Path: utzoo!utgpu!watmath!clyde!mcdchg!chinet!att!rutgers!ukma!uflorida!novavax!hcx1!hcx2!bill
From: bill@hcx2.SSD.HARRIS.COM
Newsgroups: comp.lang.fortran
Subject: Re: Brainerd Comments - Response
Message-ID: <44400027@hcx2>
Date: 28 Oct 88 14:01:00 GMT
References: <758@a.lanl.gov>
Lines: 38
Nf-ID: #R:a.lanl.gov:758:hcx2:44400027:000:2262
Nf-From: hcx2.SSD.HARRIS.COM!bill    Oct 28 10:01:00 1988


"Best effort", I think, means that each member of X3J3 did their best to
represent what they thought the "majority" of users want in FORTRAN.  But,
the exchanges posted recently in this news group make it pretty clear (to
me, at least) that there is no "majority" position among users.  We have
users strongly advocating everything, or almost everything, in the current
draft.  We have other users equally strongly protesting.  We have the
public review in which, although a clear majority disapproved of the draft,
many of the controversial features received at least some support.

FORTRAN has become a "jack of all trades" to many people.  Many scientists,
engineers, physicists, etc. want a language that is concise, easy to use,
relatively simple, and that gives them access to all the power of the
machine.  Other users, including in some cases colleagues of those same
scientists and engineers and physicists, want a "safe", powerful language,
a language well-suited to software engineering.

Isn't it time we admitted that one language cannot do all?  Yes, I
understand all the arguments for using the same language.  I suppose that
carpenters would prefer that one tool could hammer nails, saw boards,
measure distances, and smooth wood; but instead they carry hammers, saws,
rulers, and planes.

Perhaps also the arguments made in support of using one language have as
their basis an invalid assumption: that it is a good idea to try and take a
research project directly to a production-quality product without redesign.
All attempts I have seen to do that with hardware have been dismal
failures; there is no reason to believe software is immune from the causes
of those failures.

If you try to make FORTRAN be all things to all people, the effort will be
a failure.  However, I also predict that no one will listen to this
prophecy, and we (X3J3) will continue to try to do just that.  The real
shame is that X3J3 will be blamed for the failure, when we are merely
trying to do what users are asking of us.  Each member, of course, has a
different opinion of what the requests are, but each is quite honest in
believing his/her opinion is correct.  The bare fact is, however, that we
are being asked to perform an impossible task, and we will fail.