SNOBOL

[Matt Youell]

Does anyone know someone familiar with SNOBOL who would be willing and
able to give an interesting talk about it?

[Charlie Loyd]

On Nov 13, 7:34 pm, Matt Youell <m...@newmoniclabs.com> wrote:
> Does anyone know someone familiar with SNOBOL who would be willing and
> able to give an interesting talk about it?

I am not, but I would like to strongly second the request. SNOBOL/
SPITBOL seems like a fertile branch of the programming tree that never
got the exploration it deserved. Or perhaps its ideas were merged into
perl and Haskell -- I don't know; I'd have to hear a talk on it.

A math professor of my acquaintance once remarked to me:

> I did attempt a SNOBOL compiler once. This was in the last year
> of high school and I was looking (in the complete absence of any
> direction or advice or even a decent library) for a good project.
> The SNOBOL book must have turned up in a local bookshop. The
> compiler was written in MACRO-11 assembler for the PDP 11/10. It
> was pretty long as I recall. I got all the basics working and the
> string data type, but I never did all the fancy pattern-matching
> and high-level features. The biggest program I wrote in the
> compiler itself was one to play cribbage. That was interesting in
> itself. Based on which 5 cards the computer held, it scored each
> possible choice of discards and computed all the probabilities
> and it used to routinely get double my score. Everyone thought it
> was cheating!

Those were the days, I guess.

[Matt Youell]

On Nov 13, 7:50 pm, Charlie Loyd <cel...@gmail.com> wrote:
>
> I am not, but I would like to strongly second the request. SNOBOL/
> SPITBOL seems like a fertile branch of the programming tree that never
> got the exploration it deserved. Or perhaps its ideas were merged into
> perl and Haskell -- I don't know; I'd have to hear a talk on it.

Ditto. And as you noted, sounds like a challenge to compile. I'll keep
an ear to the ground.

[Matt Youell]

On Nov 14, 12:22 am, Matt Youell <m...@newmoniclabs.com> wrote:
> Ditto. And as you noted, sounds like a challenge to compile. I'll keep
> an ear to the ground.

Er... challenge to build a compiler *for*.

That's what I get for posting late at night.

[Mahmud Mohamed]

Icon is billed by its creators as a modern snobol.

[Matt Youell]

On Nov 14, 12:29 am, Mahmud Mohamed <bigthing...@gmail.com> wrote:
> Icon is billed by its creators as a modern snobol.

Icon is another language I'd like to see presented on. I didn't know
about the SNOBOL connection.

[Brian Rice]

Icon is definitely the most suitable placeholder for discussion about SNOBOL, and has a lot of available information. Someone could learn about it over the course of a few days and present effectively. No, I won't be in PDX until the weekend of December 3, before you ask. ;-)

--
-Brian T. Rice

[Matt Youell]

On Nov 14, 3:30 pm, Brian Rice <briantr...@gmail.com> wrote:
> Icon is definitely the most suitable placeholder for discussion about
> SNOBOL, and has a lot of available information. Someone could learn about
> it over the course of a few days and present effectively. No, I won't be in
> PDX until the weekend of December 3, before you ask. ;-)
>

First meeting is likely to be in February... Plenty of time. :)

[Brian Rice]

Oh, that's easier, then. I skimmed through the Icon manual and implemented >:= and <:= (comparison-conditional assignment) handily in Slate. Granted, the macro-expansion performs double-evaluation of the terms, but... :-)

FWIW, Icon generators have become relatively widely distributed, although perhaps not the feel of the programming style they engendered originally, compared to Python generators or the like.

--
-Brian T. Rice

[Bart Massey]

Local genius Mike Haertel knew Icon pretty well at one point in his life. I could ask him to talk about it if folks like.

Icon (as everyone may know) is actually a deliberate successor to SNOBOL from a project by Ralph Griswold, one of the SNOBOL creators. Unicon is the successor to Icon that gets the most attention these days. I may be mistaken, but I seem to recall that for a long time there were distribution restrictions on Griswold's Icon implementation?

[Matt Youell]

Bart,

That sounds great. Can you pull him into the list?

Matt