goer%sophist@GARGOYLE.UCHICAGO.EDU (Richard Goerwitz) (12/06/90)
(I didn't see anyone else respond, and it seems worthwhile, I'll try to answer this question. Note that I am not part of the Icon Project, and that some of what I say may not be entirely accurate :-)). Re: Is ICON the successor to SNOBOL? Sure. They started working on SL5, but suddenly felt inspired with the knowledge that the backtracking mechanisms found in string scan- ning could be superimposed on a consistent, general-purpose program- ming language with a modern, procedurally oriented program struc- ture. Icon is, at once, like Snobol string processing gone wild, and at the same time like languages such as C and Pascal. I've also heard it described as an Algolish Prolog as well, because it essentially im- plements a backtracking mechanism somewhat like that of Prolog in terms of control structures we all know and "love." It is not really quite a *successor* to Snobol, but it is certainly a logical step forward. I don't believe SL5 (which *would* have been the successor to SNOBOL4) was ever released. If you know Snobol, don't expect the conversion to be immediate. You can write a 3-line Arb() procedure, if you like. But it's not built in. Icon offers low-level tools with which to construct your own matching functions. It is very flexible, but what I'm trying to say is that, at first, Snobol programmers often feel a bit lost. Perhaps some former Snobol-ers can comment on this statement. Icon is nice in the sense that it runs on almost all small and mid-range machines, and on many high-end ones. It's very portable, and is pretty much free. -Richard