Hello,
Could someone please point out the main differences between
the ISO/IEC 13816:1997 ISLISP standard and the older
ANSI X3.226-1994 Common Lisp standard. Does the ISLISP
standard obsolete Common Lisp in any way?
Thanks,
Neil
Neil Zanella <········@garfield.cs.mun.ca> writes:
> Could someone please point out the main differences between
> the ISO/IEC 13816:1997 ISLISP standard and the older
> ANSI X3.226-1994 Common Lisp standard.
The ISLISP dialect is a tiny dialect created for various political reasons.
It's the approximate conceptual size of Scheme but is conceptually in the
style of Common Lisp.
Loosely characterized, it's about the same as if you went through ANSI
CL and then removed all but 3 of the functions in each chapter, and
then renamed about 1/3 of the functions that remained. :-) It is a VERY
spartan dialect, a bit over 100 pages; you should be able to easily read
the spec in an evening or two if you already know some other Lisp dialect.
ISLISP is of very little commercial value as a Lisp unto itself, but it
would serve as an appropriate base for extension if someone were looking
to make a smallish Lisp that conformed to some standard much smaller than
ANSI CL (e.g., perhaps for embedded applications).
I and others worked in the design of ISLISP to keep it "conceptually
compatible" with CL to the extent that it is believed that one could host
an ISLISP implementation in the same address space as CL without creating
logical contradictions, although it does have quite a different naming
scheme for a number of concepts (particularly its way of arranging arrays
in its type matrix).
> Does the ISLISP standard obsolete Common Lisp in any way?
No. It is (deliberately) for a different LISP dialect.
ISLISP does not threaten ANSI CL. There are a small handful of
implementations, but they do not compete commercially with ANSI CL in
any material sense.
In my opinion, ISLISP was important for various political reasons to
keep various Lisp efforts already in progress from being labeled as
non-standard. As they say, "the nice thing about standards is that
there are so many of them".
The long term impact of ISLISP has yet to be shown, but I think at worst
it can be regarded as harmless, and at its best it helps to raise the
awareness of Lisp internationally and creates one more arena for possible
future work.
Also, as a side-effect, I think the exercise of producing the standard
was also important to calming certain political tensions that I
perceive as having existed between various countries' Lisp efforts at
the outset. I think at a personal and committee level, it brought a
much greater understanding to those of us involved in the project of
the issues and interests of other nations, and it brought a certain
rightful degree of pride to the various nations that contributed to its
design.
Disclaimer: I used to be (but no longer am) the US international
representative for (X3)J13 to the committee SC22/WG16 that designed
ISLISP. I continue to act as project editor for ISLISP. However,
none of my remarks in the above should be seen as official positions
of ANSI, J13, ISO, SC22, nor WG16. They're just my personal impressions
based on having been close to the process.
Kent M Pitman <······@world.std.com> writes:
> I and others worked in the design of ISLISP to keep it "conceptually
> compatible" with CL to the extent that it is believed that one could host
> an ISLISP implementation in the same address space as CL without creating
> logical contradictions, although it does have quite a different naming
> scheme for a number of concepts (particularly its way of arranging arrays
> in its type matrix).
I think you're right. I once did part of the work of implementing an
ISLISP package in CL and there were no major stumble blocks.
> The long term impact of ISLISP has yet to be shown, but I think at worst
> it can be regarded as harmless, and at its best it helps to raise the
> awareness of Lisp internationally and creates one more arena for possible
> future work.
One thing I find strangely missing, given the emphasis on industrial
needs in the introduction, is a module/package system. Otherwise, it's
a nice little language that I find more appropriate for teaching
purposes than Scheme.
--
Lieven Marchand <···@wyrd.be>
She says, "Honey, you're a Bastard of great proportion."
He says, "Darling, I plead guilty to that sin."
Cowboy Junkies -- A few simple words
From: T. Kurt Bond
Subject: Re: standards: ANSI LISP vs. ISO/IEC ISLISP
Date:
Message-ID: <m3elp33huv.fsf@tkb.mpl.com>
Lieven Marchand <···@wyrd.be>, writing about ISLISP, said:
> One thing I find strangely missing, given the emphasis on industrial
> needs in the introduction, is a module/package system. Otherwise,
> it's a nice little language that I find more appropriate for
> teaching purposes than Scheme.
EuLISP is another small Lisp that originated in Europe. EuLisp
decides a lot of design design decisions differently than ISLISP or
Common Lisp. It has a module system and threads. Unlike ISLISP its
definition was never completely finished, but it makes an interesting
read, and there is at least one reasonably complete implementation
(Youtoo) and at least one subset implementation (Euscheme)
--
T. Kurt Bond, ···@tkb.mpl.com