[net.lang] Snobol and C

bzs@bu-cs.UUCP (Barry Shein) (12/17/85)

>From: abh6509@ritcv.UUCP (Xandrew)
>In article <2433@hcradm.UUCP> steve@hcradm.UUCP (Steve Pozgaj) writes:
>>Hey, c'mon people!  Something very obvious has been missed here.  Namely,
>>the damn program should have been written in SNOBOL in the first place.
>>You want to write operating systems, you use C; you want to match patterns,
>>you use SNOBOL:-)
>>--
>>Everyone's a SNOBOL freak at heart!  It's only the brave who admit it.

>How right you are! Now show us how to link a SNOBOL file into a C module......

Trivial, use a couple of pipes...(unless you think it would be faster
to re-write all of Snobol's functionality in C.)

	-Barry Shein, Boston University

perl@rdin.UUCP (Robert Perlberg) (12/20/85)

I always considered awk(1) to be a marriage of Snobol and C.  In fact,
the awk manual even mentions Snobol in describing its array features.
I always thought that if they just put a little more work into it, they
could develop awk into a real language that could call C routines and
vice versa but retained awk's pattern matching and string splitting and
string-indexed arrays and auto-type-conversion.  Does anybody know why
this wasn't done/shouldn't be done/can't be done?

Robert Perlberg
Resource Dynamics Inc.
New York
{philabs|delftcc}!rdin!perl

roger@rtech.UUCP (Roger Rohrbach) (12/30/85)

> I always considered awk(1) to be a marriage of Snobol and C...
> I always thought that if they just put a little more work into it, they
> could develop awk into a real language...  Does anybody know why
> this wasn't done/shouldn't be done/can't be done?
> 
> Robert Perlberg

	If you pick up a copy of "The Icon Programming Language" by
R.E. & M.T. Griswold (published in 1983 by Prentice-Hall), you will
see that Snobol's powerful successor fits your bill (although it
resembles C less than awk does), providing awk's capabilities and
beyond in the context of a procedural language.  Somehow I suspect
that most C programmers will prefer this to a pattern-directed language
such as awk, because it at least appears to be more general  (Kind of
like Lisp vs. Prolog).

					Roger Rohrbach
					{ucb,dec}vax!mtxinu!rtech

ark@alice.UucP (Andrew Koenig) (12/31/85)

> 	If you pick up a copy of "The Icon Programming Language" by
> R.E. & M.T. Griswold (published in 1983 by Prentice-Hall), you will
> see that Snobol's powerful successor fits your bill (although it
> resembles C less than awk does), providing awk's capabilities and
> beyond in the context of a procedural language.  Somehow I suspect
> that most C programmers will prefer this to a pattern-directed language
> such as awk, because it at least appears to be more general  (Kind of
> like Lisp vs. Prolog).

I have used both Snobol and Icon and prefer the former, once I can get
around the crummy syntax.  I am aware, though, that the problem may
be that I learned Snobol first and have not yet had the incentive
to steep myself in Icon lore.

One nice thing about Icon -- it's free.