4 ms·

History of Xerox Interlisp and BBN-Lisp

72 points by mepian 3y ago

gwern 3y ago
The description of DWIM and other early IDE-like features starting at https://interlisp.org/history/timeline/#1970 https://interlisp.org/history/timeline/#1970 is good.
- 082349872349872 3y ago
  Judging from Teitelman's post-retirement activities, a cynic might say that he found dogs to be much better than computers at DWIM.
ggm 3y ago
Amazing history. I know of people like Daniel Bobrow as somebody who was coming to dinner in Edinburgh, with amusing stories (I was like 11 at the time) in the early 70s. This, in a household which I could characterise as "LISP haters" although thats not strictly true, it just felt like it sometimes. I think we had a plastic soup ladle he burned on the cooker which we called "dannys ear" because of a bubble in the plastic which stuck out like an ear on a bald head.
gumby 3y ago
I hadn’t realised that Danny Bobrow and Alice Hartley invented the spaghetti stack — I’d thought it was RMS a couple of years later. There was a bug in the spaghetti stack code in Interlisp D which I triggered at PARC in mid 1984 (basically it spaghettied the stack every time you made a closure. I spawned hundreds…which brought things to a crawl. JonL eventually fixed it. On DWIM: Interlisp used it to do macro expansion so you could not disable it.
- kragen 3y ago
  have you tried the recent revival of interlisp and medley? i'm interested to hear what you think
  - gumby 3y ago
    Those days are in my past. Also though medley had some good ideas I preferred the MIT-style machines and commonlisp. I will say that in general PARC’s cloud services of the late 70s/mid 80s were still superior to what we have access to today. Another case of “worse is better”. Sadly it feels today that people think that means “worse is preferred”. It’s like all the people who know the title “Goto considered harmful” but never read the paper, so grossly misunderstand it.
- amszmidt 3y ago
  Maybe you are thinking of Phantom Stacks[1]. Spaghetti stacks where created by Bobrow and Wegbreit in 1972. [1]: https://dspace.mit.edu/handle/1721.1/6331 https://dspace.mit.edu/handle/1721.1/6331
  - gumby 3y ago
    Ah yes, that’s surely what I am conflating! I didn’t think the phantom stack idea was that great (but it was just an idea worthy of a working paper, so no big deal). Also I really don’t think spaghetti stacks were so great in retrospect, especially when implemented on modern architectures that lack the lightweight stack support of the PDP-10 which probably inspired the design.
pjmlp 3y ago
The quote at the end says it all > “Medley is and has been a system before its time. Through the debuggers, the graphics and window systems, and the organized structure of the underlying Lisp systems, Medley provides a marvelous tool that creates enthusiasm and motivation for programmers, applications builders, application users, teachers, and students alike.” Interlisp, Smalltalk ideas also ended up on the systems programming environment at Xerox from Mesa (initially XDE), and then Mesa/Cedar. Experiences that to this day are only fully available in (some) Lisp, and Smalltalk descents, or partially in JVM/CLR/BEAM environments (with related IDEs).
kragen 3y ago
lars brinkhoff pointed out the interlisp-11 report https://apps.dtic.mil/sti/tr/pdf/ADA066142.pdf https://apps.dtic.mil/sti/tr/pdf/ADA066142.pdf
timonoko 3y ago
TIL: Teletype structural editor was from 1968. It was a marvel. Somebody tried such approach for Algol too, I remember. I myself thought it might be possible to make similar editor for free text. You could see whole book on one TTY-line. Few character from each chapter and when you choose chapter you can see few characters from each paragraph etc. Unfortunately Finnish Government gave me video terminal and I forgot this project.

72 points by mepian 3y ago

The description of DWIM and other early IDE-like features starting at https://interlisp.org/history/timeline/#1970 https://interlisp.org/history/timeline/#1970 is good.

082349872349872 3y ago
Judging from Teitelman's post-retirement activities, a cynic might say that he found dogs to be much better than computers at DWIM.

ggm 3y ago

Amazing history. I know of people like Daniel Bobrow as somebody who was coming to dinner in Edinburgh, with amusing stories (I was like 11 at the time) in the early 70s. This, in a household which I could characterise as "LISP haters" although thats not strictly true, it just felt like it sometimes. I think we had a plastic soup ladle he burned on the cooker which we called "dannys ear" because of a bubble in the plastic which stuck out like an ear on a bald head.

gumby 3y ago

I hadn’t realised that Danny Bobrow and Alice Hartley invented the spaghetti stack — I’d thought it was RMS a couple of years later. There was a bug in the spaghetti stack code in Interlisp D which I triggered at PARC in mid 1984 (basically it spaghettied the stack every time you made a closure. I spawned hundreds…which brought things to a crawl. JonL eventually fixed it. On DWIM: Interlisp used it to do macro expansion so you could not disable it.

kragen 3y ago
have you tried the recent revival of interlisp and medley? i'm interested to hear what you think
- gumby 3y ago
  Those days are in my past. Also though medley had some good ideas I preferred the MIT-style machines and commonlisp. I will say that in general PARC’s cloud services of the late 70s/mid 80s were still superior to what we have access to today. Another case of “worse is better”. Sadly it feels today that people think that means “worse is preferred”. It’s like all the people who know the title “Goto considered harmful” but never read the paper, so grossly misunderstand it.
amszmidt 3y ago
Maybe you are thinking of Phantom Stacks[1]. Spaghetti stacks where created by Bobrow and Wegbreit in 1972. [1]: https://dspace.mit.edu/handle/1721.1/6331 https://dspace.mit.edu/handle/1721.1/6331
- gumby 3y ago
  Ah yes, that’s surely what I am conflating! I didn’t think the phantom stack idea was that great (but it was just an idea worthy of a working paper, so no big deal). Also I really don’t think spaghetti stacks were so great in retrospect, especially when implemented on modern architectures that lack the lightweight stack support of the PDP-10 which probably inspired the design.

pjmlp 3y ago

The quote at the end says it all > “Medley is and has been a system before its time. Through the debuggers, the graphics and window systems, and the organized structure of the underlying Lisp systems, Medley provides a marvelous tool that creates enthusiasm and motivation for programmers, applications builders, application users, teachers, and students alike.” Interlisp, Smalltalk ideas also ended up on the systems programming environment at Xerox from Mesa (initially XDE), and then Mesa/Cedar. Experiences that to this day are only fully available in (some) Lisp, and Smalltalk descents, or partially in JVM/CLR/BEAM environments (with related IDEs).

kragen 3y ago

lars brinkhoff pointed out the interlisp-11 report https://apps.dtic.mil/sti/tr/pdf/ADA066142.pdf https://apps.dtic.mil/sti/tr/pdf/ADA066142.pdf

timonoko 3y ago

TIL: Teletype structural editor was from 1968. It was a marvel. Somebody tried such approach for Algol too, I remember. I myself thought it might be possible to make similar editor for free text. You could see whole book on one TTY-line. Few character from each chapter and when you choose chapter you can see few characters from each paragraph etc. Unfortunately Finnish Government gave me video terminal and I forgot this project.