Path: utzoo!attcan!uunet!husc6!bloom-beacon!mit-eddie!uw-beaver!teknowledge-vaxc!sri-unix!quintus!ok
From: ok@quintus.UUCP (Richard A. O'Keefe)
Newsgroups: comp.unix.wizards
Subject: Re: Open Software Foundation
Message-ID: <1045@cresswell.quintus.UUCP>
Date: 1 Jun 88 07:22:14 GMT
References: <14976@brl-adm.ARPA> <54822@sun.uucp> <15812@uflorida.cis.ufl.EDU> <4538@vdsvax.steinmetz.ge.com>
Organization: Quintus Computer Systems, Mountain View, CA
Lines: 20
Posted: Wed Jun  1 00:22:14 1988

In article <4538@vdsvax.steinmetz.ge.com>, barnett@vdsvax.steinmetz.ge.com (Bruce G. Barnett) writes:
> As I heard it, one of the reasons was that Sun wanted to better
> understand the needs of the customers. When all of the compilers are 
> bundled (i.e. 'free'), they had no way to measure how 'popular'
> the compiler was.

One method which has been tried by other manufacturers is surveys.
I remember Burroughs did one for the mainframes that I saw (in fact I
learned about a programming technique on those machines from that survey;
they asked whether anyone was doing X, and that was the first time I ever
heard that X was possible).

Another possible source is bug reports.

If you are in the business of producing, say, a library package which is
supposed to be callable from Fortran, Pascal, and C, your package might
be written entirely in C, and the only use you might have for Fortran or
Pascal might be to test the interface every so often.  The utility of the
Fortran and Pascal compilers is then much lower to you, but the price is
set by other customers who are using them all the time.  GF77, anyone?