Xref: utzoo comp.windows.news:1033 comp.lang.postscript:1337
Path: utzoo!attcan!uunet!husc6!encore!bzs
From: bzs@Encore.COM (Barry Shein)
Newsgroups: comp.windows.news,comp.lang.postscript
Subject: Re: NeWS and DPS,  who wins?
Message-ID: <4467@xenna.Encore.COM>
Date: 23 Dec 88 17:06:53 GMT
References: <926@cmx.npac.syr.edu> <13061@ico.ISC.COM>
Organization: Encore Computer Corp, Marlboro, MA
Lines: 62
In-reply-to: rcd@ico.ISC.COM's message of 23 Dec 88 00:19:58 GMT
Posting-Front-End: GNU Emacs 18.41.15 of Tue Jun  9 1987 on xenna (berkeley-unix)


The only computer software standards are de-facto standards, it
doesn't matter where the original software comes from. Standardization
is an ex post facto activity, not something you make up one night in a
smoke-filled room out of whole cloth.

The fallacy we are all having shoved at us is that someone can sit
down and write a piece of software and declare it a standard before it
has significant following and practice. This does happen on rare
occasion when the proferred solution fills a huge vacuum, NFS came
close to this, but it's rare.

Customers are now rushing about willy-nilly looking for the
"standard", and vendors are of course selling it to them at every
turn.

A standard is something which happens after accepted usage develops.

Fortran can be standardized because there is a great deal of overlap
between the many versions. Unix has a chance of becoming standardized
(remember, you'll rarely achieve a platonic standardization with no
variation, just some large set of features generally considered
essential) again as a result of broad use and practice to base a
standard on.

As a particularly bad example, ADA and ISO protocols are both
"standards" which depend on bureaucratic muscle from government
agencies for acceptance. As such they sort of mock the notion of a
standard although they are at least defined.  There is a slight
difference between, for example, ISO protocols claiming to be standard
ISO protocols (all fine and dandy, I agree) and ISO protocols claiming
to be the standard networking protocol (but, as with ADA, it's amazing
how a few billion dollars in DOD contracts focuses the mind!) After a
decade I'm still awaiting my ISO protocol suite written in ADA...

I remember finding out years ago that a DEC-20 used a "standard" power
plug (there it was, right in the standards!) which unfortunately
required a "standard" receptacle only manufactured by one company in
the world, and they wouldn't have any for months. Not much call for
that "standard", tho it certainly was standard. I think the analogy
with current standards fever should be evident.

Let's face it, "standards" have become the buzzword of the industry
and, like most buzzwords, means whatever they want it to mean (which
isn't much.)

In most cases replacing the word "standard" with "popular" separates
the wheat from the chaff:

	NFS is a popular remote file system protocol.
	Fortran is a popular programming language.
	DOD will require ADA to be popular.
	Adobe announces popular windowing system.
	Sun declares NeWS is already popular.
	Mumble denies X-windows popularity!
	ISO to be required as the popular networking protocol.
	TCP/IP to cease to be popular.
	OSF and Unix Int'l fight over which Unix will be mandated as popular.

It's really gotten very silly.

	-Barry Shein, ||Encore||