Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!unmvax!brainerd
From: brainerd@unmvax.unm.edu (Walt Brainerd)
Newsgroups: comp.lang.fortran
Subject: Re: Character aliases are Satanic extensions
Keywords: Using SAP inclusions
Message-ID: <100@unmvax.unm.edu>
Date: 31 May 89 04:12:17 GMT
References: <598@mbph.UUCP>
Organization: University of New Mexico at Albuquerque
Lines: 48

In article <598@mbph.UUCP>, hybl@mbph.UUCP (Albert Hybl  Dept of Biophysics  SM) writes:
> 
> One thing the Army knows better than anyone else is warfare,
> but they are not allowed to declare WAR.
> 
True, but maybe not relevant. If you are trying to make an analogy
with the Fortran standards activity, you would be proposing
that generals not sit on councils that advise people who declare war.
Are you proposing that no vendors be allowed on standards committees?
I don't see any way that could happen under our
current system of standards making by ANSI, although, given what has
happened to X3J3 over the last few years, it seems tempting.
These years have convinced me that something should change, but
I am not sure I have any specific proposals that are politically feasible.
Would love to hear some.

Incidentally, some of the vendors on X3J3 have been very supportive
of the efforts and have made execellent contributions to the technical work.

> This is an excellent example of a calumniatory sentence.  Notice
> the innuendo--ONLY STUPID PROGRAMMERS USE VENDOR EXTENSIONS!

Wrong!
IGNORANT programmers also use vendor extensions (there is a lot of
difference).  Also INTELLIGENT programmers use extensions
when they need them to get a special job done.  I believe there are some
very useful things that do not belong in the standard (one reason
being that they are not portable).  Perhaps the user-defined character
facility I mentioned previously is a good example.  It is useful
to certain groups of programmers, but is _not_ portable, and therefore
does not belong in the standard.

We seem to have two proposals (please don't anyone say these are MY proposals):

1)  No vendors should be allowed on standards committees.
    If there is some more mellow variation of this,
    that might be interesting to hear.

2)  The Fortran standard should state that the processor does not
    conform to the standard if it accepts extensions.

I would like to see some rational discussion of these issues by other
readers of these items.  Another point to keep in mind is that adoption
of (2) means that the standard would not be upward compatible with
the previous standard.  Remember what happened when we tried to schedule
deletion of some important redundant features for the year 2030?

Walt Brainerd   brainerd@unmvax.unm.edu   505/275-0800