Seeking IPL-V on PDP-10 (or, for that matter, any of the retro systems)

[Jeff Shrager]

Hi. I'm actually a PiDP-8 owner, but in my "day job" I work on reanimating old AIs. (Okay, that's not really my day job -- wishful thinking -- maybe my primary hobby? Anyway... ) A while back we reanimated the original ELIZA:

    https://ieeexplore.ieee.org/document/11030922

And more recently I succeeded in reanimating the Logic Theorist (LT), literally the world's first AI:

   https://arxiv.org/abs/2603.13514

(Yes, we can argue about Samuel's check player if you like.)

This is pretty good as far as it does, but LT was written in a now-obscure language called IPL-V (IPL five), and the LT project required that I build my own IPL-V interpreter, which was fun and all, but I'm now finding more and more IPL-V code and don't really want to hassle with my emulator all the time, so I'd really like to find a real IPL-V interpreter from those days. 

I know that IPL-V ran on Dec-10s bcs the SAIL documentation mentions it! I don't have a PiDP-10, but I'm hoping that one of you who does could poke around and see whether the IPL-V is on the release you have, and if so, maybe you'd like to collaborate on reanimating some more of the world's first AIs and cognitive models.

Thanks in advance for any effort you put into looking into this.

Cheers,

'Jeff

[Clem Cole]

Have you looked online at the SAILDART archives?

--
You received this message because you are subscribed to the Google Groups "PiDP-10" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pidp-10+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/pidp-10/e5cb0cd4-68ab-45cb-a9a2-1617752a5aa3n%40googlegroups.com.

[Jeff Shrager]

Thanks. I have. There are a  couple of mentions. I also found several interesting documents regarding how to run IPL-V at Stanford (attached) but these are the only clues I've found. IPL-V would not have been in use after about 1965, so it's unlikely that it ended up on the DEC10 release. 

----------

Jeff Shrager, Pi Rebel

Chief Scientist, Bennu Climate, Inc.

[Lars Brinkhoff]

Jeff wrote:

I work on reanimating old AIs.

Have you looked into ELIZA's friend PARRY?

[Jeff Shrager]

I have, actually. I have all the code (which I got from SAIL DART!) It
is written in MLisp -- a Meta-Lisp written (I think) at SAIL -- and
LAP, a lisp assembler for the Dec10. It would not be difficult to
translate MLisp to CL, and I don't think that the LAP does a lot that
I couldn't replace with CL code. In fact, I was musing recently about
just handing the who task to Claude. But that would be extremely
in-authentic. I'm guessing that since I got it all from SAIL DART, one
of your systems can actually run it directly.

Cheers,
'Jeff

> --
> You received this message because you are subscribed to a topic in the Google Groups "PiDP-10" group.
> To unsubscribe from this topic, visit https://groups.google.com/d/topic/pidp-10/vTSghb_v74Q/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to pidp-10+u...@googlegroups.com.
> To view this discussion visit https://groups.google.com/d/msgid/pidp-10/f4c0c91f-c4bb-4805-9d7e-eecc7cd4bf1cn%40googlegroups.com.



--

[Lars Brinkhoff]

Jeff Shrager wrote:

Lars Brinkhoff wrote:
> Have you looked into ELIZA's friend PARRY?

I have, actually. I have all the code (which I got from SAIL DART!) It
is written in MLisp -- a Meta-Lisp written (I think) at SAIL -- and
LAP, a lisp assembler for the Dec10. It would not be difficult to
translate MLisp to CL, and I don't think that the LAP does a lot that
I couldn't replace with CL code. In fact, I was musing recently about
just handing the who task to Claude. But that would be extremely
in-authentic. I'm guessing that since I got it all from SAIL DART, one
of your systems can actually run it directly.

I would hope so.  I tested logging into the SAIL host, and running MLISP.

It seems to work.  There is also a PARRY executable, but it complains about

a missing file QPARRY[PAR,BLF].  That file is in the Saildart archive, so it

should be salvageable.  

 

[Jeff Shrager]

It’s all under here somewhere:

https://drive.google.com/drive/folders/12A9lXzcrdUsxT0oO1iUEuJ855Fd8tAZT

----------

Jeff Shrager, Pi Rebel

Chief Scientist, Bennu Climate, Inc.

[Jeff Shrager]

Hmmm. QPARRY is a bit of a mystery. There's no file in the PARRY (at least that I've come across so far) by that name, but there's special purpose SUMMEX-specific code in pmem5 that mentions it, which is presumably what you ran into:

EXPR SUMEXALLOW();
	BEGIN NEW FILCHAN,FLAG,STATUS;
	FILCHAN EVAL<'INPUT,DIAFILEAREA,'QPARRY>;
	INC(FILCHAN,NIL);
	STATUS READ(); FLAG READ();
	IF STATUS='OK THEN INC(NIL,T) ALSO RETURN FLAG;
	INC(NIL,NIL);
	EVAL < 'OUTPUT, DIAFILEAREA, 'QPARRY >;
	OUTC(T,NIL); PRINT(0); PRINT FLAG; OUTC(NIL,T);   % RESET FILE%
	IF STATUS=0 THEN SLEEP(10) ALSO EXIT() ALSO CAR NIL;
	RETURN STATUS;
	END;

There are numerous other <letter>PARRY files; maybe as a hail mary copy on of them into QPARRY.<whatever> and see if it does anything sensible, under the theory that some execution script copies in one of them to QPARRY ebfore running? This seems unlikely since these are all different.

Google also can't find anything called QPARRY that seems related. (There's a collision detection utility called "PARRY", which suddenly gave me a new, and very apt, reading of the name "PARRY"! :-)

[Lars Brinkhoff]

Jeff Shrager wrote:

Hmmm. QPARRY is a bit of a mystery.

Here's an example.  I put it on the SAIL file system, and PARRY seemed happy about

that.  There were also two more files necessary.  I updated this:

https://github.com/larsbrinkhoff/sailing-on-arpanet/

 

         NIL      THIS IS THE MESSAGE OF THE DAY
                 IT IS EITHER NIL, OR ONE SEXPR  (LIKE THIS)

        1
                 IS NUMBER OF NET PARRY USERS ALLOWED IN NON PRIME TIME
        ALL
                 NIL MEANS NO NET PARRY USERS DURING PRIME TIME
                 T MEANS EXCLUDE THE FOLLOWING LIST
                 ALL MEANS EXCLUDE AND LET THEM SEE THE INTERNAL WORKINGS

        NIL      NIL MEANS NO 100,100 USERS EVEN IF NOT ON LIST BELOW
                 T MEANS ALLOW 100,100 USERS IF LIST BELOW IS USED

        NIL
        L-RADT
        L-AMET
        L-RMLT
        L-RTGT
        L-ISDT
        L-CCN/
       L-GWCT
        L-LOND
        L-NBST
        L-USCT
        L-ILL1
        NIL

[Jeff Shrager]

So you made Parry run? Want to share some computational paranoia with us? :-)

[Lars Brinkhoff]

Jeff Shrager wrote:

So you made Parry run? Want to share some computational paranoia with us? :-)

Not much.  I don't know how make an interesting conversation with Parry.

But here's a little sample of a run.

.R PARRY

END INPUT PARAMETERS WITH CARRIAGE RETURN OR ALTMODE

PRINT NON VERBAL FEATURE? [Y,N]
*N

VERSION [WEAK, MILD, STRONG]
*MILD

TRACE EMOTION VARIABLES? [Y,N]
*N

DO YOU WANT THE CORE DUMPED? [Y,N]
*N
 
END INPUT WITH A PERIOD OR QUESTION MARK,
   FOLLOWED BY CARRIAGE RETURN.
TO INDICATE SILENCE, TYPE   .
   WHEN FINISHED, TYPE   BYE.
USE PERIODS ONLY AT THE ENDS OF SENTENCES,
   NOT IN ABBREVIATIONS.

READY:
*HELLO.

HI

READY:
*HOW ARE YOU?

ALL RIGHT I GUESS

READY:
*HOW DO YOU FEEL?

I'M FEELING OK PHYSICALLY

READY:
*DO YOU HAVE ANY PROBLEMS?

I'M QUITE UPSET

READY:
*WHY ARE YOU UPSET?

PEOPLE GET ON MY NERVES SOMETIMES

READY:
*ELIZA, AMIRITE?

I'VE BEEN UPSET LATELY

READY:
*WHY ARE YOU UPSET?

I THOUGHT I TOLD YOU PEOPLE GET ON MY NERVES

READY:
*YES, YOU TOLD US.

ANOTHER THING IS GANGSTERS

READY:
*GANGSTERS ARE THE WORST.

ANYTHING ELSE?

READY:
*THUGS ARE PRETTY BAD TOO.

SO WHERE ARE WE?

READY:
*

[Jeff Shrager]

Well that was easy! Congratulations!

Would you like to post something about this someplace, or should I (on
ElizaGen.org and Hacker News?) Regardless, if you do post it
someplace, let me know and I'll put a note up on Hacker News.

'Jeff

[Lars Brinkhoff]

Jeff wrote this about running PARRY:

https://sites.google.com/view/elizagen-org/blog/parry-parries-again

[Lars Brinkhoff]

Jeff Shrager wrote:

I know that IPL-V ran on Dec-10s bcs the SAIL documentation mentions it!

Can you please point to (or provide) what documentation or files mention this DEC-10 IPL-V?

Second, did you check with RAND archivists?  Much of their catalog is online, but perhaps not all, and sometimes they will scan documents that are not yet in PDF form.  They have been quite helpful with regards to unconvering JOSS-II articles and listings.

Third, what other computers did IPL-V run on?

[Jeff Shrager]

Actually, upon re-locating the document I was referring to I see that
it's for the 709/7090:

https://drive.google.com/file/d/1gkT8Oj71WDsP2aCn0M_ogX_OcCn5RNBS

We know that IPL-V ran on many machines because of page 3 of the 1963 manual:

https://drive.google.com/file/d/1i_9BkBYcLy1cNw6Nve5TvGme9t8iZjl6

And since CMU was one of the earliest and most fervent adopters of the
PDP-10, and Newell and Simon were the bigwigs at CMU (probably called
CIT at that time!) iI guess I just assumed ... more like extrapolated
or speculated .... that they would have put IPL-V on their Dec10.

(Although, as Dr. House says "Never speculate because you make a spec
of you and someone named 'late'!")

'Jeff

> --
> You received this message because you are subscribed to a topic in the Google Groups "PiDP-10" group.
> To unsubscribe from this topic, visit https://groups.google.com/d/topic/pidp-10/vTSghb_v74Q/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to pidp-10+u...@googlegroups.com.

> To view this discussion visit https://groups.google.com/d/msgid/pidp-10/0a884ea0-e283-44d8-863f-6fe5da519b4cn%40googlegroups.com.

[David Todd]

According to the "IPL-V Programmers' Reference Manual" by Allen Newell, for Rand in 1963,
    https://www.rand.org/content/dam/rand/pubs/research_memoranda/2008/RM3739.pdf
on Summary page (v), IPL-V at that time ran on

• IBM 650, 704, 709, 7090, and 7094 (my second machine!)
• Philco 2000
• Bendix G-20
• CDC-1604
• Univac 1105
• AN/FSQ-32

[Jeff Shrager]

Thanks, David.

We know that IPL-V  ran on other machines. There was even a PC version, written in BASIC! (I’ve spoken to the author, Howard Richman, who does not have the code any longer.)

Since CMU was a hotbed of IPL-V (via Simon and Newell), and also one of the earliest PDP-10 adopters, I speculated that there would, at some point, have been an IPL-V on the 10. 

But you know what Dr. House says: “Don’t speculate, because you make a spec of you and someone named late.” :-)

By 1966, when the 10 appeared, IPL-V was dying, having been killed by Lisp, so it’s completely plausible that they never created a Dec10 IPL-V. 

(Newell had actually spec’ed an IPL-VI, which also never happened. Basically, everyone went over to Lisp. 

‘Jeff

----------

Jeff Shrager, Pi Rebel

Chief Scientist, Bennu Climate, Inc.