Snobol Books

There are many interesting items in here that a programmer would find interesting even today, and I'm not just saying that because the author is my dad. However, even a person with no interest in computer programming would greatly enjoy the parts I found most interesting, namely those dealing with constructing random poems, names, jokes, etc. One wonders if an unwitting audience member would recognize that the sonnet he is hearing, talking about "lusty twigs" and other absurdities, was a random concoction made with Dr. Gimpel's algorithms rather than a Shakespearian original.

SNOBOL use to be my favorite hacking language. It was concise, cryptic, known to few, and bizarrely powerful. For example, it took only a few lines to write a symbolic differentiation program, plus one line per transcendental function. Of course, it involved a kind of "computed goto" that works only in interpreted languages. Oh, and it was somewhat column oriented, helpful whe you're using punch cards.

Remember how C++, the definitive OO language, used to be a preprocessor for non-OO K&R C? SNOBOL is even worse - it was the ultimate string language for many years, written as an intepreter in the ultimate non-string language, FORTRAN IV. SNOBOL had recursion, garbage collection, typed data (with untyped variables), and lots more to delight the hacker's soul. FORTRAN didn't. A later version came out, called SPITBOL. It was a compiler, so lacked some of the interpreter's wilder features. It was still wicked awesome anyway. (Then there was ICEBOL, a program to compact SNOBOL programs. Really.)

But when I became a man, I put away the things of a child. It's just a hacker's language, not suited to the professional world. Much that was good in it lives on in Perl and Javascript, when loosely understood. The language itself is just a happy childhood memory. I just had a few moments putting the shine back on those memories - aah.

//wiredweird

In 1974 I was hanging around the University of Toronto's High Speed Job Stream: you could write programs on punch cards and run them on the IBM mainframe without a user account.

I was running FORTRAN programs on the system's WATFIV implementation, when one day I found a big, green textbook left behind by a hapless undergrad. I started perusing the book, and right away I saw that I was no longer in Kansas.

I was looking at the SNOBOL programming language, and the style of programming--using pattern matching, back tracking, succeed, fail, fence, and so forth--was utterly new. My brain exploded and I started writing programs for the University's SPITBOL compiler.

Years later, when I saw Prolog and Regular Expressions, I came to appreciate how innovative the SNOBOL team had been. I've ordered a copy of this text for my library, and I recommend it for anyone who appreciates the work of pioneers like Griswold.