In 1997, the language ISLISP was standardized (ISO/IEC 13816:1997(E)
ISLISP Language) to "bridge the gap between the various incompatible
members of the Lisp family of languages (most notably Common Lisp,
Eulisp, LeLisp, and Scheme) by focusing on standardizing those areas
of widespread agreement." <http://islisp.info/index.html> Since that
time, Common Lisp and Scheme have gone on into the real world, whereas
ISLISP seems to have faded into obscurity.
From my own very brief experiences playing around with OpenLisp by
Eligis <http://christian.jullien.free.fr/> ISLISP feels like a small
Common Lisp subset with some gratuitious differances (ISLISP FOR is
equivalent to Common Lisp DO) and no advantages. ISLISP is nothing you
couldn't do with a few bored hours, a reasonable knowledge of macros,
and a Common Lisp implementation. If ISLISP had taken call/cc or even
hygenic macros it would at least have been interesting. As it stands,
it just feels pointless, especially given that it postdates both
Common Lisp and Scheme.
So, is anyone using ISLISP these days? Any interesting
applications/extensions of the language? Anything to make it worth
taking a more in-depth look at?
Chris Barts <··············@gmail.com> writes:
> In 1997, the language ISLISP was standardized (ISO/IEC 13816:1997(E)
> ISLISP Language) to "bridge the gap between the various incompatible
> members of the Lisp family of languages (most notably Common Lisp,
> Eulisp, LeLisp, and Scheme) by focusing on standardizing those areas
> of widespread agreement." <http://islisp.info/index.html> Since that
> time, Common Lisp and Scheme have gone on into the real world, whereas
> ISLISP seems to have faded into obscurity.
>
> From my own very brief experiences playing around with OpenLisp by
> Eligis <http://christian.jullien.free.fr/> ISLISP feels like a small
> Common Lisp subset with some gratuitious differances (ISLISP FOR is
> equivalent to Common Lisp DO) and no advantages. ISLISP is nothing you
> couldn't do with a few bored hours, a reasonable knowledge of macros,
> and a Common Lisp implementation. If ISLISP had taken call/cc or even
> hygenic macros it would at least have been interesting. As it stands,
> it just feels pointless, especially given that it postdates both
> Common Lisp and Scheme.
>
> So, is anyone using ISLISP these days? Any interesting
> applications/extensions of the language? Anything to make it worth
> taking a more in-depth look at?
Well given that Common Lisp displaced all the other lisps (but emacs
lisp and autocad lisp), the need for an ISO standard lisp is greatly
reduced.
Another thing that has changed in the situations when ISLISP was
created and now, is that twenty years ago, free software was scarcer,
and systems were more diverse. When you had an OS, you didn't have a
choice between N native implementations of lisp or scheme, and M
others you could fetch from the internet and port or just compile on
your machine. Basically, you used what the system provided gave you,
or what you paid very dearly to some rare commercial vendor. In this
situation, it was worthwhile to have an ISO Lisp or Common Lisp layer
in the implementation you had to use, since it was the only way to be
able to port easily your own development from a system to the other.
Nowadays, if you develop an application in clisp on Linux and suddenly
you want to deploy it on MS-Windows, you don't care about portability
over lisp, you just fetch clisp sources, compile them on MS-Windows,
and launch your lisp application on it. You would never try to port
your lisp code to this LeLisp or that MacLisp, they just don't exist
anymore, so you don't need an ISO Lisp layer on either. At most, you
may want to port your Common Lisp code from clisp to sbcl, or to
allegro cl, but this only means to use some portability libraries.
And it's even worse than that, if your clisp application uses some C
libraries accessing some low level linux stuff, when porting to
MS-Windows, you could very well actually put it in a user-mode-linux
virtual machine and be done without changing a token of your sources.
--
__Pascal Bourguignon__
···@informatimago.com (Pascal J. Bourguignon) writes:
> Chris Barts <··············@gmail.com> writes:
>
>> In 1997, the language ISLISP was standardized (ISO/IEC 13816:1997(E)
>> ISLISP Language) to "bridge the gap between the various incompatible
>> members of the Lisp family of languages (most notably Common Lisp,
>> Eulisp, LeLisp, and Scheme) by focusing on standardizing those areas
>> of widespread agreement." <http://islisp.info/index.html> Since that
>> time, Common Lisp and Scheme have gone on into the real world, whereas
>> ISLISP seems to have faded into obscurity.
>>
>> From my own very brief experiences playing around with OpenLisp by
>> Eligis <http://christian.jullien.free.fr/> ISLISP feels like a small
>> Common Lisp subset with some gratuitious differances (ISLISP FOR is
>> equivalent to Common Lisp DO) and no advantages. ISLISP is nothing you
>> couldn't do with a few bored hours, a reasonable knowledge of macros,
>> and a Common Lisp implementation. If ISLISP had taken call/cc or even
>> hygenic macros it would at least have been interesting. As it stands,
>> it just feels pointless, especially given that it postdates both
>> Common Lisp and Scheme.
>>
>> So, is anyone using ISLISP these days? Any interesting
>> applications/extensions of the language? Anything to make it worth
>> taking a more in-depth look at?
>
> Well given that Common Lisp displaced all the other lisps (but emacs
> lisp and autocad lisp), the need for an ISO standard lisp is greatly
> reduced.
>
Especially given that it's possible for a language to be
cross-standardized. (C, for example, has both ANSI and ISO standards,
with the ANSI coming first in 1989 and ISO copying that standard
(either word-for-word or close enough).) Not that this would stop a
full-on political fight, I suppose.
> Another thing that has changed in the situations when ISLISP was
> created and now, is that twenty years ago, free software was scarcer,
> and systems were more diverse.
Hm. I was under the impression that ISLISP was about a decade old, not
two decades. ("ISO/IEC 13816:1997(E) ISLISP Language" suggests that,
unless I'm badly misreading or missing something.)
> When you had an OS, you didn't have a
> choice between N native implementations of lisp or scheme, and M
> others you could fetch from the internet and port or just compile on
> your machine. Basically, you used what the system provided gave you,
> or what you paid very dearly to some rare commercial vendor. In this
> situation, it was worthwhile to have an ISO Lisp or Common Lisp layer
> in the implementation you had to use, since it was the only way to be
> able to port easily your own development from a system to the other.
This all makes sense. I was just expecting ISLISP to be more different
from Common Lisp given that the quote (above) implies that it was
intended to be used to help bridge the gap between Common Lisp and
Scheme (in addition to LeLisp and Eulisp). It's no easier to 'bridge'
from ISLISP to Scheme than it is to go from Common Lisp to Scheme,
given how similar ISLISP and Common Lisp are. (I suppose they
misjudged how quickly Common Lisp would become the only standalone
Lisp-2 dialect worth talking about. But, as per above, wasn't this
fairly obvious by 1997?)
>
> Nowadays, if you develop an application in clisp on Linux and suddenly
> you want to deploy it on MS-Windows, you don't care about portability
> over lisp, you just fetch clisp sources, compile them on MS-Windows,
> and launch your lisp application on it. You would never try to port
> your lisp code to this LeLisp or that MacLisp, they just don't exist
> anymore, so you don't need an ISO Lisp layer on either. At most, you
> may want to port your Common Lisp code from clisp to sbcl, or to
> allegro cl, but this only means to use some portability libraries.
Right. This is even worse/better when you start developing in PHP,
Perl, Ruby, Python, and so on, where the language is defined by a
tight-knit family of implementations that all morph over time. Perl
especially seems to want to provide a Unix-like experience regardless
of the OS, providing C-style POSIX-lite APIs as part of the
interpreter.
> And it's even worse than that, if your clisp application uses some C
> libraries accessing some low level linux stuff, when porting to
> MS-Windows, you could very well actually put it in a user-mode-linux
> virtual machine and be done without changing a token of your sources.
To be fair, Eunice[1] was pretty gross, too, and nobody thinks of porting
to MVS[2] without a compatibility layer of some kind.
[1] A Unix compatibility layer for VMS. Think Cygwin and you're about
there.
[2] The mainframe OS IBM calls z/OS now, and z/OS includes Unix
compatibility out of the box.
Chris Barts wrote:
> This all makes sense. I was just expecting ISLISP to be more different
> from Common Lisp given that the quote (above) implies that it was
> intended to be used to help bridge the gap between Common Lisp and
> Scheme (in addition to LeLisp and Eulisp).
A standard's goals seldom coincide with its achievements.
> (I suppose they
> misjudged how quickly Common Lisp would become the only standalone
> Lisp-2 dialect worth talking about. But, as per above, wasn't this
> fairly obvious by 1997?)
The ISLISP standard has its roots in the situation circa
1987. It took about ten years to go from idea to ISO/IEC
standard.
Will
Chris Barts <··············@gmail.com> writes:
> ···@informatimago.com (Pascal J. Bourguignon) writes:
>
>> Chris Barts <··············@gmail.com> writes:
>>
>>> So, is anyone using ISLISP these days? Any interesting
>>> applications/extensions of the language? Anything to make it worth
>>> taking a more in-depth look at?
>>
>> Well given that Common Lisp displaced all the other lisps (but emacs
>> lisp and autocad lisp), the need for an ISO standard lisp is greatly
>> reduced.
>>
>
> Especially given that it's possible for a language to be
> cross-standardized. (C, for example, has both ANSI and ISO standards,
> with the ANSI coming first in 1989 and ISO copying that standard
> (either word-for-word or close enough).) Not that this would stop a
> full-on political fight, I suppose.
>
>> Another thing that has changed in the situations when ISLISP was
>> created and now, is that twenty years ago, free software was scarcer,
>> and systems were more diverse.
>
> Hm. I was under the impression that ISLISP was about a decade old, not
> two decades. ("ISO/IEC 13816:1997(E) ISLISP Language" suggests that,
> unless I'm badly misreading or missing something.)
Yes, but it takes time to finalize a standard document, and also
before that it takes time for people to realize the situation has
changed.
>> When you had an OS, you didn't have a
>> choice between N native implementations of lisp or scheme, and M
>> others you could fetch from the internet and port or just compile on
>> your machine. Basically, you used what the system provided gave you,
>> or what you paid very dearly to some rare commercial vendor. In this
>> situation, it was worthwhile to have an ISO Lisp or Common Lisp layer
>> in the implementation you had to use, since it was the only way to be
>> able to port easily your own development from a system to the other.
>
> This all makes sense. I was just expecting ISLISP to be more different
> from Common Lisp given that the quote (above) implies that it was
> intended to be used to help bridge the gap between Common Lisp and
> Scheme (in addition to LeLisp and Eulisp). It's no easier to 'bridge'
> from ISLISP to Scheme than it is to go from Common Lisp to Scheme,
> given how similar ISLISP and Common Lisp are. (I suppose they
> misjudged how quickly Common Lisp would become the only standalone
> Lisp-2 dialect worth talking about. But, as per above, wasn't this
> fairly obvious by 1997?)
I think that the help in bridging that gap is mostly related to dynamic
variables, that in ISLISP are clearly denoted with the dynamic special
operator.
--
__Pascal Bourguignon__