Differences between Interlisp and Common Lisp

[Lassi Kortela]

Has anyone thought of publishing a guide to Interlisp for people who
already know Common Lisp? A bit like a reverse "what's new" document.

[Alexander Shendi]

Hi,

A bit like "What You Get By Downgrading from ANSI Common Lisp to INTERLISP-D"?
- D-EDIT/S-EDIT a wholesome improvement over emacs or any othe "IDE", for that matter
- Image Based Lisp Programming
- #Traditonal Dynamic Scoping, instead of boring new Lexical Scoping
- Case Sensitivity
- A Whole New & Different Representation of Path Names

(For the humour impaired: Humour/Satire -- The Surgeon General Warns: Abuse of Laughter may lead stiches, shortage of breath and thinkcrime)

NB: I don't actually know enough about Interlisp or Lisp History to write such a document.

Am 16. September 2021 11:24:26 MESZ schrieb Lassi Kortela <la...@lassi.io>:
>Has anyone thought of publishing a guide to Interlisp for people who
>already know Common Lisp? A bit like a reverse "what's new" document.
>

--
Diese Nachricht wurde von meinem Android-Gerät mit K-9 Mail gesendet.

[Tim Bradshaw]

It's worth remembering that you can use DEdit for CL almost always. I really only ever wrote CL on D-machines: I only used Interlisp when I really had to as I didn't know it at all well. There are some limitations, one at least used to be #+/#-, but for many purposes it's fine. And it is some kind of object lesson in how good a structure editor could be, I think.

--tim

> On 16 Sep 2021, at 11:16, Alexander Shendi <Alexande...@web.de> wrote:
>
> Hi,

> --
> https://Interlisp.org for more details
> ---
> You received this message because you are subscribed to the Google Groups "Medley Interlisp" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to interlisp+...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/interlisp/5D788BC8-BBE7-41A9-8C00-F12AAEC3739A%40web.de.

[Pascal Bourguignon]

Le 16/09/2021 à 12:16, Alexander Shendi a écrit :
> Hi,
>
> A bit like "What You Get By Downgrading from ANSI Common Lisp to
INTERLISP-D"?
> - D-EDIT/S-EDIT a wholesome improvement over emacs or any othe "IDE",
for that matter
> - Image Based Lisp Programming
> - #Traditonal Dynamic Scoping, instead of boring new Lexical Scoping
> - Case Sensitivity
> - A Whole New & Different Representation of Path Names
>
> (For the humour impaired: Humour/Satire -- The Surgeon General Warns:
Abuse of Laughter may lead stiches, shortage of breath and thinkcrime)
>
> NB: I don't actually know enough about Interlisp or Lisp History to
write such a document.

Already, all those features but the loss of lexical scoping make us
salivate!


http://informatimago.com/develop/lisp/com/informatimago/small-cl-pgms/sedit/index.html

http://informatimago.com/develop/lisp/com/informatimago/small-cl-pgms/ibcl/

informatimago says: (SETF (READTABLE-CASE *READTABLE*) :PRESERVE)

informatimago says: long live to logical pathnames!

I've got a serious case of InterLisp envy!

--
__Pascal Bourguignon__

[Alexander Shendi]

Hi Tim,

How do you do it? "(IL:DF 'my-nice-fn)"? Do you have to call the function as "(IL:my-nice-fn ...)" or is it automagically imported into CL-USER?

TIA, HAND,

Alexander

[Alexander Shendi]

Hi Pascal,

But few people seem to want to use it. Certain Editor Macros or Electron apps are more popular. I guess full screen editors have won! Vae Victis!

[Nick Briggs]

If you're running the full.sysout, you'll have the "who-line" window up, and you can see which readtable and which package you'll get by default for the process with the input focus, and you can change them. If you're typing at an Interlisp exec instead of an XCL exec (new ones of either flavor from the right-button background menu) you'll default to the IL package.

> To view this discussion on the web visit https://groups.google.com/d/msgid/interlisp/8616BE4E-A635-4D37-A841-D7D88EFFD12D%40web.de.

[Larry Masinter]

> Has anyone thought of publishing a guide to Interlisp for people who already
> know Common Lisp? A bit like a reverse "what's new" document.

I'm hoping we can provide enough raw material to aid software historians who want to tell the stories of evolution of computer language and systems.
Yes, there are papers and discussions you can read https://ml.cddddr.org/lisp@parc/maillist.html and
https://ml.cddddr.org/cl-cleanup/maillist.htmlbut, but what were we talking about then? Now you can try it first-hand.

These days, it seems like the way to explain things would be youtube videos.
https://github.com/Interlisp/medley/issues/4
that would be useful for people who know other programming languages and systems, not just Lisps.
--
https://LarryMasinter.net https://interlisp.org

[Alexander Shendi]

Hi Nick,

Thanks for your reply. I was referring to writing CL programs with Interlisp DF/DEDIT, so I guess that means calling them from an XCL EXEC.

[Nick Briggs]

If you're in an XCL Exec, the default (which you can see from the who-line display) would be that unqualified atoms are created in the XCL-USER package, so (IL:DF my-nice-fn) will create XCL-USER::MY-NICE-FN.

Interlisp knows about CommonLisp packages, so in an Interlisp exec you could use XCL-USER::MY-NICE-FN, and throughout the entire system you can freely intermix the two languages.

[Tim Bradshaw]

Sadly I no longer remember: this is all thirty years ago now. I suspect strongly that either ED (ie CL:ED) just worked or that I trained it to the right thing. It's also the case that I never used a 'raw' sysout (I think they existed on the fileservers, but Dandytigers only really had enough storage for one sysout) but rather some inherited one with a load of now-useless speech-related code in it so it may be that someone previously had hacked it to work. I need to find time and organisation to get a Medley system working again...

--tim

[Alexander Shendi]

Hi,

thanks! One more silly question: What "level" of Common Lisp is supported, i.e. CLtL1, CLtL2, ANSI (I suppose not)?

[John Cowan]

CLtL1 and parts of CLtL2 are supported: in particular, there is a CLOS implementation.  Getting it up to ANSI standard is I think very important for the future of Medley, but so far no one has the time/energy to work on it.  Volunteering would be a Good Thing.

The right place to start, I think, is with the ANSI CL test suite at <https://gitlab.common-lisp.net/ansi-test>: see how much passes, fails, or errors out.  Some code probably needs to be commented out because READ is not fully ANSI yet.  There is lots of code in other Common Lisps that can be brought in, especially from SBCL and SICL (both BSD licensed).

To view this discussion on the web visit https://groups.google.com/d/msgid/interlisp/5BFD1EFD-090B-4643-BB08-BA53C05BBF3E%40web.de.

[Tim Bradshaw]

>
> On 17 Sep 2021, at 16:04, John Cowan <co...@ccil.org> wrote:
>
> CLtL1 and parts of CLtL2 are supported: in particular, there is a CLOS implementation.

What's the history of the CLOS implementation? When I used (an earlier) Medley on D-machines I spent a great deal of time building PCL releases (or, the machine did, I think it was straightforward but slow), so I'd be interested in knowing if it derives from PCL or is of independent origin. That whole history must be interesting in fact, since PCL came from ?Common LOOPS? I think which came from LOOPS which was Interlisp (again, I think). I wonder if any of the PCL release history survives? They had good names: I remember rainy day, Victoria day, and probably others if I think.

--tim

[Alexander Shendi]

Forwarding to list.


-------- Ursprüngliche Nachricht --------
Von: John Cowan <co...@ccil.org>
Gesendet: 17. September 2021 19:04:18 MESZ
An: Alexander Shendi <Alexande...@web.de>
Betreff: Re: Differences between Interlisp and Common Lisp

On Fri, Sep 17, 2021 at 12:55 PM Alexander Shendi <Alexande...@web.de>
wrote:

Dear list,
>

You sent this just to me, but no matter.

> What are are the goals of the "Medley Restoration Project"?
> 1. An environment to run historic SYSOUTs on modern machines and OSes
> (conserve Interlisp/XCL)
>

Definitely.

> OR
>

AND

> 2. Yet another ANSI CL implementation
>

That too, but not just "yet another"; rather, a CL embedded in the Medley
environment, which allows a "develop CL code in Medley, deploy it in SBCL"
mode. For that purpose, speed doesn't matter so much, any more than it
does for Elisp.

[Larry Masinter]

> What's the history of the CLOS implementation? 

 

https://github.com/Interlisp/medley/blob/master/clos/README.MD

 

This is based on the '91 PCL. In order to load this into Medley 3.5, load t the file DEFSYS.DFASL, and then execute (CLOS::LOAD-CLOS). After this all the CLOS functionality is in the package CLOS.

 

 

>  I spent a great deal of time building PCL releases (or,

> the machine did, I think it was straightforward but slow

 

[Tim Bradshaw]

On 18 Sep 2021, at 02:40, Larry Masinter <L...@acm.org> wrote:
<very short load times for CLOS>

Is that to load it or build it from source?  Either way things are clearly a lot faster than they were in 1990...

--tim