AT&T clean UNIX

[Dick Dunn]

rbraun at spdcc.COM (Rich Braun) writes:
> dab at ubitrex.mb.ca (Danny Boulet) writes:
[news thing in _UNIX_World_ about Berkeley Software Design, etc.]
> _Allow_ them to sell it at this price, notably higher than any existing
> 386-based Unix O/S products?...

The article announced an initial price of $995.  Now, there's no real
detail there, but let's consider that:
	- It's unlikely to contain a 1-2 user limitation!  (Putting that
	  sort of limit in a source-licensed product would be an exercise
	  in futility.)
	- It's pretty likely to contain networking, since that's the first
	  batch of code Berkeley ever cut loose on its own.
	- It's got to have a development system.
Who knows whether it'll contain X, but even with the above, the price is
better than all the SysV 386-UNIX vendors.  There's 386 Mach from mt Xinu
which is currently $995 for a full system (w/dev sys, networking, X), and
that seems a good deal for the UNIX-knowledgable purchaser.

But the main difference is that this BSD package is supposed to contain
source.  I shouldn't have to remind you that the source license for AT&T
V.n for 386 is a tad more than a kilobuck (like about two orders of mag-
nitude more).  Sounds like a pretty good deal if/when it's real.

>...And using the word Berkeley in the company
> name is just plain un-cool:  I hope the folks at the real Berkeley cry
> foul.  This company even ripped off the BSD acronym:  it's called
> "Berkeley Software Design Inc.", and the article on p.16 of UnixWorld
> makes no mention of any affiliation or non-affiliation with BSD...

I'll have to go along with Rich that when someone first mentioned the
article to me, I was awfully skeptical.  In fact, I said "frankly I don't
believe it."  After I got a look at the full article (such as it was), I
was a lot less skeptical.

The two names associated with BSD, Inc. in the article are Bill Jolitz and
Donn Seeley.  Bill has worked with the Berkeley CSRG folks off and on for
quite a while; if you remember 2.9 BSD (?a decade ago?), a lot of the
kernel work for that was his, ??I believe done while he was with USGS??
Bill did the bulk of the port to the 386.  He was showing 386 BSD on a
laptop at Anaheim (summer) USENIX last year.  There's a series about the
port and the internals of the system, by Bill and Lynne Jolitz, running in
_Dr_Dobbs_ right now.  It started in January of this year.  Donn has also
worked with the Berkeley folks for some time.  I believe he did work with
shared libraries and dynamic loading, although I'm not sure about that.

So anyway, conjecturing wildly, I'd guess that a name like "Berkeley Soft-
ware Design" and the abbreviation to BSD was probably done with the Berk-
eley CSRG folks knowing all about it.  It would be hard for them to have
been underhanded about it, given the close association.

>...article does point out that BSD Inc.'s engineers have been "loaned" to
> Berkeley for the 4.4 O/S (my assumption, correct me if I'm wrong, is that
> BSD Inc. provided a couple of consultants to Berkeley for an hourly fee.)

See above.  These folks have been associated with Berkeley CSRG for some
time.  I have no idea what financial arrangement exists; hard to see how/
why it matters.

> Sorry I'm such a cynic, but when I read that article I came away with a
> rather bad feeling about this.  Were the list price significantly under
> $1995, perhaps I might not have this reaction.

Again, I don't see how $2k (initially $1k) for full source is anything less
than an incredible bargain.  Nothing in the description indicates that it's
oriented to being an end-user system, and it's not set up to compete with
current commercial UNIXes (yet!), but it looks like a very interesting
tool for education and research as soon as it's out.

That's as much as I can contribute, and a lot of it is conjecture.  I'll
second Rich's plea for more info; I hope someone like Bill or Lynne can
give us an explanation and an update on the 386 BSD-based system.
-- 
Dick Dunn     rcd at ico.isc.com -or- ico!rcd       Boulder, CO   (303)449-2870
   ...Simpler is better.