vintage screwlisp accountMar 10, 2024

@danhalbert
https://danhalbert.org/
I'm interested in your bachelor's thesis, since I think it also relates or complements somewhat @kentpitman 's https://www.nhplace.com/kent/Papers/Condition-Handling-2001.html for example; I think Kent has worked on python's sophistication in this department. (Actually, I don't use inspect particularly so who knows). Well, I have to look over both.

I see you've spanned BBN.. Xerox .. things I don't know .. more BBN .. modern times around AI. I wonder if you could comment about c. 1980 and then now

Dan Halbert

Show threadvintage screwlisp accountMar 10, 2024
@danhalbert
forgive me if it's tiresome; in Erik Sandewall's unpublished text/handbook draft from 10 years ago, he recollected that around 1980 the world became increasingly indifferent to programming language choice and this led to at least a splintering of different AI skeins. I've been trying to imagine my way through this a bit after reading the anthology of Barstow et al. Interactive programming environments with help from @kentpitman , larry, and others. I guess your life traversed this.
Show threadDan Halbert

@screwtape @kentpitman

> around 1980 the world became increasingly indifferent to programming language choice

I don't really understand this at all. There has been a ton of work on new programming languages and concepts since then, and nearly all of the most popular programming languages in use today were developed well after 1980.

What is true is that there is a lot more reuse of existing software than there used to be, so interoperability is more important.

Mar 10, 2024 at 2:13pmWeb
Show threadvintage screwlisp accountMar 10, 2024
@danhalbert
aha thanks. I probably overread a throwaway line in Sandewall's draft (chapter 7, choices for implementing AI - he says that in the 60s and 70s, lisp was clearly preferred for AI, but after about 1980 this stopped being true). Albeit his Leonardo system (host for his Software Individuals) was written in lisp. He does say the python world got to be similar to lisp, but he views lisp as being more specialised for AI than python (lisp symbols cf python normalised strings)
@kentpitman
Show threadKent PitmanMar 10, 2024

@screwtape @danhalbert

For a LONG time, people were dissing lisp for inefficiency, as if that were the world's only dimension of virtue. (The programming analog of shareholder capitalism, optimizing for nothing but profit as if nothing else in the world mattered.)

Google groups recently became harder to reference, but I was quoted by copy here at Slashdot with a story I once told about the kinds of things Lisp was up against: https://slashdot.org/comments.pl?sid=1238471&cid=28018063
There were many such small oppositions scattered about, enough that it eroded some of Lisp's hold.

My view (just one person's opinion) is that the switch away from lisp wasn't people nog caring as much about languages, but rather leberaging that people did still care and tarring lisp with criticisms far out of proportion with anything legit.

Later, as AI was oversold, promising to solve all problems for all people, managers explained that had these projects only been done in C++, capable of better efficiency, they'd have done better. Of course, the oriducts probably could never have been prototyped in C++, but being far along, some were readily recoded, which took time and kept projects alive as expectations were justified. After similar years of C++, of course, all the world's problenx weren't solved either, but attention had moved on to other issues.

That's not to say Lisp was best at everything.. no language is. There was way to much pressure to be all things to all people, but failing at that was something that should not have meant failing as a language or we'd have no languages at all.

In my later (2001) Slashdot interview (weird I have occasion to even mention it but twice in one essay is especially odd--so it goes), I tried to take a live and let live posture on languages that I think was well accepted. See URLs here since it was split over multiple editions: https://nhplace.com/kent/commentary.html

All of this just my personal opinion. It's fair for others to summarize this era very differently. We all see the world through the lens of our experience.

Wolfram|Alpha's Surprising Terms of Service - Slashdot

This whole "new kind of [whatever]" meme might be really funny if it weren't so sad -- not because Wolfram doesn't really think he is smarter than almost everybody else (he does), but because - reportedly - he can't be prevailed upon to care about what most other people think, let alone how his choi...