From: Philip Chu
Subject: copy-file
Date: 
Message-ID: <PCHU.94May12110013@kariba.bbn.com>
I just noticed in CLtL2 that although RENAME-FILE and DELETE-FILE are
available in Common Lisp, there is no function for copying a file. Is
there are particular reason for such an omission, or was it just not
considered important?
-- 
Phil Chu
Internet: ····@bbn.com

From: Mike McDonald
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <Cqo6DJ.5xC@jabba.ess.harris.com>
In article <·········@nz12.rz.uni-karlsruhe.de>, ······@ma2s2.mathematik.uni-karlsruhe.de (Bruno Haible) writes:
|> Fernando Mato Mira <········@di.epfl.ch> wrote:
|> 
|> > > CLX is distributed with X11R6. The directory rearrangement makes it
|> > > harder to find in the 'contrib' directory tree. 
|> >
|> > OK. Here is what I really meant: that it is now a non-standard part
|> > of X11. It's not just a directory move.
|> 
|> I don't think so. It *is* just a directory move. The most likely
|> explanation is that the X Consortium does not want to maintain this
|> *sample* implementation of an interface between CL and X11. In fact, most
|> changes to CLX in the past few years were ports to different CL platforms,
|> done by the CL vendors. The X11 protocol won't change, so maintaining CLX
|> is the vendors' duty, not the X Consortium's.
|> 
|> 
|>                     Bruno Haible
|>                     ······@ma2s2.mathematik.uni-karlsruhe.de
|> 

  I believe it is more of an indication of a lack of interest in making CLX a
Consortium standard. There was an effort a few years back at making it one. It
died an early death due to a lack of interest. I think Bob Schiefler was
probably the only Consortium member who had an real interest in Lisp. He
probably doesn't have the time or energy to continue pushing it. The
Consortium's membership seems to be solely concerned with C (with a few pushing
C++). If I were to make a prediction by looking into my fuzzy crytal ball, I'd
guess that CLX will be from even the Consortium's contrib directory in a few
years. 

  Mike McDonald				Advanced Technology Dept.	
					Harris Corp.
  Email: ···@trantor.harris-atd.com	M.S. 16-1912
  Voice: (407) 727-5060			P.O. Box 37
  Fax:   (407) 729-3363			Melbourne, Florida 32902
From: Marco Antoniotti
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <MARCOXA.94Jun1112330@graphics.cs.nyu.edu>
'(Dear Lispers)

Mike McDonald is right when he does not see a future for CLX with the
X Consortium. As it is the situation *is* grim.

At the same time, I disagree with Bruno when he says that the user
community should expect "the Vendors" to keep up the support of
CLX. It does not look like they have done it for years. Moreover, also
the "non vendors" do not seem interested in keeping up a "common
resource". I surely do not want to be unpolite toward people who have
put a lot of work into making PD or Free CL's, but, if you check the
CLX version released with CMUCL 17e and the CLX released with X11R6,
it turns out that they ARE slightly different and that the CMU
specific code is not included into the X Consortium version.

I do not know why and how this happened.

As an "end user" I would just like to see a "higher" degree of
homogeneity for what is certainly a common ground for the Common
Lisp community.

Yet, it is also clear that there is not anymore enough "free" support for
Common Lisp. Hence, the community *must* rely on "the Vendors" good
will to have access to good pieces of basic software like CLX (why not
CLIM? A question asked many times) for free (or at least at a
reasonable price  - i.e. what a *small* company or university lab could
pay without problems).

Happy Lisping
--
Marco Antoniotti - Resistente Umano
-------------------------------------------------------------------------------
Robotics Lab		| room: 1220 - tel. #: (212) 998 3370
Courant Institute NYU	| e-mail: ·······@cs.nyu.edu

...e` la semplicita` che e` difficile a farsi.
...it is simplicity that is difficult to make.
				Bertholdt Brecht
From: Mike McDonald
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <Cqs1Jn.5t9@jabba.ess.harris.com>
In article <····················@graphics.cs.nyu.edu>, ·······@graphics.cs.nyu.edu (Marco Antoniotti) writes:
|> 
|> '(Dear Lispers)
|> 
|> Mike McDonald is right when he does not see a future for CLX with the
|> X Consortium. As it is the situation *is* grim.
|> 
|> At the same time, I disagree with Bruno when he says that the user
|> community should expect "the Vendors" to keep up the support of
|> CLX. It does not look like they have done it for years. Moreover, also
|> the "non vendors" do not seem interested in keeping up a "common
|> resource". I surely do not want to be unpolite toward people who have
|> put a lot of work into making PD or Free CL's, but, if you check the
|> CLX version released with CMUCL 17e and the CLX released with X11R6,
|> it turns out that they ARE slightly different and that the CMU
|> specific code is not included into the X Consortium version.
|> 
|> I do not know why and how this happened.

  As the Consortium lost interest and support for CLX, a gentleman at Symbolics
(sorry, I can't remember his name. Too much C++ coding ...) volunteered to
maintain CLX by incorporating bug fixes submitted by interested individuals. With
Symbolics going the way of the Doodoo bird, I don't believe that this individual
is still there. There's also ben a drastic decrease in interest in using lisp
with X (and in lisp in general) since the time we tried to standardize CLX. I
think in the last year I've received on bug report and maybe two source code
requests on the bug-clx mailing list. The lisp vendors seemed to only have
interest in the proprietary CLIM effort. 

  Mike McDonald				Advanced Technology Dept.	
					Harris Corp.
  Email: ···@trantor.harris-atd.com	M.S. 16-1912
  Voice: (407) 727-5060			P.O. Box 37
  Fax:   (407) 729-3363			Melbourne, Florida 32902


P.S. Seems like the name might be Charles Hornig <······@stony-brook.scrc.symbolics.com>.
From: Marco Antoniotti
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <MARCOXA.94Jun3105005@graphics.cs.nyu.edu>
In article <··········@jabba.ess.harris.com> ···@trantor.harris-atd.com (Mike McDonald) writes:

   Newsgroups: comp.lang.lisp
   From: ···@trantor.harris-atd.com (Mike McDonald)
   Sender: ······@jabba.ess.harris.com (Usenet News Feed)
   Nntp-Posting-Host: v3a.ess.harris.com
   Organization: Advanced Technology Dept, Harris Corp, Melbourne, FL
   References: <··················@kariba.bbn.com> <··········@underdog.jpl.nasa.gov> <····················@graphics.cs.nyu.edu>
   Date: Thu, 2 Jun 1994 15:54:59 GMT
   Lines: 37
   Xref: slinky.cs.nyu.edu comp.lang.lisp:2649 comp.sys.sgi.graphics:3935

   In article <····················@graphics.cs.nyu.edu>, ·······@graphics.cs.nyu.edu (Marco Antoniotti) writes:
	...

   |> I do not know why and how this happened.

     As the Consortium lost interest and support for CLX, a gentleman
   at Symbolics 
   (sorry, I can't remember his name. Too much C++ coding ...) volunteered to
   maintain CLX by incorporating bug fixes submitted by interested
   individuals. With 
   Symbolics going the way of the Doodoo bird, I don't believe that
   this individual 
   is still there. There's also ben a drastic decrease in interest in
   using lisp 
   with X (and in lisp in general) since the time we tried to
   standardize CLX. I 
   think in the last year I've received on bug report and maybe two source code
   requests on the bug-clx mailing list. The lisp vendors seemed to only have
   interest in the proprietary CLIM effort. 

Again, what you say is right. Yet, allow me to note that I did not know
of any effort to standardize CLX. This means that no posting on the
subject were send over to comp.lang.lisp (of course I might have
missed them - but I think to be pretty good at monitoring this
newsgroup).

This is a Comma 22 situation. Most people over here would REALLY like
to see a pretty standard CLX (and CLIM or similar) environment, but
they do not have the energy to work on it (myself included). At the
same time the vendors are not interested in sharing bits an pieces of
their code, that is CLX bug fixes etc.

The future seems grim, unless
1) some volunteer will be able to take upon her/himself the burden of
   maintaining PD pieces of CL*
2) pressure will be put on the vendors to be more cooperative in their
   development of the language and environment.

Happy Lisping

--
Marco Antoniotti - Resistente Umano
-------------------------------------------------------------------------------
Robotics Lab		| room: 1220 - tel. #: (212) 998 3370
Courant Institute NYU	| e-mail: ·······@cs.nyu.edu

...e` la semplicita` che e` difficile a farsi.
...it is simplicity that is difficult to make.
				Bertholdt Brecht
From: Bruno Haible
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <2sv63f$3r9@nz12.rz.uni-karlsruhe.de>
Mike McDonald <···@trantor.harris-atd.com> wrote:
>
>  ... a gentleman at Symbolics
> (sorry, I can't remember his name. Too much C++ coding ...) volunteered to
> maintain CLX by incorporating bug fixes submitted by interested individuals.
> ...
> P.S. Seems like the name might be Charles Hornig
> <······@stony-brook.scrc.symbolics.com>.

Mr. Hornig tells me:
  Your information is out of date.  I had to give CLX up when I changed
  employers about two years ago.  I don't know who works with CLX any more.

So, who volunteers to take over the maintainment of CLX ?

Not me  --  Bruno Haible <······@ma2s2.mathematik.uni-karlsruhe.de>
From: Mike McDonald
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <CqzJty.MD1@jabba.ess.harris.com>
In article <··········@nz12.rz.uni-karlsruhe.de>, ······@ma2s2.mathematik.uni-karlsruhe.de (Bruno Haible) writes:
|> Mike McDonald <···@trantor.harris-atd.com> wrote:
|> >
|> >  ... a gentleman at Symbolics
|> > (sorry, I can't remember his name. Too much C++ coding ...) volunteered to
|> > maintain CLX by incorporating bug fixes submitted by interested individuals.
|> > ...
|> > P.S. Seems like the name might be Charles Hornig
|> > <······@stony-brook.scrc.symbolics.com>.
|> 
|> Mr. Hornig tells me:
|>   Your information is out of date.  I had to give CLX up when I changed
|>   employers about two years ago.  I don't know who works with CLX any more.
|> 
|> So, who volunteers to take over the maintainment of CLX ?
|> 
|> Not me  --  Bruno Haible <······@ma2s2.mathematik.uni-karlsruhe.de>
|> 

  Yes, that was my point. The CLX standardization effort was 4 years ago. Charles
vounteered to maintain CLX after the effort failed. Nobody as far as I know is
maintaining a central copy of CLX. I've sent Email to Bob Scheifler at the X
Consortium asking him what the statis of CLX is. I'll post whatever I find out.

  Mike McDonald				Advanced Technology Dept.	
					Harris Corp.
  Email: ···@trantor.harris-atd.com	M.S. 16-1912
  Voice: (407) 727-5060			P.O. Box 37
  Fax:   (407) 729-3363			Melbourne, Florida 32902
From: Marcus Daniels
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <2svsvn$6iv@ursula.ee.pdx.edu>
·······@graphics.cs.nyu.edu (Marco Antoniotti) writes:

>Yet, it is also clear that there is not anymore enough "free" support for
>Common Lisp.

Today, is it realistic or even desirable to think that LISP users have LISP
interface toolkits?  Is it even worth putting effort into CLX?

Yet, other than Fresco, NEXTSTEP, and perhaps Win32 (Chicago?),
there doesn't seem to be much potential for aggreement in the OOP GUI
world.

For all the neglect this area has supposedly received recently by the
LISP community, I'm somewhat underwhelmed by the `modern' GUI alternatives
(commercial or otherwise).
From: Mike McDonald
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <Cr1144.GM3@jabba.ess.harris.com>
In article <··········@ursula.ee.pdx.edu>, ······@ee.pdx.edu (Marcus Daniels) writes:
|> ·······@graphics.cs.nyu.edu (Marco Antoniotti) writes:
|> 
|> >Yet, it is also clear that there is not anymore enough "free" support for
|> >Common Lisp.
|> 
|> Today, is it realistic or even desirable to think that LISP users have LISP
|> interface toolkits?  Is it even worth putting effort into CLX?

  Realistic? No. Desirable? Yes. Worth the effort? Probably not. I don't believe
that the vendors would support the effort. And I don't think there are enough
independant lisp groups to support the effort either. If CMU could have stayed
with Lisp instead of having to jump ship for Dylan, there might still have been a
chance.

|> Yet, other than Fresco, NEXTSTEP, and perhaps Win32 (Chicago?),
|> there doesn't seem to be much potential for aggreement in the OOP GUI
|> world.
|> 
|> For all the neglect this area has supposedly received recently by the
|> LISP community, I'm somewhat underwhelmed by the `modern' GUI alternatives
|> (commercial or otherwise).
|> 

  I am too. But then I haven't seen anything much better in lisp.

  Mike McDonald				Advanced Technology Dept.	
					Harris Corp.
  Email: ···@trantor.harris-atd.com	M.S. 16-1912
  Voice: (407) 727-5060			P.O. Box 37
  Fax:   (407) 729-3363			Melbourne, Florida 32902
From: Fernando Mato Mira
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <2t21qo$mep@disuns2.epfl.ch>
In article <··········@jabba.ess.harris.com>, ···@trantor.harris-atd.com (Mike McDonald) writes:
> |> 
> |> Today, is it realistic or even desirable to think that LISP users have LISP
> |> interface toolkits?  Is it even worth putting effort into CLX?
> 
>   Realistic? No. Desirable? Yes. Worth the effort? Probably not. I don't believe

Given that the CLX data structures are opaque, wouldn't it be more reasonable
to just implement the CLX API on top of Xlib? The fact that it's not so has only
given me headaches (no Xinput support, no GLX support, no
straight-to-pipeline OpenGL support).
 
  I don't know how the Mac and Win32 work, so let me ask a question:
Would it be possible to redefine CLX as a low-level platform-independent
GUI API for taking care of things like getting events, opening windows and
making requests, but not implementing any high-level concepts (presentations,
interactors, etc.), or trying to remap generic widgets to the ones
available on the target system?

  I dont't care about CLIM. At least not while I keep on feeling that
"closed world AI-only lisp machine attitude" from the vendors/proponents.


-- 
F.D. Mato Mira                           
Computer Graphics Lab    ········@epfl.ch
EPFL                     FAX: +41 (21) 693-5328

"AIer by training, VRer by hacking"
From: Marco Antoniotti
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <MARCOXA.94Jun7142550@graphics.cs.nyu.edu>
In article <··········@disuns2.epfl.ch> ········@di.epfl.ch (Fernando Mato Mira) writes:


   From: ········@di.epfl.ch (Fernando Mato Mira)
   Newsgroups: comp.lang.lisp,comp.sys.sgi.graphics,comp.windows.garnet
   Date: 7 Jun 1994 14:56:24 GMT
   Organization: Ecole Polytechnique Federale de Lausanne
   Lines: 37
   Distribution: world



   Given that the CLX data structures are opaque, wouldn't it be more
   reasonable 
   to just implement the CLX API on top of Xlib? The fact that it's
   not so has only 
   given me headaches (no Xinput support, no GLX support, no
   straight-to-pipeline OpenGL support).

I do not know how the GL support is being tied into X, but the issues
here are different.

CLX is the equivalent of XLIB. As such it is mainly an implementation
of the X protocol. What would be reasonable is to add to CLX those
extensions to XLIB that have appeared in the recent releases.

     I don't know how the Mac and Win32 work, so let me ask a question:
   Would it be possible to redefine CLX as a low-level platform-independent
   GUI API for taking care of things like getting events, opening windows and
   making requests, but not implementing any high-level concepts
   (presentations, 
   interactors, etc.), or trying to remap generic widgets to the ones
   available on the target system?

That is certainly possible, but it is not going to be CLX. The goal of
CLIM seems to be exactly that. Yet, CLIM is a closed world for those
without a serious budget. The goal of CLUE and CLIO was almost
the same and it is a shame that they were not adopted as the standard
CL Implementation of a widget set.

The other major problem lies in the "standards" out there. Look and
Feel are NOT things to be overlooked. CMUCL and CLM take a step into
this direction by providing a "language" binding for Motif (and, I
suppose, for any other Window System / Widget Set running on
UN*X and sockets). But CLX is still needed in order to do the low
level stuff (and it should be). Macs and Windows systems are beyond my
knowledge.

   I dont't care about CLIM. At least not while I keep on feeling that
   "closed world AI-only lisp machine attitude" from the vendors/proponents.

Once again the vendors should at least realize that a release of their
code to the Net would be beneficial for the entire community and for
themselves.

Happy Lisping
--
Marco Antoniotti - Resistente Umano
-------------------------------------------------------------------------------
Robotics Lab		| room: 1220 - tel. #: (212) 998 3370
Courant Institute NYU	| e-mail: ·······@cs.nyu.edu

...e` la semplicita` che e` difficile a farsi.
...it is simplicity that is difficult to make.
				Bertholdt Brecht
From: Mike McDonald
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <Cr30LE.FDA@jabba.ess.harris.com>
In article <····················@graphics.cs.nyu.edu>, ·······@graphics.cs.nyu.edu (Marco Antoniotti) writes:
|> In article <··········@disuns2.epfl.ch> ········@di.epfl.ch (Fernando Mato Mira) writes:
|> 
|> 
|>    From: ········@di.epfl.ch (Fernando Mato Mira)
|>    Newsgroups: comp.lang.lisp,comp.sys.sgi.graphics,comp.windows.garnet
|>    Date: 7 Jun 1994 14:56:24 GMT
|>    Organization: Ecole Polytechnique Federale de Lausanne
|>    Lines: 37
|>    Distribution: world
|> 
|> 
|> 
|>    Given that the CLX data structures are opaque, wouldn't it be more
|>    reasonable 
|>    to just implement the CLX API on top of Xlib? The fact that it's
|>    not so has only 
|>    given me headaches (no Xinput support, no GLX support, no
|>    straight-to-pipeline OpenGL support).
|> 
|> I do not know how the GL support is being tied into X, but the issues
|> here are different.

  Since OpenGl is anothe rone of those proprietary "open" systems, I'd be against
supporting it. I'd rather see PEX supported by lisp. (Actually, I'd like OpenGL
to be donated to the X Consortium as a truely open standard.)

|> CLX is the equivalent of XLIB. As such it is mainly an implementation
|> of the X protocol. What would be reasonable is to add to CLX those
|> extensions to XLIB that have appeared in the recent releases.

  Reasonable? No. Desirable? Yes. :-) There are about a half dozen or so
"standard" extensions to the X protocol. Most are relatively small and (quite
niavely he says) probably "easy" to implement. Others are huge (PEX). Others are
so convoluted that they're hard to understand (Fernando Mato Mira says the XInput
extension is "poorly documented").

|>      I don't know how the Mac and Win32 work, so let me ask a question:
|>    Would it be possible to redefine CLX as a low-level platform-independent
|>    GUI API for taking care of things like getting events, opening windows and
|>    making requests, but not implementing any high-level concepts
|>    (presentations, 
|>    interactors, etc.), or trying to remap generic widgets to the ones
|>    available on the target system?
|> 
|> That is certainly possible, but it is not going to be CLX. The goal of
|> CLIM seems to be exactly that. Yet, CLIM is a closed world for those
|> without a serious budget. The goal of CLUE and CLIO was almost
|> the same and it is a shame that they were not adopted as the standard
|> CL Implementation of a widget set.
|> 
|> The other major problem lies in the "standards" out there. Look and
|> Feel are NOT things to be overlooked. CMUCL and CLM take a step into
|> this direction by providing a "language" binding for Motif (and, I
|> suppose, for any other Window System / Widget Set running on
|> UN*X and sockets). But CLX is still needed in order to do the low
|> level stuff (and it should be). Macs and Windows systems are beyond my
|> knowledge.

  I think one has to make a distinction between finding a Lisp solution to X and
finding one for GUIs in general. If all you want to do is to be able to take full
advantage of the existing X protocols and implementations, that leads you down
one road. If you want to be able to write a single GUIed application that will
work across X, Windows, and Macs, that takes you down a different path. I believe
that there is sufficient differences between these windowing systems that trying
to solve both problems with one solution wouldn't be feasible. Putting in enough
code to smooth over the incompatibilities would leave you too far from the native
system. If you're willing to distance yourself from the underlying system, I
think it is possible (and may even be reasonable).

  Mike McDonald				Advanced Technology Dept.	
					Harris Corp.
  Email: ···@trantor.harris-atd.com	M.S. 16-1912
  Voice: (407) 727-5060			P.O. Box 37
  Fax:   (407) 729-3363			Melbourne, Florida 32902
From: Fernando Mato Mira
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <2t4pl3$5fo@disuns2.epfl.ch>
In article <··········@jabba.ess.harris.com>, ···@trantor.harris-atd.com (Mike McDonald) writes:

> so convoluted that they're hard to understand (Fernando Mato Mira says the XInput
> extension is "poorly documented").

Wow, wow, wow. What's "poorly documented" is the CLX extension-definition
facility, not XInput. And then, it's not powerful enough, so you have to start
wandering about the internal CLX code to try to figure out how that could
be extended. It would be so much easier to just call XNextEvent or whatever, 
and let it take care of tranlating from wire to app!

--
F.D. Mato Mira                           
Computer Graphics Lab    ········@epfl.ch
EPFL                     FAX: +41 (21) 693-5328
From: Fernando Mato Mira
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <2t4rml$7ft@disuns2.epfl.ch>
In article <··········@disuns2.epfl.ch>, ········@di.epfl.ch (Fernando Mato Mira) writes:

> be extended. It would be so much easier to just call XNextEvent or whatever, 
> and let it take care of tranlating from wire to app!

Er, of course you still have to define how to access the extra
info in the event. Lisps with a define-c-type facility can 
easily get inside the event directly. For others, someone will have
to add some silly C struct access code and FFI decls.

-- 
F.D. Mato Mira                           
Computer Graphics Lab    ········@epfl.ch
EPFL                     FAX: +41 (21) 693-5328
From: Mike McDonald
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <Cr3Eo8.Kos@jabba.ess.harris.com>
In article <··········@disuns2.epfl.ch>, ········@di.epfl.ch (Fernando Mato Mira) writes:
|> In article <··········@jabba.ess.harris.com>, ···@trantor.harris-atd.com (Mike McDonald) writes:
|> 
|> > so convoluted that they're hard to understand (Fernando Mato Mira says the XInput
|> > extension is "poorly documented").
|> 
|> Wow, wow, wow. What's "poorly documented" is the CLX extension-definition
|> facility, not XInput. And then, it's not powerful enough, so you have to start
|> wandering about the internal CLX code to try to figure out how that could
|> be extended. It would be so much easier to just call XNextEvent or whatever, 
|> and let it take care of tranlating from wire to app!
|> 
|> --
|> F.D. Mato Mira                           
|> Computer Graphics Lab    ········@epfl.ch
|> EPFL                     FAX: +41 (21) 693-5328

  Whoops! BIG whoops! My mistake. Maybe I should have said English is "poorly
documented"! :-) 


  Mike McDonald				Advanced Technology Dept.	
					Harris Corp.
  Email: ···@trantor.harris-atd.com	M.S. 16-1912
  Voice: (407) 727-5060			P.O. Box 37
  Fax:   (407) 729-3363			Melbourne, Florida 32902
From: Mike McDonald
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <Cr1IGM.Mu1@jabba.ess.harris.com>
In article <··········@disuns2.epfl.ch>, ········@di.epfl.ch (Fernando Mato Mira) writes:
|> In article <··········@jabba.ess.harris.com>, ···@trantor.harris-atd.com (Mike McDonald) writes:
|> > |> 
|> > |> Today, is it realistic or even desirable to think that LISP users have LISP
|> > |> interface toolkits?  Is it even worth putting effort into CLX?
|> > 
|> >   Realistic? No. Desirable? Yes. Worth the effort? Probably not. I don't believe
|> 
|> Given that the CLX data structures are opaque, wouldn't it be more reasonable
|> to just implement the CLX API on top of Xlib? The fact that it's not so has only
|> given me headaches (no Xinput support, no GLX support, no
|> straight-to-pipeline OpenGL support).

  Linking to the C version of Xlib only will work on machines that have C. Not
all Lisps have access to C. CLX was an attempt to build a lisp equivilant of
Xlib. Extensions used to work with CLX. Now, one had to actually implement the
extensions interface inorder to use it. Unfortunately, no one has been
implementing the standard X extensions.

|>   I don't know how the Mac and Win32 work, so let me ask a question:
|> Would it be possible to redefine CLX as a low-level platform-independent
|> GUI API for taking care of things like getting events, opening windows and
|> making requests, but not implementing any high-level concepts (presentations,
|> interactors, etc.), or trying to remap generic widgets to the ones
|> available on the target system?

  I don't know much about programming windows and graphics on the Mac. Franz has
an interface to Windows called Common Graphics. Having looked at docs for Common
Graphics, it doesn't seem like it would be too hard to implement a version for X.
I've thought about doing this but always stop with "What's the point?".

|>   I dont't care about CLIM. At least not while I keep on feeling that
|> "closed world AI-only lisp machine attitude" from the vendors/proponents.

  Me too!

  Mike McDonald				Advanced Technology Dept.	
					Harris Corp.
  Email: ···@trantor.harris-atd.com	M.S. 16-1912
  Voice: (407) 727-5060			P.O. Box 37
  Fax:   (407) 729-3363			Melbourne, Florida 32902
From: Marcus Daniels
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <MARCUS.94Jun7204345@tdb.ee.pdx.edu>
In-reply-to: ···@trantor.harris-atd.com's message of Tue, 7 Jun 1994 18:38:45 GMT


mike>  I don't know much about programming windows and graphics on the Mac. Franz has
mike> an interface to Windows called Common Graphics. Having looked at docs for Common
mike> Graphics, it doesn't seem like it would be too hard to implement a version for X.
mike> I've thought about doing this but always stop with "What's the point?".

Franz also had an Objective C interface that worked with NEXTSTEP (GUI) objects.
(not an implementation of a display server protocol, like CLX -- a glorified FFI).
IMHO, protocols are fine, but not as a substitute for integration features.

fernando>   I dont't care about CLIM. At least not while I keep on feeling that
fernando> "closed world AI-only lisp machine attitude" from the vendors/proponents.

mike> Me too!

Ditto.

On the face of it, Garnet seems to be a remarkable piece of work. 
Odd that there isn't more interest in it.  Has anyone who has developed with it
comment on this state of affairs?
From: Marco Antoniotti
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <MARCOXA.94Jun8095605@graphics.cs.nyu.edu>
In article <···················@tdb.ee.pdx.edu> ······@ee.pdx.edu (Marcus Daniels) writes:


   From: ······@ee.pdx.edu (Marcus Daniels)
   Newsgroups: comp.lang.lisp,comp.sys.sgi.graphics,comp.windows.garnet
   Date: 08 Jun 1994 03:43:44 GMT
   Organization: Portland State University
   Lines: 20
   Reply-To: marcus


   Franz also had an Objective C interface that worked with NEXTSTEP
   (GUI) objects. 
   (not an implementation of a display server protocol, like CLX -- a
   glorified FFI). 

CLX is not a glorified FFI. It is one of the two extra-C
interfaces of the X protocol.

   IMHO, protocols are fine, but not as a substitute for integration
   features.

That is another problem. The discussion on comp.std.lisp died out on
this subject because nobody was willing to volunteer (at least this is
what I gathered) to come up with a comparative study of different FFI
interfaces. I had access only to CMUCL documentation at that time.
Work seems still to be needed.

   fernando>   I dont't care about CLIM. At least not while I keep on
               feeling that 
   fernando> "closed world AI-only lisp machine attitude" from the
             vendors/proponents. 

   mike> Me too!

   Ditto.

The CLIM specification is good. What it lack is the recognition that either
you tie in MOTIF (or similar) or you are dead.

   On the face of it, Garnet seems to be a remarkable piece of work. 
   Odd that there isn't more interest in it.  Has anyone who has
   developed with it 
   comment on this state of affairs?

What has always kept me from installing and using the >8Mb Garnet
stuff is that they do not use CLOS. KR may be a fine language, but
once again it is an extra layer that introduces extra headaches.
No offense intended to the Garnet folks.

Happy lisping
--
Marco Antoniotti - Resistente Umano
-------------------------------------------------------------------------------
Robotics Lab		| room: 1220 - tel. #: (212) 998 3370
Courant Institute NYU	| e-mail: ·······@cs.nyu.edu

...e` la semplicita` che e` difficile a farsi.
...it is simplicity that is difficult to make.
				Bertholdt Brecht
From: Richard Billington
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <BUFF.94Jun10165710@pravda.cc.gatech.edu>
   fernando> "closed world AI-only lisp machine attitude" from the
             vendors/proponents. 

   mike> Me too!

   Ditto.

 The CLIM specification is good. What it lack is the recognition that either
 you tie in MOTIF (or similar) or you are dead.

I'd like to interject in this line of talk: CLIM 2 does specifically tie into
toolboxes (for look and feel issues) - MOTIF, the Mac toolbox, and Windows. It
does so at a high enough level that code can be readily ported from one
platform to another - this was a specific objective.

As to the "closed world" etc. There's a conference once a year called LUV (the
Lisp Users and Vendors Conference). At the first such conference the users made
it very clear that they wanted CLIM adopted as a defacto standard by the lisp
industry. This was USER pressure, not vendor pressure.

As an originator and organizer of the earliest of these conferences and still
being in touch with the organization of the conference today, I'd like to say
that we have done EVERYTHING WE CAN THINK OF to open this conference up to ALL
LISP USERS (not just cl, but scheme, dylan, elisp, etc). We have tutorials on
using elisp (in fact, RMS is a featured speaker this year along with John
McCarthy) as well as on cl topics.  We had David Moon talk about Dylan last
year. We've had not just Symbolics, Harlequin, Franz, and Lucid represented as
vendors, but also Apple, Kyoto, Gold Hill, Sapiens, and others. We've had talks
from the CMU lisp group.

If you want things to change in the lisp world, come to the conference, make
your voice heard. The vendors are there PRIMARILY to listen. They send their
development people and their top-management, NOT their sales force. Come to the
conference - get in their face about prices. Open up the "closed world" of CLIM
- the major developers will be there, and they've given (and I mean really
given) a lot to get CLIM to where it is - come talk to them and vent your
frustrations. Come to LUV and do something constructive about the future of
Lisp.

for more info on LUV (Aug 15 - 19, 1994, Berkeley, CA)
·············@ai.sri.com
···@harlequin.com           <= co-chair
····@chesapeake.ads.com     <= co-chair
From: Marco Antoniotti
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <MARCOXA.94Jun11154649@graphics.cs.nyu.edu>
Many thanks to Richard Billington for his posting. I find it
very interesting.

Most probably I will not go to LUV-94 (I know what it is and why it is
there): my decision to go back to school (hence my limited finances)
and my *real* research interests lead me in other places.

I am not interested whether CLIM 2 ties in with various toolboxes
and/or widget sets. The specification is really vague (at least the
one I downloaded following the FAQ directions) on the subject of
"available" toolbox support.

But this is off the track. Since I will not be able to be at the
conference in August, I would like to vent my frustration to the
vendors here. I sincerely hope that some of them are monitoring this
newsgroup. (Especially the "development people and their
top-management, NOT their sales force") (Note: actually this is the
wrong attitude: it should be the "sales force" that should be
bombarded by the lusers).

First of all I think that the vendors should take upon themselves the
burden to maintain and MAKE AVAILABLE pieces of free software that
have been floating around for years. CLX is the major one. If CLIM
wants to be a substitute (as it looks like) they should provide
support to the independent groups like the CMU folks, the GNU CL
group (ex AKCL) and Bruno Haible (CLISP) in order to make them able to
produce a free CLIM version, e.g. by donating the "high" level code.

The objective of such policy is to spread the use of Lisp. The vendors
are now suffering from a shrinking market, not by competiton among
themselves.

Having said so, I summarize my requests for the vendors and developers
here:

1) What should I - an end user and developer of applications - do with
   CLX?

2) How can I get a LOW priced implementation of CLIM? (e.g. zerop ?).

3) Will I ever be able to use a single FFI from every CL
   implementation (at least towards C - C++ being totally out of the
   question) ?

Since I am at it, here my list continues (I know I am boring: I
produced the same list before):

- "defsystem": why not use MK defsystem?

- Why not make CLX a CLOS based library by default?

- Why doesn't Apple release the source code of MCL since they do not
  want to port it to the PowerPC?

- and so on...

To those with ready answers that the practices I suggested are not
"business wise" I ask whether they have a better alternative proposal
that would help the lisp community as a whole. Please abstain from
suggesting to move to C/C++ :-)

Thanks to everybody for your patience

Happy Lisping
--
Marco Antoniotti - Resistente Umano
-------------------------------------------------------------------------------
Robotics Lab		| room: 1220 - tel. #: (212) 998 3370
Courant Institute NYU	| e-mail: ·······@cs.nyu.edu

...e` la semplicita` che e` difficile a farsi.
...it is simplicity that is difficult to make.
				Bertholdt Brecht
From: Marcus Daniels
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <MARCUS.94Jun11164734@tdb.ee.pdx.edu>
In-reply-to: ·······@graphics.cs.nyu.edu's message of 11 Jun 1994 19:46:49 GMT

>>>>> "MA" == Marco Antoniotti <·······@graphics.cs.nyu.edu> writes:
In article <·····················@graphics.cs.nyu.edu> ·······@graphics.cs.nyu.edu (Marco Antoniotti) writes:

MA> The objective of such policy is to spread the use of Lisp. The
MA> vendors are now suffering from a shrinking market, not by
MA> competiton among themselves.

!!!

MA> Having said so, I summarize my requests for the vendors and
MA> developers here:  
..
MA> 2) How can I get a LOW priced implementation of CLIM? (e.g. zerop ?).

Heck, how about releasing code to an old implementation?  Although
most universities probably have a recent version of Allegro or Lucid
floating around, which includes CLIM, only a small percentage of
graduate students would have a chance to use it.  It's not like there
is a big LISP curriculum nowadays.

It is easy to explain away LISP's small following to relatively minor
technical flaws (which mostly have been fixed in one form or
the other).  The obvious damning factor seems to be
cost.  How many students (or even professionals) can afford
to shell out $5000 for a development environment they have little
or no experience with?  

If CLIM is all the vendors had to offer then things are would be in a
sad state.  That there are only proprietary implementations of CLIM
greatly reduces my (and I'd guess many others) interest in dealing with 
commercial LISP vendors.  It's not like CLIM is all the vendors had to offer.
Certainly, professional users must be more interested in the vendors'
super-sophisticated compilers, garbage collection,
development-environment niceities, and *good support relationship*,
then they would be in CLIM.

Would X11 be where it is today, if there weren't a shared (public)
code base?

MA> - Why doesn't Apple release the source code of MCL since they do
MA> not want to port it to the PowerPC?

Perhaps so as not to distract from the wonder that is Dylan?
From: Lawrence G. Mayka
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <LGM.94Jun20174634@polaris.ih.att.com>
In article <·····················@graphics.cs.nyu.edu> ·······@graphics.cs.nyu.edu (Marco Antoniotti) writes:

   First of all I think that the vendors should take upon themselves the
   burden to maintain and MAKE AVAILABLE pieces of free software that
   have been floating around for years. CLX is the major one. If CLIM
   wants to be a substitute (as it looks like) they should provide
   support to the independent groups like the CMU folks, the GNU CL
   group (ex AKCL) and Bruno Haible (CLISP) in order to make them able to
   produce a free CLIM version, e.g. by donating the "high" level code.

An alternative that might be more palatable to vendors would be to
define a subset of CLIM, small enough to:

- Bundle into commercial development environments at no extra charge,
so as to whet the developer's desire for full CLIM.

- Port to free-of-charge implementations that can't afford the
development and support costs of full CLIM.

- Run on small machines like PCs and Macs that can't afford the RAM
requirements of full CLIM.
--
        Lawrence G. Mayka
        AT&T Bell Laboratories
        ···@ieain.att.com

Standard disclaimer.
From: Phil Mercurio
Subject: Getting into Garnet (was Re: Is CLX dead?)
Date: 
Message-ID: <2t505m$t4l@riscsm.scripps.edu>
:In-reply-to: ···@trantor.harris-atd.com's message of Tue, 7 Jun 1994 18:38:45 GMT
:
:On the face of it, Garnet seems to be a remarkable piece of work. 
:Odd that there isn't more interest in it.  Has anyone who has developed with it
:comment on this state of affairs?

I came back from CHI '94 all excited about using Garnet here, enough
so that I was determined to relearn Lisp and deal with the problems
of using Lisp in a primarily C++/C environment.  Since then, I've
spent a lot of my off hours trying to get Garnet to run under one
of the freely available Lisps, and this has turned out to be my
biggest stumbling block.  CMUCL doesn't work on some Sparcs (or
maybe it just doesn't like Solaris 1.1.1B) and is unsupported,
while CLISP compiled Garnet fairly easily, but is way, way too slow.
I was able to get everything but Lapidary to compile under AKCL,
and it's faster than CLISP, but nowhere near enough to be useful
(on a Sparc 1+, 24M).

So I'm stuck, the next step in evaluating Garnet would be to try
and get a commercial Lisp in on an evaluation basis.  Garnet looks
like a great piece of work, and it's freely-available, but that
doesn't help if none of the freely-available Lisps are useful.
I can see where a Lisp shop would jump on Garnet, but if you're
not already Lisping, just getting a good look at Garnet is
quite difficult.

---
Phil Mercurio                 The Neurosciences Institute
········@nsi.edu              La Jolla, California

"It's never too late to have a happy childhood."  -- Tom Robbins
From: Fernando Mato Mira
Subject: Re: Getting into Garnet (was Re: Is CLX dead?)
Date: 
Message-ID: <2t57du$c8r@disuns2.epfl.ch>
In article <··········@riscsm.scripps.edu>, ···@nsi.edu (Phil Mercurio) writes:

> I was able to get everything but Lapidary to compile under AKCL,
> and it's faster than CLISP, but nowhere near enough to be useful
> (on a Sparc 1+, 24M).
> 
> So I'm stuck, the next step in evaluating Garnet would be to try
> and get a commercial Lisp in on an evaluation basis.  Garnet looks
> like a great piece of work, and it's freely-available, but that
> doesn't help if none of the freely-available Lisps are useful.
> I can see where a Lisp shop would jump on Garnet, but if you're
> not already Lisping, just getting a good look at Garnet is
> quite difficult.

Yeah. I found out that Allegro CL seems to be much faster than
Harlequin Lispworks when it comes to Garnet. With Lispworks,
the scrollbars lag a lot behind the cursor (hey, ACL does not lag
but my R4000 is putting all its muscle just for that).
Even with ACL, poping up the menu for an option-button takes noticeable time,
even when the menu has already been built.
Note that optimization and declarations are quite pushed to the limit
in the internal Garnet functions, so that setting some global speed/
safety switches does not change things that much.

My best bet to get Garnet working blazing fast (well, almost) would
be to implement the Motif look and feel with calls to the Motif library.
That would take a lot of constraints out of the network, 
(maybe suboptimal) lisp code out of the critical path, memory to check
and reclaim from the garbage collector, and redundant and probably
non-100% look&feel compliant Motif emulation code from your text space.
You could even get lucky and some real Motif code used by other apps 
might end up hanging around in the caches, if you have shared libraries!

GINA is quite fast because of that, but still, it does Motif by
socket communication with a separate process!

--

F.D. Mato Mira                           
Computer Graphics Lab    ········@epfl.ch
EPFL                     FAX: +41 (21) 693-5328

"Wanna be mainstream? Get NATIVE!"
From: Brad Myers
Subject: Re: Getting into Garnet (was Re: Is CLX dead?)
Date: 
Message-ID: <Cr6xBx.CJr.3@cs.cmu.edu>
We made significant attempts to make Garnet run as fast as possible on 
many Lisps.   The "compile-to-C" lisps (AKCL and CLISP) are just inherently
slower than the native lisps.  Most users find Garnet fast enough on modern
machines (SPARC 2 or better, HP 735 or better) with a commercial Lisp 
compiler (e.g., Lucid or Allegro).

> My best bet to get Garnet working blazing fast (well, almost) would
> be to implement the Motif look and feel with calls to the Motif library.

Garnet explicitly is a research project and one goal is to investigate how best
to BUILD widgets.  Therefore, having the widgets implemented in Garnet is
important.  Also, many users find that being able to customize widgets
and make new widgets very useful.

Brad Myers
From: Fernando Mato Mira
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <2t3ueh$iug@disuns2.epfl.ch>
In article <··········@jabba.ess.harris.com>, ···@trantor.harris-atd.com (Mike McDonald) writes:

>   Linking to the C version of Xlib only will work on machines that have C. Not
> all Lisps have access to C. CLX was an attempt to build a lisp equivilant of
> Xlib. Extensions used to work with CLX. Now, one had to actually implement the
> extensions interface inorder to use it. Unfortunately, no one has been
> implementing the standard X extensions.

  Have you tried to implement Xinput? The poorly documented and wizard-lacking
extension definition part of CLX has been programmed with a 
"one wire event = one app event" assumption in mind, which is not the case
with Xinput.

  Who is going to pay in order to have a fully functional, fast, 100% compatible
Xlib written in Lisp (CLX)? People still running lisp machines cannot complain
about the lack of development of CLX more than anything else, and given that
the thing has hibernated for so long means that it's doing pretty well for them,
I guess. If somebody comes with a lisp-based OS again, they are free to add
#+,#- to the sources (OK, maybe the Oberon machines do not have C either, but
then, how many people use that?).

  Any decent commercial lisp on stock hardware has some way of accessing C.
For the people without $$$$, CMU CL and AKCL do too.

  Finally, why not using some of the shared text space already in memory?
Recreating an OS in a Unix process does not favor Lisp appeal in the
marketplace (read also as: use NATIVE threads whenever available!!).

-- 
F.D. Mato Mira                           
Computer Graphics Lab    ········@epfl.ch
EPFL                     FAX: +41 (21) 693-5328
From: Martin Rodgers
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <771109002snz@wildcard.demon.co.uk>
In article <··········@disuns2.epfl.ch>
           ········@di.epfl.ch "Fernando Mato Mira" writes:

> I guess. If somebody comes with a lisp-based OS again, they are free to add
> #+,#- to the sources (OK, maybe the Oberon machines do not have C either, but
> then, how many people use that?).

Most people I know would ask the same thing about Lisp. They usually
ignore me when I answer, "lots of people". What should I tell them?
 
>   Any decent commercial lisp on stock hardware has some way of accessing C.
> For the people without $$$$, CMU CL and AKCL do too.

Which of those will run under DOS or Windows, on my 386?

-- 
Martin Rodgers, WKBBG, London UK   AKA "Cyber Surfer"

Assuming that Clipper will need to be global to be effective, write
to ················@cpsr.org and tell them you oppose Clipper. Now.
This is a shareware .signature  -- please pass it on!
From: Fernando Mato Mira
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <2t6va8$86f@disuns2.epfl.ch>
In article <············@wildcard.demon.co.uk>, ············@wildcard.demon.co.uk (Martin Rodgers) writes:

> > #+,#- to the sources (OK, maybe the Oberon machines do not have C either, but
> > then, how many people use that?).
> 
> Most people I know would ask the same thing about Lisp. They usually
> ignore me when I answer, "lots of people". What should I tell them?

"lots of people" using _lisp_ on an Oberon machine? Please raise your hands!

[What's more, how many fingers do I need to count
the Oberon machines outside of ETHZ?]

-- 
F.D. "Cyber Kid" Mato Mira                           
Computer Graphics Lab    ········@epfl.ch
EPFL                     FAX: +41 (21) 693-5328
From: Martin Rodgers
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <771341196snz@wildcard.demon.co.uk>
In article <··········@disuns2.epfl.ch>
           ········@di.epfl.ch "Fernando Mato Mira" writes:

> > Most people I know would ask the same thing about Lisp. They usually
> > ignore me when I answer, "lots of people". What should I tell them?
> 
> "lots of people" using _lisp_ on an Oberon machine? Please raise your hands!

No, I meant that most people I talk to find the idea of Lisp still being
used very strange. They assume that coz they don't read or hear about it
that it's dead. I'm pleased that it isn't, but what can I offer as proof
that Lisp is still alive?

XLISP? I think not. They can dismiss XLISP as too slow (slower than VB,
and less support for Windows). CLISP is too big (bigger than VB, and no
support for Windows).

> [What's more, how many fingers do I need to count
> the Oberon machines outside of ETHZ?]

I never mentioned Oberon machines. I simply suggested that Lisp has a
similar problem to Oberon, except that a C programmer might not say,
"What's Lisp?", while they might well wonder about Oberon, which gets
almost no coverage at all in the computer magazines a C programmer might
read. I feel like I'm saying "Fortran II" when I say "Lisp".

-- 
Martin Rodgers, WKBBG, London UK   AKA "Cyber Surfer"

Assuming that Clipper will need to be global to be effective, write
to ················@cpsr.org and tell them you oppose Clipper. Now.
This is a shareware .signature  -- please pass it on!
From: Mike McDonald
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <Cr4x6w.Esq@jabba.ess.harris.com>
In article <············@wildcard.demon.co.uk>, ············@wildcard.demon.co.uk (Martin Rodgers) writes:
|> In article <··········@disuns2.epfl.ch>

|> >   Any decent commercial lisp on stock hardware has some way of accessing C.
|> > For the people without $$$$, CMU CL and AKCL do too.

  Unfortunately, the FFI of CMU-CL doesn't allow the C code to call back into
Lisp. Xt, at least, is intimately tied to the notion of callbacks. I guess that's
why everyone is using a C based daemon for their Xt displays.

 (Note: CMU's doc actually says: "There is currently a mechanism for calling Lisp
functions from C, but it is rather restricted, and is scheduled for replacement.
If you need to call Lisp functions from C, contact us and we will let you know
what capabilities are available in the system you have." I'm sending them Email
to find out more.)

|> Which of those will run under DOS or Windows, on my 386?

  Under DOS, I think CLisp is available. Unfortunately, it is missing a few
things like a useful loop macro. No real show stoppers so far though.

  Windows is a problem. I'm currently planning on using ACL/PC which, although not
free, was reasonably priced, especially if you can catch Franz on a good day when
they drop the price by nearly half.

  Mike McDonald				Advanced Technology Dept.	
					Harris Corp.
  Email: ···@trantor.harris-atd.com	M.S. 16-1912
  Voice: (407) 727-5060			P.O. Box 37
  Fax:   (407) 729-3363			Melbourne, Florida 32902
From: Martin Rodgers
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <771341514snz@wildcard.demon.co.uk>
In article <··········@jabba.ess.harris.com>
           ···@trantor.harris-atd.com "Mike McDonald" writes:

>   Unfortunately, the FFI of CMU-CL doesn't allow the C code to call back into
> Lisp. Xt, at least, is intimately tied to the notion of callbacks. I guess
>  that's
> why everyone is using a C based daemon for their Xt displays.

This is acedemic for me, as I have a 386 machine. Unless I'm mistaken,
CMUCL doesn't run on 386 machines. I don't even use Unix.

>   Under DOS, I think CLisp is available. Unfortunately, it is missing a few
> things like a useful loop macro. No real show stoppers so far though.

I have a "useful" loop macro, which I find runs with CLISP very well.
However, I rarely use loop.

>   Windows is a problem. I'm currently planning on using ACL/PC which, although
>  not
> free, was reasonably priced, especially if you can catch Franz on a good day
>  when
> they drop the price by nearly half.

Thanks, but I'm limited to freeware at the moment. Franz have contacted
me about Allegro CL/PC, but that's way too much for me, even at the recent
discount. I've been told by someone who reviewed CL/PC that it should run
well on my machine.

-- 
Martin Rodgers, WKBBG, London UK   AKA "Cyber Surfer"

Assuming that Clipper will need to be global to be effective, write
to ················@cpsr.org and tell them you oppose Clipper. Now.
This is a shareware .signature  -- please pass it on!
From: Martin Rodgers
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <771072895snz@wildcard.demon.co.uk>
In article <··········@disuns2.epfl.ch>
           ········@di.epfl.ch "Fernando Mato Mira" writes:

>   I don't know how the Mac and Win32 work, so let me ask a question:
> Would it be possible to redefine CLX as a low-level platform-independent
> GUI API for taking care of things like getting events, opening windows and
> making requests, but not implementing any high-level concepts (presentations,
> interactors, etc.), or trying to remap generic widgets to the ones
> available on the target system?

It wouldn't be CLX, and who would do it? It would have to be someone
who knows enough about various systems to know what would work, what
kind of abstractions are necessary etc.

A number of people have done this for C and C++, and their libraries
(the commercial ones that get reviewed) always lose a few features of
what ever platforms they support that many programmers will say, "I need
support for that feature, but I don't see it in that library". It could
be the same with a "standard". Any DIY solution, like extending a library,
would be non-standard.

-- 
Martin Rodgers, WKBBG, London UK   AKA "Cyber Surfer"

Assuming that Clipper will need to be global to be effective, write
to ················@cpsr.org and tell them you oppose Clipper. Now.
This is a shareware .signature  -- please pass it on!
From: Kevin Smith
Subject: GL Libraries, Public Domain Lisp implementations
Date: 
Message-ID: <Cr3K27.Hq1@wti.com>
I'm looking for a good public domain lisp implementation for an SGI IRIX
platform that has some library support for GL.  Primary interest is 
educational.  Can anyone recommend something ?  Faqs list many 
compilers/interp choices - not sure which one is best for  graphics.

Thanks
From: Barry Margolin
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <2t7lcbINNkje@early-bird.think.com>
In article <··········@disuns2.epfl.ch> ········@di.epfl.ch (Fernando Mato Mira) writes:
>  I don't know how the Mac and Win32 work, so let me ask a question:
>Would it be possible to redefine CLX as a low-level platform-independent
>GUI API for taking care of things like getting events, opening windows and
>making requests, but not implementing any high-level concepts (presentations,
>interactors, etc.), or trying to remap generic widgets to the ones
>available on the target system?

You can define an API like that, but it wouldn't be CLX.  There's too much
stuff in CLX that's really specific to the X protocol.  What you probably
want is a subset of the work that was done by Xerox, and incorporated as
the GUI interface layer of CLIM.  It would be a subset because Xerox's work
included access to widgets.

However, I happen to think that widget access is an important, almost
critical component.  If you port an application to the Mac or Windows
environment, users are going to expect it to have a Mac-like or
Windows-style user interface.  In some environments users will take
whatever they get, but in the commercial world having a non-standard GUI is
the kiss of death if there's any competition.
-- 
Barry Margolin
System Manager, Thinking Machines Corp.

······@think.com          {uunet,harvard}!think!barmar
From: Mike McDonald
Subject: Re: Is CLX dead?
Date: 
Message-ID: <Cr6rAI.BrM@jabba.ess.harris.com>
In article <············@early-bird.think.com>, ······@think.com (Barry Margolin) writes:
|> In article <··········@disuns2.epfl.ch> ········@di.epfl.ch (Fernando Mato Mira) writes:
|> >  I don't know how the Mac and Win32 work, so let me ask a question:
|> >Would it be possible to redefine CLX as a low-level platform-independent
|> >GUI API for taking care of things like getting events, opening windows and
|> >making requests, but not implementing any high-level concepts (presentations,
|> >interactors, etc.), or trying to remap generic widgets to the ones
|> >available on the target system?
|> 
|> You can define an API like that, but it wouldn't be CLX.  There's too much
|> stuff in CLX that's really specific to the X protocol.

  True, it wouldn't be CLX anymore. Maybe CLX is the "wrong" thing.

|>                                                         What you probably
|> want is a subset of the work that was done by Xerox, and incorporated as
|> the GUI interface layer of CLIM.  It would be a subset because Xerox's work
|> included access to widgets.

  Something along those lines is what I was thinking of.

|> However, I happen to think that widget access is an important, almost
|> critical component.  If you port an application to the Mac or Windows
|> environment, users are going to expect it to have a Mac-like or
|> Windows-style user interface.  In some environments users will take
|> whatever they get, but in the commercial world having a non-standard GUI is
|> the kiss of death if there's any competition.

  That's why I think CLX is the "wrong" thing. It is specific to X, it doesn't
incorporate any of the industry standard GUIs, and it's out of date. My question
to the Lisp communittee is portability across Unix, Windows, and Macs a feature
that people really need? Are you willing to give up some functionality and or
speed for that portability? Or are you willing to spend the effort on rewriting
the GUI portion of your application for each platform? Or are you going to pick
one platform and ignore the rest? Or are you going to just complain that it's all
too hard and do nothing? 

  Mike McDonald				Advanced Technology Dept.	
					Harris Corp.
  Email: ···@trantor.harris-atd.com	M.S. 16-1912
  Voice: (407) 727-5060			P.O. Box 37
  Fax:   (407) 729-3363			Melbourne, Florida 32902
From: Marco Antoniotti
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <MARCOXA.94Jun7110337@butsomi.cs.nyu.edu>
In article <··········@jabba.ess.harris.com> ···@trantor.harris-atd.com (Mike McDonald) writes:


   From: ···@trantor.harris-atd.com (Mike McDonald)
   Sender: ······@jabba.ess.harris.com (Usenet News Feed)
   Nntp-Posting-Host: v3a.ess.harris.com
   Organization: Advanced Technology Dept, Harris Corp, Melbourne, FL

   |> Today, is it realistic or even desirable to think that LISP
      users have LISP 
   |> interface toolkits?  Is it even worth putting effort into CLX?

This is a transitive argument given the nature of the beast. CLX is
the equivalent of XLIB, i.e. it is an implementation of the X
protocol, which is, by definition, language independent. If we stretch
this argument we'd end up stating that no effort is worthwhile to add
a X protocol implementation to any language.

  Realistic? No. Desirable? Yes. Worth the effort? Probably not.

Unfortunately I tend to agree with this statement. But I agree even
more with the following one.

  I don't believe that the vendors would support the effort.

This has really always puzzled me. Why shouldn't they? What keeps them
from understanding that they will have more market shares if the
language is more in use?

Besides, do you really think that all the proprietary effort on CLIM
is really doing away with CLX? I really think the opposite.

  And I don't think there are enough
  independant lisp groups to support the effort either. If CMU could
  have stayed 
  with Lisp instead of having to jump ship for Dylan, there might
  still have been a 
  chance.

This still herts! :(

   |> Yet, other than Fresco, NEXTSTEP, and perhaps Win32 (Chicago?),
   |> there doesn't seem to be much potential for aggreement in the OOP GUI
   |> world.

You are leaving Motif out. Which is, once again on a totally different
level than CLX.

   |> 
   |> For all the neglect this area has supposedly received recently by the
   |> LISP community, I'm somewhat underwhelmed by the `modern' GUI
      alternatives 
   |> (commercial or otherwise).
   |> 

     I am too. But then I haven't seen anything much better in lisp.

I really believe that the "Symbolics/Interlisp" syndrome has done a
great deal of harm to the CL GUI world. The overall dismissal of
CLUE/CLIO in their early stages is a sign of this attitude. (Yes, I know
that Harlequin developed LispWorks on CLUE)

Happy Lisping

--
Marco Antoniotti - Resistente Umano
-------------------------------------------------------------------------------
Robotics Lab		| room: 1220 - tel. #: (212) 998 3370
Courant Institute NYU	| e-mail: ·······@cs.nyu.edu

...e` la semplicita` che e` difficile a farsi.
...it is simplicity that is difficult to make.
				Bertholdt Brecht
From: Mike McDonald
Subject: Re: Is CLX dead? (Re: C machines (was: copy-file))
Date: 
Message-ID: <Cr2zGr.EyG@jabba.ess.harris.com>
In article <····················@butsomi.cs.nyu.edu>, ·······@butsomi.cs.nyu.edu (Marco Antoniotti) writes:
|> In article <··········@jabba.ess.harris.com> ···@trantor.harris-atd.com (Mike McDonald) writes:

|> Unfortunately I tend to agree with this statement. But I agree even
|> more with the following one.
|> 
|>   I don't believe that the vendors would support the effort.
|> 
|> This has really always puzzled me. Why shouldn't they? What keeps them
|> from understanding that they will have more market shares if the
|> language is more in use?
|> 
|> Besides, do you really think that all the proprietary effort on CLIM
|> is really doing away with CLX? I really think the opposite.

  The vendors are placing all of their hopes and efforts into CLIM. They support
CLX only enough to statisfy CLIM's needs. Heck, I don't know if CLIM even uses
CLX anymore. With CLIM's binding to native C widget sets, I wouldn't be surprised
if it didn't.
 
|>   And I don't think there are enough
|>   independant lisp groups to support the effort either. If CMU could
|>   have stayed 
|>   with Lisp instead of having to jump ship for Dylan, there might
|>   still have been a 
|>   chance.
|> 
|> This still herts! :(
|> 
|>    |> Yet, other than Fresco, NEXTSTEP, and perhaps Win32 (Chicago?),
|>    |> there doesn't seem to be much potential for aggreement in the OOP GUI
|>    |> world.
|> 
|> You are leaving Motif out. Which is, once again on a totally different
|> level than CLX.

  I agree that XLib (CLX) and Motif are on different levels. But for a usable
GUI, you really need both. (It'd be nice if they worked well together too.) Even
in the Motif world, you still XLib for lots of things. (Graphics being the most
obvious.) The question is how does a very small group gain access to the large
amount of existing but hostile graphical libraries?

|>    |> 
|>    |> For all the neglect this area has supposedly received recently by the
|>    |> LISP community, I'm somewhat underwhelmed by the `modern' GUI
|>       alternatives 
|>    |> (commercial or otherwise).
|>    |> 
|> 
|>      I am too. But then I haven't seen anything much better in lisp.
|> 
|> I really believe that the "Symbolics/Interlisp" syndrome has done a
|> great deal of harm to the CL GUI world. The overall dismissal of
|> CLUE/CLIO in their early stages is a sign of this attitude. (Yes, I know
|> that Harlequin developed LispWorks on CLUE)

  I could write tons on why I think Lisp is dead but that's another ball of wax.
In some ways, I agree with you. In others, I disagree. Yes, I think that the
preoccupation with presentation types ala dynamic windows has been a detriment to
the lisp communitee. I also think some of Symbolics' totally brain dead business
decisions have really hurt lisp. I also think that Common Lisp (and the recent
ANSI Lisp standard) have helped done more than their fair share to the detriment
of lisp. (ie. the kitchen sink syndrome for starters.) But I'm wandering ...

  Mike McDonald				Advanced Technology Dept.	
					Harris Corp.
  Email: ···@trantor.harris-atd.com	M.S. 16-1912
  Voice: (407) 727-5060			P.O. Box 37
  Fax:   (407) 729-3363			Melbourne, Florida 32902