Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!utgpu!water!watmath!clyde!cbatt!ucbvax!jade!eris!mwm From: mwm@eris.UUCP Newsgroups: comp.sys.amiga Subject: Re: Language Wars Message-ID: <3245@jade.BERKELEY.EDU> Date: Tue, 21-Apr-87 06:28:36 EST Article-I.D.: jade.3245 Posted: Tue Apr 21 06:28:36 1987 Date-Received: Thu, 23-Apr-87 23:47:45 EST References: <8704210744.AA10632@ingres.Berkeley.EDU> Sender: usenet@jade.BERKELEY.EDU Reply-To: mwm@eris.BERKELEY.EDU (Mike (My watch has windows) Meyer) Organization: Missionaria Phonibalonica Lines: 47 In article <8704210744.AA10632@ingres.Berkeley.EDU> hatcher@INGRES.BERKELEY.EDU (Doug Merritt) writes: >To an average programmer, every language and dialect is importantly >different than all others. True. And to a good programmer, those differences are what determines which one gets used for what jobs. >To a language designer, most languages look more similar than different, >because he considers more abstract issues. Thanx for the words of support, Doug. But I'd like to point out that the problems I mentioned are things I see as a programmer, not as "language designer." Never tried GP language design, and probably won't for a long while. There are still to many language I haven't had a chance to play with (for instance, anyone got an Occam for the Amiga? :-). >For pragmatic reasons, C and Modula-2 can both be good languages to >implement things in. For theoretical reasons, languages like Smalltalk >and FP are much cleaner. FP is kinda clunky; even major functional programming fans admit that you need some way to model the passage of time. On the other hand, smalltalk is slick, and gives one of the three "good" typing models (any object that supports what we wanna do is of the correct type). As an example, there's a v8 tool called "mc". It reads in a file (usually stdin), finds the widest line, and then outputs the result in a multi-columned format, each column being wide enough to hold the widest line. I did a version in C, just to play with it. I then rewrote it in CLU (CLU is to modula/ada/euclid as C is to pascal/algol/fortran :-). The only real change was to put in the set of arguments that playing with the first version had indicated were usefull, and tweaks to the algorithm for handling them. The CLU version was notably shorter and easier to read. The resulting binary ran faster, and was more robust. Now CLU is a language I could see laying out a couple of hundred dollars for, over and above C. Anyone got one they want to sell?