that implementation and while its
true these implementations are available on a lot of platforms, you
still don't have the level of portability that CL can provide. While
it is true that because bigloo is open source you can always port it
or modify it, the point is you probably don't have to do this with
something in CL.
To some extent, I get the feeling Kenny's argument could be extended
to the more generalised debate concerning how much of a language
should be controlled by the spec and how much should be left to
different implementors to do as they see fit. I remember the arguments
in the 80's about C standards (and lack of them) and notice the
popularity of Java and its large selection of centrally controlled
APIs and the number of people who cursed Ada for being too large and
difficult to understand/learn/know. Swings and roundabouts, changing
hemlines, religious wars - its such a fun world!
I suspect much of the other stuff Kenny said was done with his tongue
firmly planted in his cheek (or perhaps he was just playing with you -
his middle name possibly starts with 'A', making his full initials
KAT.
Don't fret little mouse - different strokes for different folks and
I'm sure there are plenty who can help you when you need to port to
another implementation!
Kenny, if I've misrepresented you - well, this is the net and shit
happens. Its not intentional of course.
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!
Kenny Tilton wrote:
> (2) Aw, jeez, you are just a student? Not even a hobbyist? I would have
> had some respect for a hobbyist. Well, at least you are not a professor.
> Nothing could be worse than that.
>
> :)
I hope the :) means that this isn't meant insulting. I don't
think that working coders or professors are inherently better than
one another. And student? Well, what's wrong with that? It's
the almost mandatory route these days to get a job as an
Oh-so-superior programmer.
Maybe my learning CL now is a sign that indeed I shouldn't go into
academics ;)
--
Don't let school interfere with your education. -- Mark Twain
Ulrich Hobelmann <···········@web.de> wrote:
> Kenny Tilton wrote:
>> (2) Aw, jeez, you are just a student? Not even a hobbyist? I would
>> have had some respect for a hobbyist. Well, at least you are not a
>> professor. Nothing could be worse than that.
>> :)
>
> I hope the :) means that this isn't meant insulting. I don't think
> that working coders or professors are inherently better than one
> another. And student? Well, what's wrong with that? It's the almost
> mandatory route these days to get a job as an Oh-so-superior
> programmer.
>
> Maybe my learning CL now is a sign that indeed I shouldn't go into
> academics ;)
Alas, there's some tendency for some varieties of academic involvement
to have a pretty negative effect on one's understanding of things.
Consider the "Torvalds/Tanembaum" flame war. Tanembaum indicated, as
"eminent professor," that he wouldn't give good marks to someone
creating a new monolithic OS kernel like Linux.
Of course, hardly anyone uses a micro-kernel these days; the academic
interest of the early 1990s didn't actually lead to widespread
practical use.
Students led by these sorts of academic interests wind up pretty
desperately misdirected. And the issue _isn't_ about micro-kernels;
it's about the near-myopia about those sorts of issues.
The Lisp-relevant thing is that academics seem wont to make grand
statements about what Lisp is like based on what was true in Lisp 1.5
that is devoid of any understanding of what has happened since.
--
"cbbrowne",·@","gmail.com"
http://linuxdatabases.info/info/emacs.html
"We come to bury DOS, not to praise it."
-- Paul Vojta <·····@math.berkeley.edu>, paraphrasing a quote of
Shakespeare
Christopher Browne wrote:
> Alas, there's some tendency for some varieties of academic involvement
> to have a pretty negative effect on one's understanding of things.
>
> Consider the "Torvalds/Tanembaum" flame war. Tanembaum indicated, as
> "eminent professor," that he wouldn't give good marks to someone
> creating a new monolithic OS kernel like Linux.
I wouldn't either. I'd be much happier if we had a nice system
based on L4 instead of that huge Linux kernel blob. DragonflyBSD
seems a nice alternative, while not microkernelized (which isn't
my point) but certainly modular and multithreaded. It's not about
context-switching, it's about modularity, and Linux isn't exactly
great in that area (see the binary drivers mess, both technical
and legal (license)).
> Of course, hardly anyone uses a micro-kernel these days; the academic
> interest of the early 1990s didn't actually lead to widespread
> practical use.
Well, Mach (Mac OS) is widely deployed, even though I wouldn't
consider Mach elegant. L4 is much newer and seems much better and
more advanced. The hype was a hype and the time not yet come for
microkernels (plus, everybody wanted to mix MKs with OOP).
> Students led by these sorts of academic interests wind up pretty
> desperately misdirected. And the issue _isn't_ about micro-kernels;
> it's about the near-myopia about those sorts of issues.
Yes, it's pretty off-topic ;)
> The Lisp-relevant thing is that academics seem wont to make grand
> statements about what Lisp is like based on what was true in Lisp 1.5
> that is devoid of any understanding of what has happened since.
I think it's more that professors try to cover a wide range of
prog. paradigms, which leads them to use Scheme for functional
prog (and Scheme *is* much more FP-oriented than Lisp). That way
students see the practical value. The lack of standard module
systems, and fragmentation in general doesn't help either. And as
an FP-language I prefer ML to Scheme. Lisp is a totally different
thing, and with a much different programming style, as seen in PCL.
OTOH my opinion is that universities (except perhaps American BSc
degrees and German FHs) shouldn't teach *practical* languages, but
teach principles/essentials of PLs (like the well-known courses of
the same name). Learning practical skills is the student's
responsibility (as one programmer told me when I was just out of
high school: you don't learn programming there; that you have to
do yourself. Turns out he was wrong, unfortunately).
There's two different routes to take: the scientific, fundamental
route, which makes you study the stuff behind the scenes, and
later think about practical aspects. That's what I have been
interested in all along. That's why I went to a regular
(non-practical) university. That's why I studied Scheme back
then. Now I'm happy with what I learnt and want to do more
practical stuff, also because graduation comes nearer and I have
to become more practical.
I agree that the academic way isn't right for everyone (in fact,
most people), but it's right for some. It's unfortunate that
German universities try to be Practical Schools (like FHs) and
give up on all academic aspects.
--
Don't let school interfere with your education. -- Mark Twain
Christopher Browne <········@acm.org> writes:
> Consider the "Torvalds/Tanembaum" flame war. Tanembaum indicated, as
> "eminent professor," that he wouldn't give good marks to someone
> creating a new monolithic OS kernel like Linux.
The marks might have been even worse if Linus had not used recursion
only--extra penalty for using arrays.
Paolo
--
Why Lisp? http://lisp.tech.coop/RtL%20Highlight%20Film
Recommended Common Lisp libraries/tools (see also http://clrfi.alu.org):
- ASDF/ASDF-INSTALL: system building/installation
- CL-PPCRE: regular expressions
- UFFI: Foreign Function Interface
F�rster vom Silberwald wrote:
> What I do not understand: where do you get your distorted misconception
> that a typical Scheme user is watching out for a tiny standard and
> nothing more.
That's not what Kenny said. What Kenny said is that "folks looking for
Lisp don't want a two-page spec with everything else unspecified." What
you have said in this thread indicates that you fall under this
category. (He also said that they are looking for Common Lisp. You
obviously don't fall under that category.)
It's a fact that Scheme code is harder to port across Scheme
implementations. It's also a fact that it's easier to port code across
Common Lisp implementations. That may or may not be relevant to you, but
it doesn't matter wrt to the truth of these statements.
I agree with Kenny that I can't imagine that people want to program
against nothing else but R5RS.
Here is what a Schemer has to say about this issue:
http://groups-beta.google.com/group/comp.lang.scheme/msg/7cf2275a837caf25
> Schneewittchen
> PS: I do not use something else than Bigloo, because I haven't found
> any other Scheme implementation which has all the features of what
> Bigloo actually will deliver me. But we know: different strokes for
> different people. Oh, I do not have to repeat it again: I haven't found
> any Lisp either which possess the features of Bigloo.
How thoroughly have you checked this? A quick glance at the Bigloo
website reveals things that are missing in Bigloo but that every Common
Lisp implementation provides. (For example multiple inheritance,
multimethods and method combinations.)
It's hard to understand what you want to say here, unless you only want
to make a statement that's trivially true. (Implementations of different
languages have different features?!?)
Pascal
--
2nd European Lisp and Scheme Workshop
July 26 - Glasgow, Scotland - co-located with ECOOP 2005
http://lisp-ecoop05.bknr.net/
Pascal Costanza wrote:
> How thoroughly have you checked this? A quick glance at the Bigloo
> website reveals things that are missing in Bigloo but that every
Common
> Lisp implementation provides. (For example multiple inheritance,
> multimethods and method combinations.)
Such kind of features might be important for you but not necessarily
for me. Much more important for me are things the way Bigloo
accomplishes its type system. Giving types in CommonLisp clutters up
code but that is not the case in Bigloo. Or pattern matching is another
case (Bigloo its pattern matching facility resembles any modern
functional language), etc.
Different strokes for different people. However, I get upset how
Lispers typically behave: please love us and our CommonLisp otherwise
you will end up in hell.
A more important question actually will be: if there won't be any
Scheme languages out there would that mean that all the current Scheme
users will use CommonLisp then?
Geschirrspülmittel
F�rster vom Silberwald wrote:
> Pascal Costanza wrote:
>
>>How thoroughly have you checked this? A quick glance at the Bigloo
>>website reveals things that are missing in Bigloo but that every
>
> Common
>
>>Lisp implementation provides. (For example multiple inheritance,
>>multimethods and method combinations.)
>
> Such kind of features might be important for you but not necessarily
> for me. Much more important for me are things the way Bigloo
> accomplishes its type system. Giving types in CommonLisp clutters up
> code but that is not the case in Bigloo. Or pattern matching is another
> case (Bigloo its pattern matching facility resembles any modern
> functional language), etc.
I think these things are very easily fixable in Common Lisp, but I also
understand that you probably don't want to spend your time on fixing
these things.
> Different strokes for different people. However, I get upset how
> Lispers typically behave: please love us and our CommonLisp otherwise
> you will end up in hell.
My impression is that "typical" Common Lispers and Schemers are alike:
They just want to be pragmatic and use a tool that does the best job for
their problems at hand. Many Schemers and Lispers switch between those
two languages under different circumstances and requirements, just as
they also use completely different languages. Some do this to a larger
extent than others. These are not breathtaking insights.
However, there is a history behind the Scheme vs. Common Lisp thing. If
you carefully read the Scheme specs and compare their release dates with
the timeline of the Common Lisp publications (CLtL1, CLtL2, ANSI Common
Lisp), you will notice that there are many remarks in the Scheme specs
that indicate that Scheme is "clearly" superior to Common Lisp in
several respects, and these are clearly reactions to the success of
Common Lisp. Common Lispers are just making fun of the fact that the
"official" Scheme (i.e., its spec) doesn't really meet its goals.
You are obviously using "full" Bigloo instead of "just" Scheme as well,
so the Scheme spec doesn't really help you either, right?
> A more important question actually will be: if there won't be any
> Scheme languages out there would that mean that all the current Scheme
> users will use CommonLisp then?
I can't speak for anyone else, but in a hypothetical world without
Common Lisp, I would definitely use Scheme. ;)
Pascal
--
2nd European Lisp and Scheme Workshop
July 26 - Glasgow, Scotland - co-located with ECOOP 2005
http://lisp-ecoop05.bknr.net/
Pascal Costanza wrote:
> You are obviously using "full" Bigloo instead of "just" Scheme as
well,
> so the Scheme spec doesn't really help you either, right?
>
> > A more important question actually will be: if there won't be any
> > Scheme languages out there would that mean that all the current
Scheme
> > users will use CommonLisp then?
>
> I can't speak for anyone else, but in a hypothetical world without
> Common Lisp, I would definitely use Scheme. ;)
As I wrote elsewhere I am not here to convince anone of anything.
I must honestly say: I haven't studied the Scheme spec in detail. When
I was starting my PhD (I wrote my Master thesis with Yorick) I have
been evaluating Clean whether it suits me or not. But the Clean
community is a dead horse. However, Clean was the first language which
I actually learned on my own expenses. Surely, we had some Fortran
classes but Fortran 95 is crap.
After that I tried CommonLisp. I didn't really like it. And then I
found Bigloo. And Bigloo is has been my working horse for 3 or 4 years
now.
By the way: I concede to all what you write (more or less). I have
never said that the Scheme community isn't cluttered up or porting
Scheme code is a bless.
Quite to the contrary: I learned Scheme (my Scheme is not that deep) by
means of the Bigloo manual. And there it is clearly noted when a
function or method belongs to the standard R5Rs or is Bigloo native. I
haven't count it but 70% or 80% in that particular manual are Bigloo
extensions. Hence it was always clear for me that the Scheme standard
itself is rather useless.
Schneewittchen
F�rster vom Silberwald wrote:
> After that I tried CommonLisp. I didn't really like it. And then I
> found Bigloo. And Bigloo is has been my working horse for 3 or 4 years
> now.
Nope. You did not try "Common Lisp" You tried *one* implementation of
Common Lisp which you did not like. I can live with that, there are
implementations of Common Lisp which I do not like. You just have not
tried the right one.
> By the way: I concede to all what you write (more or less). I have
> never said that the Scheme community isn't cluttered up or porting
> Scheme code is a bless.
>
> Quite to the contrary: I learned Scheme (my Scheme is not that deep) by
> means of the Bigloo manual. And there it is clearly noted when a
> function or method belongs to the standard R5Rs or is Bigloo native. I
> haven't count it but 70% or 80% in that particular manual are Bigloo
> extensions. Hence it was always clear for me that the Scheme standard
> itself is rather useless.
Yep. The problem is that if you factor out the parts that are not in
the CL ANSI standard, you will see that the remaining BigLoo
"extensions" will get you close to the ANSI standard itself.
Having said that, if you are happy with BigLoo, be happy with it. My
only advice is that you check out another CL implemetantion.
Cheers
--
Marco
F�rster vom Silberwald schrieb:
> You guys are permanetly commiting a huge mistake: everyone should fall
> in love with CommonLisp.
It is nice and shocking to see how humans function.
As long there is no common enemy (Java, C++) the Lisp fraction fights
itself.
Look at Linux - only a few people are using it. As long linux-people
discuss with someone from the windows world they all can work together
as a team and bash on Windows.
But don't leave them alone in a room. It would only take a short
amount of time until a Gentoo vs. Debian debate starts.
Same happens here. When no Javalist is in here a debate "CL vs. Scheme"
begins.
Andr�
--
Andr� Thieme <······························@justmail.de> writes:
> When no Javalist is in here a debate "CL vs. Scheme" begins.
It seems to me that it only happens when Scheme people
eschew their own newsgroup and come here trolling;
it never happens in the Scheme newsgroup, because no
CL people go over there trolling.
······@news.dtpq.com (Christopher C. Stacy) writes:
>> When no Javalist is in here a debate "CL vs. Scheme" begins.
>
> It seems to me that it only happens when Scheme people
> eschew their own newsgroup and come here trolling;
> it never happens in the Scheme newsgroup, because no
> CL people go over there trolling.
They are trolling on their own newsgroup instead.
Sorry, couldn't resist :-)
--
__("< Marcin Kowalczyk
\__/ ······@knm.org.pl
^^ http://qrnik.knm.org.pl/~qrczak/
F�rster vom Silberwald wrote:
> CommonLisp tried to find a least common denominator. But the latter is
> all what can be said about CommonLisp. I haven't understood the reasons
> why one would prefer CommonLisp over Scheme. What will CommonLisp
> deliver me what Bigloo, or Dr. Scheme or Chicken, etc. cannot deliver
> me on my platforms: Mac OSX, Linux, Unix.
Better portability across different Common Lisp implementations.
It's important to measure a language against its original goals. You can
find the goals of Common Lisp listed in Chapter 1.1 of "Common Lisp the
Language, 2nd Edition". (Google for it.)
To quote from that Chapter, "all the goals stated above have been
achieved, most notably that of portability. Moving large bodies of Lisp
code from one computer to another is now routine." This includes porting
from one CL implementation to another.
Compare this with the goal of Scheme, as stated in the Introduction of
R5RS: "Programming languages should be designed not by piling feature on
top of feature, but by removing the weaknesses and restrictions that
make additional features appear necessary." Do you really think that
Scheme has met that goal? Especially in the light of the proposed
additions for R6RS?
Of course, languages can have advantages and disadvantages beyond those
of the stated goals and non-goals...
Pascal
--
2nd European Lisp and Scheme Workshop
July 26 - Glasgow, Scotland - co-located with ECOOP 2005
http://lisp-ecoop05.bknr.net/
Jamie Border wrote:
> (I'm climbing into my asbestos suit as I write...)
>
> I dream of a LISP that...
>
> 1) Is open-source
> 2) Is cross-platform
> 3) Can deliver some kind of GUI (on X and Win32)
>
> I can't find one (and I've looked).
Where? You did not like http://common-lisp.net/project/cells-gtk/ with
CLisp? Not sure where the Gtk2 port for Mac OS X stands, nor CLisp's FFI
for Mac OS X, but I am sure volunteers are welcome.
kenny
--
Cells? Cello?: http://www.common-lisp.net/project/cells/
Cells-Gtk?: http://www.common-lisp.net/project/cells-gtk/
Why Lisp? http://lisp.tech.coop/RtL%20Highlight%20Film
"Doctor, I wrestled with reality for forty years, and I am happy to
state that I finally won out over it." -- Elwood P. Dowd
"Kenny Tilton" <·······@nyc.rr.com> wrote:
>
> Where? You did not like http://common-lisp.net/project/cells-gtk/ with
> CLisp? Not sure where the Gtk2 port for Mac OS X stands, nor CLisp's FFI
> for Mac OS X, but I am sure volunteers are welcome.
>
Unfortunately, due to cruel and unusual punishment (having to post using
Oulook Express - aaargh!) I failed to read the cells-gtk suggestion before
posting.
I don't think that I will be qualified to give a verdict on cells for some
time yet. I'm going to go away and write as much code as I can using cells
before I come back and whinge that nobody cares :-)
Jamie
> kenny
>
> --
> Cells? Cello?: http://www.common-lisp.net/project/cells/
> Cells-Gtk?: http://www.common-lisp.net/project/cells-gtk/
> Why Lisp? http://lisp.tech.coop/RtL%20Highlight%20Film
>
> "Doctor, I wrestled with reality for forty years, and I am happy to state
> that I finally won out over it." -- Elwood P. Dowd
>
On Mon, 9 May 2005 19:51:35 +0000 (UTC), Jamie Border <·····@jborder.com> wrote:
> (I'm climbing into my asbestos suit as I write...)
>
> I dream of a LISP that...
>
> 1) Is open-source
> 2) Is cross-platform
> 3) Can deliver some kind of GUI (on X and Win32)
4) Supports all the posix calls (esp select and the socket calls).
--
Please excuse my spelling as I suffer from agraphia. See
http://dformosa.zeta.org.au/~dformosa/Spelling.html to find out more.
Free the Memes.
>> I dream of a LISP that...
>>
>> 1) Is open-source
>> 2) Is cross-platform
>> 3) Can deliver some kind of GUI (on X and Win32)
>
DF> 4) Supports all the posix calls (esp select and the socket calls).
That would be next on my list. I'd like to think that I could do select
calls without having to use FFI.
Jamie
I suppose PLT-Scheme doesn't count not being Common Lisp.
Still, it meets requirements 1-3,
it has lots of Libraries and an active community,
(and you read SICP so you already know Scheme.)
s.
Jamie Border wrote:
> (I'm climbing into my asbestos suit as I write...)
>
> I dream of a LISP that...
>
> 1) Is open-source
> 2) Is cross-platform
> 3) Can deliver some kind of GUI (on X and Win32)
>
> I can't find one (and I've looked).
>
> Maybe this puts people off learning LISP. It nearly did the same to
me.
>
> It took me quite a while before I realised that learning LISP would
be the
> last language I would need to learn. After a couple of days reading
(On
> Lisp, SICP, PCL), I realised that I would rather use LISP than
anything
> else.
>
> Now I can't build a GUI without using a commercial LISP.
>
> I am currently poor, but am _still_ considering the $1100 for a
non-crippled
> LispWorks because I feel that strongly about the need for GUI.
>
> Richard Stallman and his crowd have given us an excellent free C
> compiler[1]. With the glut of [free] libraries that let me produce
> X-Windows applications, I can give somebody a list of URLs and they
can
> teach themselves how to use these. Then they can write useful
software.
>
> I can download Python and build some pretty little applications.
>
> I can download Java with a fairly good IDE and start developing
straight
> away.
>
> How hard would it be to provide something like this for an
open-source LISP?
>
> Would it help people's perception of LISP?
>
> I think so.
>
> Jamie
Jamie Border wrote:
> My work consists mostly of writing web front-ends for things, and I can see
> how I can win here with LISP (talking to stupid, inflexible, feature-poor
> legacy apps on minicomputers).
>
> The problem is, I can't find a good way of building GUIs.
>
> Before shouting, consider my predicament:
[...]
> Anybody got any ideas?
Have a look at Ltk (see link below) :)
Peter
--
pet project: http://dawn.netcologne.de
homepage: http://www.peter-herth.de
lisp stuff: http://www.peter-herth.de/lisp.html
get Ltk here: http://www.peter-herth.de/ltk/
Jamie Border schrieb:
Hallo
> My work consists mostly of writing web front-ends for things, and I can see
> how I can win here with LISP (talking to stupid, inflexible, feature-poor
> legacy apps on minicomputers).
> * I need something that I can use relatively unchanged on Linux/*BSD/Win32
I use newlisp version 6.3. It runs on win and with wine under linux (bsd
I do not use). It is good for my needs. It is small, easy to use and
quick. I found nothing like this till now. Lispworks-Capi is a bit
bigger, ... and commercial.
> Anybody got any ideas?
stefan
some need 3GHz-CPU for a keystroke every 30ms.
From: Friedrich Dominicus
Subject: Re: A (non-negative) 'what is LISP good for?'
Date:
Message-ID: <87y8aninpt.fsf@flarge.here>
"Jamie Border" <·····@jborder.com> writes:
> My work consists mostly of writing web front-ends for things, and I can see
> how I can win here with LISP (talking to stupid, inflexible, feature-poor
> legacy apps on minicomputers).
>
> The problem is, I can't find a good way of building GUIs.
>
> Before shouting, consider my predicament:
>
> * I can't use commercial LISPs, as the software _must_ be royalty-free, and
> preferably not reliant on closed-source systems (not my stipulation, but I
> can see the thinking)
That does not exclude commercial Lisps.
>
> * I need something that I can use relatively unchanged on
> Linux/*BSD/Win32
Well than the choices are quite limited.
>
> * Ideally, I'd like something that looks and feels like the target windowing
> system. I'd settle for something that looks like Java Swing.
This does not make sense. You want a web frontend that looks like
Windows?
Regards
Friedrich
--
Please remove just-for-news- to reply via e-mail.
"Friedrich Dominicus" <···················@q-software-solutions.de> wrote in
message ···················@flarge.here...
> "Jamie Border" <·····@jborder.com> writes:
>
>> My work consists mostly of writing web front-ends for things, and I can
>> see
>> how I can win here with LISP (talking to stupid, inflexible, feature-poor
>> legacy apps on minicomputers).
>>
>> The problem is, I can't find a good way of building GUIs.
>>
>> Before shouting, consider my predicament:
>>
>> * I can't use commercial LISPs, as the software _must_ be royalty-free,
>> and
>> preferably not reliant on closed-source systems (not my stipulation, but
>> I
>> can see the thinking)
> That does not exclude commercial Lisps.
>>
>> * I need something that I can use relatively unchanged on
>> Linux/*BSD/Win32
> Well than the choices are quite limited.
>>
>> * Ideally, I'd like something that looks and feels like the target
>> windowing
>> system. I'd settle for something that looks like Java Swing.
> This does not make sense. You want a web frontend that looks like
> Windows?
No. I have to build Win32 and X-based clients as well. If it was only a
question of web, then I wouldn't have had a problem. I managed to prove to
myself I could build a (simle-minded) web application with CLISP. Hence I
was encouraged to see if I could use lisp as a general-purpose replacement.
Incidentally, after all the talk I've seen about "Oh, Lisp is only good for
AI", I'd like to add my $0.02. I think that Lisp is not only good for AI,
it's good for human intelligence too.
Jamie
>
> Regards
> Friedrich
>
> --
> Please remove just-for-news- to reply via e-mail.
From: Friedrich Dominicus
Subject: Re: A (non-negative) 'what is LISP good for?'
Date:
Message-ID: <87mzr3gqck.fsf@flarge.here>
"Jamie Border" <·····@jborder.com> writes:
>
> No. I have to build Win32 and X-based clients as well. If it was only a
> question of web, then I wouldn't have had a problem. I managed to prove to
> myself I could build a (simle-minded) web application with CLISP. Hence I
> was encouraged to see if I could use lisp as a general-purpose
> replacement.
Ok, then I would opt for LispWorks. You can start using the Free
version and see if CAPI does it for you. I just can tell that CAPI is
supported on at least Windows, Linux, Mac OS X, you can do web
programming with LispWorks as you can with CLISP.
Regards
Friedrich
--
Please remove just-for-news- to reply via e-mail.
"Jamie Border" <·····@jborder.com> writes:
> My work consists mostly of writing web front-ends for things, and I can see
> how I can win here with LISP (talking to stupid, inflexible, feature-poor
> legacy apps on minicomputers).
>
> The problem is, I can't find a good way of building GUIs.
I don't understand. How could a way to build Graphical User Interface
help you in writting Web User Interfaces, front-ends for things?
If you want to write WUI, consider Uncommon-Web
http://common-lisp.net/project/ucw/
there are other possibilities, check: http://www.cliki.net/Web
If you want to write portable GUI I see three choices:
- clisp + ltk + tcl/tk, which gives the native look thanks to tcl wish.
clisp is the best choice for portability.
- clisp + any clx based lisp GUI (garnet, CLIM, etc) + a X server on
your target workstation (there are X servers on unix, on MacOSX,
on MS-Windows, etc).
- GTK (lgtk or lambda-gtk), may be with a Cell layer above, might
be a good solution but they don't work with clisp so you're
restricted to the target hosts of the CL implementations they
work on.
--
__Pascal Bourguignon__ http://www.informatimago.com/
Litter box not here.
You must have moved it again.
I'll poop in the sink.
PB> I don't understand. How could a way to build Graphical User Interface
PB> help you in writting Web User Interfaces, front-ends for things?
I can see that I did not explain my problem very well.
Actually, I typically have to build common web and GUI front-ends to stuff.
As far as GUI goes, I've started playing with both cells and the Tk stuff.
One particular product already has some Tk front-end, so for this, Peter
Hearth's stuff looks good.
When I understand cells, I'll hopefully be able to build some pretty gtk
apps too.
Thanks to everybody who's chipped in. I did not expect such a great
response to such a newbie question. I think I said this earlier in the
thread, but the reputation of Lispers as unapproachable and smug is totally
undeserved, IMHO.
Jamie
From: Christophe Rhodes
Subject: Re: A (non-negative) 'what is LISP good for?'
Date:
Message-ID: <sqzmv3bftq.fsf@cam.ac.uk>
"Jamie Border" <·····@jborder.com> writes:
> When I understand cells, I'll hopefully be able to build some pretty gtk
> apps too.
There are several other gtk wrappers which do not involve the cells
infrastructure and learning curve: for example, a relatively
one-to-one binding called lambda-gtk[1], and a strongly CLOS-based
version called clg[2].
Christophe
[1] <http://common-lisp.net/project/lambda-gtk/>
[2] <http://sourceforge.net/projects/clg>
From: Robert Uhl
Subject: Re: A (non-negative) 'what is LISP good for?'
Date:
Message-ID: <m3is1rup9w.fsf@4dv.net>
Christophe Rhodes <·····@cam.ac.uk> writes:
>
> There are several other gtk wrappers which do not involve the cells
> infrastructure and learning curve: for example, a relatively
> one-to-one binding called lambda-gtk[1], and a strongly CLOS-based
> version called clg[2].
Hmmm...
Thanks for the direct link to the clg page; CLiki only had info on the
old version (it's updated now). It appears that there's a recent (April
'05) gtk+ 2 version out.
Anyone written any comparisons of cells-gtk, lambda-gtk and clg?
--
Robert Uhl <http://public.xdi.org/=ruhl>
Eybershter undzer iz geshtanen! Avade er iz ufgeshtanen!
Pascal Bourguignon <···@informatimago.com> wrote:
> If you want to write portable GUI I see three choices:
>
> - clisp + ltk + tcl/tk, which gives the native look thanks to tcl wish.
> clisp is the best choice for portability.
>
> - clisp + any clx based lisp GUI (garnet, CLIM, etc) + a X server on
> your target workstation (there are X servers on unix, on MacOSX,
> on MS-Windows, etc).
>
> - GTK (lgtk or lambda-gtk), may be with a Cell layer above, might
> be a good solution but they don't work with clisp so you're
> restricted to the target hosts of the CL implementations they
> work on.
Are there any bindings for Qt, BTW ? CLOS would be the place to do so.
--
Didier Verna, ······@lrde.epita.fr, http://www.lrde.epita.fr/~didier
EPITA / LRDE, 14-16 rue Voltaire Tel.+33 (1) 44 08 01 85
94276 Le Kremlin-Bic�tre, France Fax.+33 (1) 53 14 59 22 ······@xemacs.org
For many years I have been very sucessfully been using an interface I
built to
tcl/tk in both CMUCL and Allegro on Linux/X86, Solaris/Sparc,
MacOSX/PPC,
and WinNT/X86. I have had virtually no problems moving from release to
release
of tcl/tk.
I evaluated the use of GTK1.x but then found that GTK2.x had major
incompatibilities and decided that GTK was still a moving target.
Pascal Bourguignon wrote:
> "Jamie Border" <·····@jborder.com> writes:
>
> > My work consists mostly of writing web front-ends for things, and I
can see
> > how I can win here with LISP (talking to stupid, inflexible,
feature-poor
> > legacy apps on minicomputers).
> >
> > The problem is, I can't find a good way of building GUIs.
>
> I don't understand. How could a way to build Graphical User
Interface
> help you in writting Web User Interfaces, front-ends for things?
>
>
> If you want to write WUI, consider Uncommon-Web
> http://common-lisp.net/project/ucw/
> there are other possibilities, check: http://www.cliki.net/Web
>
>
> If you want to write portable GUI I see three choices:
>
> - clisp + ltk + tcl/tk, which gives the native look thanks to tcl
wish.
> clisp is the best choice for portability.
>
> - clisp + any clx based lisp GUI (garnet, CLIM, etc) + a X server
on
> your target workstation (there are X servers on unix, on MacOSX,
> on MS-Windows, etc).
>
> - GTK (lgtk or lambda-gtk), may be with a Cell layer above, might
> be a good solution but they don't work with clisp so you're
> restricted to the target hosts of the CL implementations they
> work on.
>
>
> --
> __Pascal Bourguignon__
http://www.informatimago.com/
> Litter box not here.
> You must have moved it again.
> I'll poop in the sink.
"Jamie Border" <·····@jborder.com> writes:
> The problem is, I can't find a good way of building GUIs.
Check out LTK (http://www.peter-herth.de/ltk/).
> Before shouting, consider my predicament:
>
> * I can't use commercial LISPs, as the software _must_ be royalty-free, and
> preferably not reliant on closed-source systems (not my stipulation, but I
> can see the thinking)
>
> * I need something that I can use relatively unchanged on Linux/*BSD/Win32
LTK works on Win, Linux, *BSD and MacOSX. Probably more (that is: any platform
that has Tk and an ANSI CL).
You won't have to change the GUI code when jumping between platforms.
> * Ideally, I'd like something that looks and feels like the target windowing
> system. I'd settle for something that looks like Java Swing.
Tk looks (and is) more native than Swing.
LTK is (at the time) using sockets to talk to Tk. You should'nt expect it to
be as fast as FFI-ized user interfaces. I guess (without knowing) that you can
compare LTK with Java Swing, speedwise.
Also, there's Armed Bear CL. A Lisp within Java, that might suit your needs.
--
Oyvin
Oyvin Halfdan Thuv wrote:
> LTK is (at the time) using sockets to talk to Tk. You should'nt expect it to
> be as fast as FFI-ized user interfaces. I guess (without knowing) that you can
> compare LTK with Java Swing, speedwise.
>
Allow me some nitpicking :) Ltk uses pipes, not sockets (unless we talk
about ltk-remote). Ltk starts Tk as a subprocess of the Lisp process and
talks directly to its stdin/out via the stream associated with the
subprocess.
Peter
--
pet project: http://dawn.netcologne.de
homepage: http://www.peter-herth.de
lisp stuff: http://www.peter-herth.de/lisp.html
get Ltk here: http://www.peter-herth.de/ltk/
Hi Jamie.
I'm not sure if you're asking for a GUI builder or just a GUI library.
If just the latter, then possibly GNU Common Lisp (GCL) with the binding to
JAPI provided as standard with the Windows binary installer. JAPI is a java
GUI library:
http://savannah.gnu.org/projects/gcl/
http://ftp.gnu.org/gnu/gcl/binaries/stable/gcl_2.6.6.mingw32_ansi_japi_20050210.exe
See the example program here:
http://savannah.gnu.org/cgi-bin/viewcvs/gcl/gcl/
http://savannah.gnu.org/cgi-bin/viewcvs/gcl/gcl/japitest.lsp?rev=1.8&content-type=text/vnd.viewcvs-markup
for how to use the binding. No documentation other than that provided with
that example and the original C library, but both the library and binding
are dead simple to use:
http://www.japi.de/
> * I need something that I can use relatively unchanged on Linux/*BSD/Win32
I haven't tested this GUI binding under anything other than Windows but as
JAPI is cross-platform I expect it should work fine on those Unixy platforms
you mention.
> * Ideally, I'd like something that looks and feels like the target
> windowing system. I'd settle for something that looks like Java Swing.
That's JAPI.
Cheers
Mike Thomas.