Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83 v7 ucbtopaz-1.8; site ucbtopaz.CC.Berkeley.ARPA
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!genrad!decvax!ucbvax!ucbtopaz!mwm
From: mwm@ucbtopaz.CC.Berkeley.ARPA
Newsgroups: net.lang
Subject: (qualityp BASIC) -> NIL
Message-ID: <733@ucbtopaz.CC.Berkeley.ARPA>
Date: Sat, 16-Feb-85 23:38:17 EST
Article-I.D.: ucbtopaz.733
Posted: Sat Feb 16 23:38:17 1985
Date-Received: Tue, 19-Feb-85 08:33:19 EST
References: <7873@brl-tgr.ARPA> <706@ucbtopaz.CC.Berkeley.ARPA> <467@houxj.UUCP>
Reply-To: mwm@ucbtopaz.UUCP (Praiser of Bob)
Organization: Missionaria Phonibalonica
Lines: 43
Summary: 


Yes, you can find external reasons for using a language. Including BASIC,
COBOL, PL/I and FORTRAN (the whipping boys of the CS community). Just such
reasons have had me writing machine code.

However, first and foremost, a programming language is a tool for using a
tool (your computer) to solve problems. [Ohmigod, I said that like it was a
fact! Gulp! Get out the asbestos suite.] Like any other class of tools, some
languages are better for solving a specific problem than other languages.

For example:

I've as yet to find a better tool for most raw number crunching than FORTRAN
(but give me 77!).

For building tools to work on my tools, where speed is important but not
vital (kernel hackery & such), C does a good job.

For shuffling symbols, and breadboarding algorithms, LISP is marvelous
(but I may change my mind, now that Little Smalltalk is here...).

For tweaking strings, Icon.

For quick and dirty calculations, one of calc, bc, or LISP.

As a side flame, I'm bothered by the number of people who don't seem to
understand this, and start thinking about the code before they consider
the language. Does nobody teach that choosing the right tool for a job
is important?

Now, back to the subject line (maligning BASIC :-). I know of *no* class
of problems for which anything even close to ANSI minimal BASIC is the
best tool for the job. True, there are extended BASICs that are useful,
but most of those look more like Pascal than BASIC. However, those aren't
portable to other BASICs, and some of them choke on ANSI minimal BASIC.
It's like the Pascal problem: you either get something inadaquate, or you
get something unportable.

Even COBOL does better than that, as there is a small class of problems
where COBOL would be preferable - at least until somebody puts the
CORRESPONDING keyword into a modern language. (Hmm x CORR= y?).

	<mike