marketing unix

[Erik E. Fair]

After being quoted so widely (and defended very well; Thanks Guy & friends),
I feel somewhat obligated to respond. Guy Harris expanded on my terse jab
eloquently, but the following pokes a *very* sore spot:

>> From: jpl at sftig.UUCP
>> Date: Mon, 16-Jul-84 06:35:45 PDT
>> Organization: AT&T Bell Laboratories, Summit, NJ
>> 
>> "Paging Mr. Harris, paging Mr. Harris, please come down to earth."
>> 
>> Readers will recall from the last referenced article that one of
>> Guy's major point's was that adopting other people's ideas is a
>> useful and sensible thing to do.  Quite so.  However, this
>> entails evaluating the ideas and the particular instantiation
>> of those ideas, and not just swallowing them whole.
>> 
>> Also, in future, when Mr. Harris plays the Great Reconciler,
>> will he please refrain from baiting USDL (? Unix System Development Lab ?)
>> members?  They are certainly NOT -- what was the phrase? -- Berkeleyphobics.
>> Au contrair, they generally seem capable of building on the best ideas
>> for UNIX enhancements, whether originated at UCB or other universities,
>> internal or external research labs, or even the occaisional original
>> idea.
>> 
>> jeff lankford	sftig!jpl

Mr. Lankford,

	I have seen more evidence of an epidemic class case of
the famous ``Not-Invented-Here'' syndrome in the people who bring us
System V than any other. I have never met anyone (at least not
knowingly) who works for USG/USDL/(whatever-they're-calling-it-this-week)
but I read their C code both in the Kernel and in the utilities.  It is
sloppy. It is in some cases unportable (a very important thing where I
work; we do UNIX on a 68000, not a VAX or a 3B20). It does not often
pass lint unscathed. But worst of all, it reflects a disturbing
thing: that these people don't know how to complete something
properly (e.g.  the System III/V tty driver changes from V7: they
changed the kernel, but they didn't do a complete conversion of the
utilities, so they left a horrid compatability hack for stty(2) and
gtty(2) in the kernel & C library, which didn't always work.
Fercrissakes, ``getty'' still had stty/gtty calls in it!)

I'm terribly afraid that AT&T is turning into IBM in the sense that they
do backward compatability forever, and are terrified of fixing anything
that is broken, just because it has always behaved that way. I cite the
4.2 BSD signal mechanism (although the implementation is messier) is the
*right* way to do signals, and they should have been done that way in the
first place. I very much doubt that USG/USDL will ever try to fix it. I
echo the doubt of others that System V will ever do networking decently.
Will they ever fix the filesystem? (and so on...)

And if they're not Berkeleyphobic, then why don't System V VAXEN page?
They've had three years to evaluate it...

	(flame off!)

	Erik E. Fair	ucbvax!fair	fair at ucb-arpa.ARPA

	dual!fair at BERKELEY.ARPA
	{ihnp4,ucbvax,hplabs,decwrl,cbosgd,sun,nsc,apple,pyramid}!dual!fair
	Dual Systems Corporation, Berkeley, California

P.S.	I wish I could see the look on their faces when IBM announces 4.2BSD
	for their entire line of computers...