Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!cbosgd!gatech!ut-sally!im4u!info-sequent-request
From: info-sequent-request@im4u.UUCP
Newsgroups: mod.computers.sequent
Subject: re: departure from the list, with discussion
Message-ID: <8603201812.AA10394@icst-cmr.ARPA>
Date: Thu, 20-Mar-86 13:12:11 EST
Article-I.D.: icst-cmr.8603201812.AA10394
Posted: Thu Mar 20 13:12:11 1986
Date-Received: Sat, 22-Mar-86 05:31:59 EST
Lines: 69
Approved: info-sequent-request@im4u.UUCP
Posted-Date: Thu, 20 Mar 86 13:12:11 est

From: rbj@icst-cmr.arpa (Root Boy Jim)

/*
	To suggest, as Henry Spencer does, that Sequent is using a system of
	equations of the form:
	
		rich = integrity
		poor = dishonesty
	
	to measure access to its sources is to ignore completely the issue of
	protecting a valuable trade secret.

I don't buy this. Who are they worried about? Encore? Alliant? Those guys
probably already have the source. Admittedly, they may have had to *pay*
for it, but the only true security is total secrecy. Penalizing your
legitimate customers because of competition paranoia is regrettable.
	
	I speak as someone whose organization has a Balance 8000 installed, a
	Balance 21000 on order, sources for Dynix 2.0.1, and experience with
	multi-processor systems.  We have four dual VAXs of the Purdue design,
	and several older CDC 6000 systems - 6500, 6600.  I have written a
	multi-processor operating system for the CDC 6000s and have installed
	George Goble's dual CPU UNIX system on the VAXs.
	
	Someone who has does not have that kind of experience might not
	appreciate that Dynix is an advanced and innovative multi-processor
	operating system.  It is unique.  Sequent should not surrender its
	competitive edge causually.

Big deal. Multiprocessor systems are not much more complicated than
uniprocessors, at least not in the software. Designing a hardware
configuration that will make it run fast is much more innovative.

The approach has been discussed at least twice in the famed $27 BSTJ issue
and probably as well in USENIX literature. Besides, it's no secret.
	
	I think that Sequent has taken a reasonable approach - charge enough
	for the advantage to recover its costs.  That's why Dynix is expensive;
	that's why the HP-35 cost so much when it came out, etc., etc.  The
	real equation is:
	
		price = cost of development
	
As Henry said, the advantage is mostly in the binarys. Sources enable the
end user to make fixes or enhancements that the vendor either doesn't
have the time or expertise for or won't. They're also educational.

UNIX (when there was only one) used to come with sources. Had this not
been the case, I doubt UNIX would have succeeded.  The fact that
Sequent, because it has a `unique' (multiprocessor), or SUN, because it
has a proprietary memory management board (or whatever) refuses to make
its sources available at a reasonable price is to revoke a legacy. The
fact that TPC also does is no excuse.

	We spent a lot of time discussing this issue of access to sources with
	Sequent.  They were informative discussions.  Sequent listened to us;
	we listened to them.  In the end, we came to an arrangement that
	satisfied both of us.  I'm sure that anyone else who might take the
	time to do the same will be able to afford the Dynix sources.
	
	Vic Abell, abe@asc.purdue.edu, ...!pur-ee!pucc-j!abe
	Purdue University Computing Center
	
Perhaps you should give Henry a call & coach him on what to say.
While I find Henry somewhat stuffy occaisionally, his integrity and
persona are well known on the net. Sequent doesn't need enemies like him.

	(Root Boy) Jim Cottrell		<rbj@cmr>
*/