Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!clyde.concordia.ca!uunet!tut.cis.ohio-state.edu!cs.utexas.edu!longway!std-unix
From: std-unix@longway.TIC.COM (Moderator, John S. Quarterman)
Newsgroups: comp.std.unix
Subject: Re: P1003.1 "Trial Use"
Message-ID: <472@longway.TIC.COM>
Date: 11 Dec 89 03:51:21 GMT
References: <471@longway.TIC.COM>
Reply-To: Doug Gwyn <uunet!brl.mil!gwyn>
Organization: Ballistic Research Lab (BRL), APG, MD.
Lines: 40
Approved: jsq@longway.tic.com (Moderator, John S. Quarterman)

From: Doug Gwyn <uunet!smoke.brl.mil!gwyn>

In article <471@longway.TIC.COM> alan@s5000.RSVL.UNISYS.COM writes:
>The P1003.1 "POSIX" standard went thru a 1 year "trial use" period. Was this
>a useful and productive process? What were the results of the "trial use" 
>period, and how were they incorporated into (or omitted from) the final 
>standard?  Does this meet some of the criticism that is currently being 
>brought against 1003 that it is going too fast too soon? 

I think the trial use period was utterly useless.  There was not
sufficient time for the tentative standard to be implemented and
made widely available commercially, and certainly not enough time
to make conformance to the tentative standard a keystone of
software development or system specification efforts.

The "Interim FIPS" also hurt the quality of the standard by forcing
completion at too rapid a rate.  For evidence of this, consider the
drastic nature of the changes that occurred in the proposed standard
DURING THE BALLOTING PROCESS.

>P1003.2 and P1003.3 are currently in balloting; P1003.4 will be balloted
>beginning in January. I have not heard of any "trial use" period for these
>standards. Should the "trial use" concept be applied to these standards?

No.  What I think SHOULD be done is to publish the proposed standards
for public review and comment, rather than keeping discussion limited
to a small number of people most of whom who wrote the standards.

I don't think any standard that has gone through the hasty, unchecked
procedures these 1003.n standards are going through should be adopted
in any mandatory context (e.g. FIPS).  In fact I think some of the
1003.n standards are entirely uncalled for, for example the ones for
graphical user interfaces and "transparent network file access".
(POSIX file semantics are already covered in 1003.1, and we were
careful to consider what reasonable network file systems should be
required to do.  NFS was not considered POSIX-conforming.)  How does
one block adoption of an unwarranted standard, anyhow?  Can this
juggernaut be stopped?

Volume-Number: Volume 17, Number 99