Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 30 Jan 1998 11:55:16 -0000
From:      Paul Richards <paul@originat.demon.co.uk>
To:        freebsd-alpha@FreeBSD.ORG
Subject:   RE: FreeBSD/Alpha 
Message-ID:  <E191A8FCC38DD111A84900E029115B40A1F5@OCTOPUS>

next in thread | raw e-mail | index | archive | help
On Wednesday, January 28, 1998 3:53 AM, Jordan K. Hubbard
[SMTP:jkh@time.cdrom.com] wrote:
> This is not to say that I hate Digital - far from it - just about
> every machine I've ever really admired came from that company.  It
> simply vexes me that there couldn't have been a lot MORE to my
> dealings with such a promising company over the years, the limitations
> on this generally being a direct result of various shortcomings in
> Digital's upper-level management.  Olson, for example, was openly
> hostile to the whole workstation concept and his "there can be Only
> One OS and its name is VMS" attitude hardly helped the company's Unix
> strategy.  Palmer, in turn, seems to have spent more of his time
> dodging falling masonry than in trying to articulate a workstation &
> server software strategy, and all the while you have these various DEC
> IBUs wandering around essentially rudderless on the whole issue.  "Are
> we selling hardware?  Are we selling software?  Is that OS in
> competition with us or helping us?  We don't know and the people up to
> don't know either."

Vey true.

> Perhaps the acquisition by Compaq will result in the right kind of
> shakeup in DEC's management structure, I don't know.  One can only
> hope so.
> 

Unfortunately I don't think so. All I've read over the last few days,
not all
relating to Compaq's aquisition, looks very bad for commercial Unix.
From
the press releases I've seen from Compaq the main benefit of the
aquisition is
Digital's support infrastructure. Digital had already shed large parts
of it's corporation.
The alpha chip looked doomed anyhow, the fabrication side of the
business had been sold
to Intel and Digital had already stated that they yould be supporting
the IA-64 chip in the
future. They've also sold off their networking arm as well. It looked to
me at the time like
they were shedding parts of the company that a prospective buyer didn't
want in order to make
them more attractive to said buyer.

Compaq, immediately following the press release on the aquisition
launched a marketing
campaign for E2000, their enterprise range that is specifically
targetted at ousting Unix from
the corporate market. With Digital gone, that leaves Sun, HP and IBM.
Only Sun are wholly dependent
on Unix, the other two have a lot of NT interest. I don't think Sun's
future is too sound either, they're hardware
is vastly more expensive than NT hardware and these days the quality
isn't significantly better. You'll find a lot of the
parts inside a Sun workstation are the same as in a NT box. Solaris
sucks, I actually get more done on my NT
box than I do under Solaris!

I think the upper management of Digital had given up trying to compete
and for the last
12 months have been looking to cash in their assets. At the end of the
day these people are
driven by share price and for the shareholders this looks a good deal.
Digital may cease to exist
in any real sense but the Digital shares get converted to 0.945 Compaq
shares (which says a lot about comparitive value of the companies) and
the shareholders will be happy with that so therefore in business terms
upper management are doing a good job. Business rarely makes
technological sense, we techies would much rather see the best processor
there is on every desktop running Unix but the business people don't
care about the technical issues and sometimes cashing in the chips makes
more business sense. I think this is the option Digital have taken.

Compaq aren't interested long-term in the alpha chip or unix, what they
were after was the technical support infrastructure, I suspect that in
the mid to long term they will wind down the Unix side of the company in
favour of NT (Digital itself was already moving this way in any case)
and the alpha's future already looked bleak since it didn't look like it
was going to be around after the IA-64 and the way things have panned
out it looks like this was intentional. I think Digital sued Intel with
the intention of clearing up loose ends before a sale and they had every
intention of passing responsibilty for the alpha on to Intel so that the
fabrication part of the company was out of the way for the Compaq
negotiations to continue.

One of the press releases I saw said that this deal had been on and off
for many years and the reason Compaq had prevaricated was because they
weren't interested in the bits that Digital eventually sold off.

In other reports (all my reports are from CNET by the way, get lots of
useful news if you subscrib to their daily
mailings) it looks like Unix sales are in very serious decline. NT
outsold unix by a huge margin last year (I forget the figures but you
could find them at CNET I guess). It's not really a question of whether
NT can do the job or not, it's a question of
business. I was at Elsevier Science when they signed a five year deal
with Microsoft. From what I could see it didn't have
a lot to with the technical issues and had a lot more to do with share
price of the respective companies.

Back to the original point which has been totally lost. Is there any
point to an Alpha port of FreeBSD? As with all things
within FreeBSD, that's dependant upon the enthusiasm of the volunteers,
very little gets done in FreeBSD because of a commercial need, whether
the alpha's going to be an ongoing success isn't really relevant. If
people are enthusiastic about porting FreeBSD to the alpha then it
should go ahead. While this project is more successful than most it
shouldn't lose it's roots as a fun place for hackers to do things they
don't normally get a chance to do, porting to other architectures is
certainly one of those things.

Besides, increasingly I think projects like FreeBSD may have a longer
term role to play in the world of computing. The big
commercial players are going where the money takes them but there still
will be a need for solid technical architecture that is not driven by
the mass market, and for independent research platforms. FreeBSD is well
placed to fill both those roles and expanding it's market to other
platforms would be a very healthy thing to do.

Paul Richards.
Originative Solutions Ltd.




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?E191A8FCC38DD111A84900E029115B40A1F5>