List:       axp-redhat
Subject:    Maddog's talk in Dallas
From:       "Christopher B. Browne" <cbbrowne () knuth ! brownes ! org>
Date:       1997-05-03 10:42:30

Maddog's talk in Dallas on Thursday was quite good, by the way.  

The most interesting commentary was concerning the real economic value to 
Digital of Alpha Linux.  Some of it is decidedly not intuitively obvious:

a) Alpha Linux sells more *new* boxes with Digital Unix than it does with 
Linux.

b) The ideal target (from the Digital perspective) for Linux is in the 
aftermarket, allowing "obsolete" hardware to stay in use.  Systems that get 
redeployed using Linux are more valuable than systems that get crushed, which 
*reduces the cost of ownership* of the Digital UNIX systems.  Best of all, 
this takes place without Digital losing any sales dollars or profit margin.  
Making the hardware cheaper to the customers without cutting prices is "a good 
thing."

c) People have come to Digital having heard that Alpha was a good place to run 
Linux, and come away concluding they really want DU.  That's what you call 
"unsolicited sales," that are worth a mint.
--
Christopher B. Browne, cbbrowne@unicomp.net, chris_browne@sdt.com 
Web: http://www.conline.com/~cbbrowne/  SAP Basis Consultant, UNIX Guy 
Windows NT - How to make a 100 MIPS Linux workstation perform like an 8 MHz 286

List:       axp-redhat
Subject:    Re: Maddog's talk in Dallas
From:       "Maurice W. Hilarius" <maurice () ellpspace ! math ! ualberta ! ca>
Date:       1997-05-03 23:10:45

At 10:42 AM 05-03-97 -0500, you wrote:
>Maddog's talk in Dallas on Thursday was quite good, by the way.  
>
>The most interesting commentary was concerning the real economic value to 
>Digital of Alpha Linux.  Some of it is decidedly not intuitively obvious:
>
>a) Alpha Linux sells more *new* boxes with Digital Unix than it does with 
>Linux.
>
>b) The ideal target (from the Digital perspective) for Linux is in the 
>aftermarket, allowing "obsolete" hardware to stay in use.  Systems that get 
>redeployed using Linux are more valuable than systems that get crushed,
which 
>*reduces the cost of ownership* of the Digital UNIX systems.  Best of all, 
>this takes place without Digital losing any sales dollars or profit margin.  
>Making the hardware cheaper to the customers without cutting prices is "a
good 
>thing."
>
>c) People have come to Digital having heard that Alpha was a good place to
run 
>Linux, and come away concluding they really want DU.  That's what you call 
>"unsolicited sales," that are worth a mint.
>--
Intersting comments there, Chris.
The truly unfortunate thing that I deduce from this is that therefore
Digital is absolutely not interested in things like improved libraries for
Linux to be used for research purposes. As many who buy Alpha systems and
run Linux want to do exactly that (research) it is a bit frustrating.

Cheers,
	Maurice

+--------------------------------------------------------+
| Maurice Hilarius           | The Past is History       |
| President                  |  The Future is Mystery    |
| Hard Data Ltd.             |   Today is a Gift         |
| Phone/FAX (403) 456-1510   | That is why they call it  |
| email:harddata@bigfoot.com |   The Present             |
+--------------------------------------------------------+

List:       axp-redhat
Subject:    Re: Maddog's talk in Dallas
From:       "Bryan W. Headley" <bheadley () interaccess ! com>
Date:       1997-05-04 1:54:56

On May 3, 11:10pm, Maurice W. Hilarius wrote:
> Subject: Re: Maddog's talk in Dallas
> At 10:42 AM 05-03-97 -0500, you wrote:
> >Maddog's talk in Dallas on Thursday was quite good, by the way.
> >
> >The most interesting commentary was concerning the real economic value to
> >Digital of Alpha Linux.  Some of it is decidedly not intuitively obvious:
> >
> >a) Alpha Linux sells more *new* boxes with Digital Unix than it does with
> >Linux.
> >
> >b) The ideal target (from the Digital perspective) for Linux is in the
> >aftermarket, allowing "obsolete" hardware to stay in use.  Systems that get
> >redeployed using Linux are more valuable than systems that get crushed,
> which
> >*reduces the cost of ownership* of the Digital UNIX systems.  Best of all,
> >this takes place without Digital losing any sales dollars or profit margin.
> >Making the hardware cheaper to the customers without cutting prices is "a
> good
> >thing."
> >
> >c) People have come to Digital having heard that Alpha was a good place to
> run
> >Linux, and come away concluding they really want DU.  That's what you call
> >"unsolicited sales," that are worth a mint.
> >--
> Intersting comments there, Chris.
> The truly unfortunate thing that I deduce from this is that therefore
> Digital is absolutely not interested in things like improved libraries for
> Linux to be used for research purposes. As many who buy Alpha systems and
> run Linux want to do exactly that (research) it is a bit frustrating.
>
> Cheers,
> 	Maurice


Okay, but that's weird. When our company bought our Alpha's, we found out that
virtually no-one bought them with Digital UNIX; NT was the platform of choice.
If they want to say they sold more Digital UNIXs than unit Linux sales, that's
fine, but I am suspicious of all second-hand numbers, and face it, DEC doesn't
sell Linux.

The one major AXP machine where the availability of Linux saved DEC from
destroying unsaleable stock has got to be the Multia/UDB, which did not run
Digital UNIX in the first place. I'm not aware of any other model line where
the lines were drawn out so clearly.

It has already gotten to the point, IMHO, where maintaining proprietary,
license-encumbered UNIX ports no longer makes sense. Hell, DEC could go the
mklinux route using their own Mach port from DEC UNIX, and just stick on top
the linux server! How stable do you think their underlying Mach is by now?

-- 
Bryan W. Headley                    Home:   bheadley@interaccess.com
211 East Pineview Drive             Office: bwh@abnamro.com
Round Lake Park, IL 60073           Phone:  847 740-1446

List:       axp-redhat
Subject:    Re: Maddog's talk in Dallas
From:       Martin Radford <M.P.Radford () exeter ! ac ! uk>
Date:       1997-05-04 15:42:19

>
>It has already gotten to the point, IMHO, where maintaining proprietary,
>license-encumbered UNIX ports no longer makes sense. Hell, DEC could go the
>mklinux route using their own Mach port from DEC UNIX, and just stick on top
>the linux server! How stable do you think their underlying Mach is by now?
>

I think that DU, as well as a lot of (most?) other "proprietary,
license-encumbered UNIX ports" have features which Linux doesn't have,
and won't have for a least a few years.  Certainly this will be less
and less of an issue as time goes by, but at the moment I can't see
any major Unix OS producer ditching their product for Linux.

Martin
-- 
Martin Radford                       |      M.P.Radford@exeter.ac.uk

"Only wimps use tape backup: _real_ men just upload their important stuff
  on ftp, and let the rest of the world mirror it ;)"  - Linus Torvalds


List:       axp-redhat
Subject:    Re: Maddog's talk in Dallas
From:       "Bryan W. Headley" <bheadley () interaccess ! com>
Date:       1997-05-04 20:37:36

On May 4,  3:42pm, Martin Radford wrote:
> Subject: Re: Maddog's talk in Dallas
> >
> >It has already gotten to the point, IMHO, where maintaining proprietary,
> >license-encumbered UNIX ports no longer makes sense. Hell, DEC could go the
> >mklinux route using their own Mach port from DEC UNIX, and just stick on top
> >the linux server! How stable do you think their underlying Mach is by now?
> >
>
> I think that DU, as well as a lot of (most?) other "proprietary,
> license-encumbered UNIX ports" have features which Linux doesn't have,
> and won't have for a least a few years.  Certainly this will be less
> and less of an issue as time goes by, but at the moment I can't see
> any major Unix OS producer ditching their product for Linux.
>
> Martin
> --
> Martin Radford                       |      M.P.Radford@exeter.ac.uk
>
> "Only wimps use tape backup: _real_ men just upload their important stuff
>   on ftp, and let the rest of the world mirror it ;)"  - Linus Torvalds
>
> --
> To unsubscribe: send e-mail to axp-list-request@redhat.com with
> 'unsubscribe' as the subject.  Do not send it to axp-list@redhat.com
>
>-- End of excerpt from Martin Radford

Outside of solaris, most unices are fairly stagnant. DU's major advantage
are it's better tuned libraries and compiler, neither of which is "unix" per
se.




-- 
Bryan W. Headley                    Home:   bheadley@interaccess.com
211 East Pineview Drive             Office: bwh@abnamro.com
Round Lake Park, IL 60073           Phone:  847 740-1446

List:       axp-redhat
Subject:    Re: Maddog's talk in Dallas
From:       Eric Smith < eric () brouhaha ! com>
Date:       1997-05-05 8:50:36

"Bryan W. Headley" <bheadley@interaccess.com>
> Outside of solaris, most unices are fairly stagnant.

And inside Solaris, it's too dark to see.  :-)

Software being stagnant is sometimes good.  New releases of Solaris break
far too many things.  At my last job we had lots of machines running various
versions of Solaris, and it was a pain supporting them all.  I would really
have loved to upgrade them all to SunOS 4.1.3, or better yet to NetBSD.

Cheers,
Eric

List:       axp-redhat
Subject:    Re: Maddog's talk in Dallas
From:       "Bryan W. Headley" <bheadley () interaccess ! com>
Date:       1997-05-05 11:15:42

On May 5,  8:50am, Eric Smith wrote:
> Subject: Re: Maddog's talk in Dallas
> "Bryan W. Headley" <bheadley@interaccess.com>
> > Outside of solaris, most unices are fairly stagnant.
>
> And inside Solaris, it's too dark to see.  :-)
>
> Software being stagnant is sometimes good.  New releases of Solaris break
> far too many things.  At my last job we had lots of machines running various
> versions of Solaris, and it was a pain supporting them all.  I would really
> have loved to upgrade them all to SunOS 4.1.3, or better yet to NetBSD.

And now I hear the puppy's very stable. Must be a QA nightmare. But at least
the state of the art was pushed along. Hypothetical: would anyone do VM for
UNIX at this time (you have to assume BSD boys didn't do the work a few years
ago).
I'm guessing not, due to fear, lethargy, comparison of features versus NT, etc.

-- 
Bryan W. Headley                    Home:   bheadley@interaccess.com
211 East Pineview Drive             Office: bwh@abnamro.com
Round Lake Park, IL 60073           Phone:  847 740-1446

List:       axp-redhat
Subject:    Re: Maddog's talk in Dallas
From:       "Jon 'maddog' Hall, USG Senior Leader" < hall () zk3 ! dec ! com>
Date:       1997-05-05 13:50:45


bheadley@interaccess.com said:
> Hypothetical: would anyone do VM for UNIX at this time (you have to 
> assume BSD boys didn't do the work a few years ago). I'm guessing 
> not, due to fear, lethargy, comparison of features versus NT, etc. 

Hmmm....as a commercial vendor, if I found out that my competitors would
have a feature which would allow them to run server systems much, much more
efficiently than a swapping system, would I develop it?  I think so.

And it is probably why Digital (and other) companies are spending so much time
and energy in NUMA development for UNIX today.

The real question would be if there was only ONE operating system in the world,
and if it came out of a commercial environment, THEN would that vendor
actively implement new technologies?  Perhaps not.

md
-- 
=============================================================================
Jon "maddog" Hall                     Internet: maddog@zk3.dec.com
Senior Leader, UNIX Software Group    Executive Director, Linux International

Digital Equipment Corporation         Linux International
Mailstop ZK03-2/U15                   80 Amherst St.
110 Spit Brook Rd.                    Amherst, N.H. 03031-3032 U.S.A.
Nashua, N.H. 03062-2698 U.S.A.
WWW: http://www.unix.digital.com      WWW: http://www.li.org
Voice: +1.603.881.1341                Voice: +1.603.672.4557
FAX: +1.603.881.6424

List:       axp-redhat
Subject:    Re: Maddog's talk in Dallas
From:       Paul Anderson <andersop () agapesystems ! com>
Date:       1997-05-06 7:02:18

Bryan W. Headley wrote:

> Okay, but that's weird. When our company bought our Alpha's, we found out that
> virtually no-one bought them with Digital UNIX; NT was the platform of choice.
> If they want to say they sold more Digital UNIXs than unit Linux sales, that's
> fine, but I am suspicious of all second-hand numbers, and face it, DEC doesn't
> sell Linux.
> 
> The one major AXP machine where the availability of Linux saved DEC from
> destroying unsaleable stock has got to be the Multia/UDB, which did not run
> Digital UNIX in the first place. I'm not aware of any other model line where
> the lines were drawn out so clearly.
> 
> It has already gotten to the point, IMHO, where maintaining proprietary,
> license-encumbered UNIX ports no longer makes sense. Hell, DEC could go the
> mklinux route using their own Mach port from DEC UNIX, and just stick on top
> the linux server! How stable do you think their underlying Mach is by now?

Let's really have some fun in this area.  I work as a consdultant for
Digital and have a lot of knowledge of where they are going.  I know
that Maddog is supposed to be the top of the Unix heap, and he is.  The
fact is what was once a mountain is now a molehill.  Digital has sold
nearly everyone of its major Unix products to third party vendors and
has lost key partnerships with other third party vendors, because it
refuses to pay to see the stuff ported to the Alpha.  Examples: 
CohesionWorxs was sold to Bentley and Digital Tools no longer has any
products that run on Digital machines.  Unix and VMS are dying at DEC
which is why they even allow the Linux trade.  If they were a truly
viable Unix platform they would not devote time and resources to this
effort.  Digital is banking on NT, not Unix, to take it to its future.

I should have fun answering the mail from this one.

Paul Anderson
-- 
"Good morning, doctors.  I have taken the liberty of removing
Windows 95 from my hard drive."

Arthur C Clarke, author of 2001:A Space Odyssey, on the first words
he would have liked HAL, his fictional computer, to speak

List:       axp-redhat
Subject:    Re: Maddog's talk in Dallas
From:       "Jon 'maddog' Hall, USG Senior Leader" <hall () zk3 ! dec ! com>
Date:       1997-05-06 10:31:32

Paul,

>Let's really have some fun in this area.  I work as a consdultant for
>Digital and have a lot of knowledge of where they are going.  I know
>that Maddog is supposed to be the top of the Unix heap, and he is.  The
>fact is what was once a mountain is now a molehill.  Digital has sold
>nearly everyone of its major Unix products to third party vendors and
>has lost key partnerships with other third party vendors, because it
>refuses to pay to see the stuff ported to the Alpha.  Examples: 
>CohesionWorxs was sold to Bentley and Digital Tools no longer has any
>products that run on Digital machines.  Unix and VMS are dying at DEC
>which is why they even allow the Linux trade.  If they were a truly
>viable Unix platform they would not devote time and resources to this
>effort.  Digital is banking on NT, not Unix, to take it to its future.

I will not pretend that there are not people inside of Digital that believe
WNT is the future of Digital.  Nevertheless there are also people inside of
Digital that believe UNIX will be around a long time and that we should capture
a significant portion of that market as well.

Unfortunately for the UNIX market, a lot of people believe that Microsoft
products will be the desktop of the future, with UNIX systems being reserved
for "large servers".  I think that this thought process is shared by a lot
of UNIX vendors, not just Digital.  They look at the growth market of the
UNIX desktop (6-8%) versus the growth market of the WNT desktop (est. at
20-30%) and they make their decisions on that.  If you ask them if
they are going to abandon that 6-8% UNIX growth, the answer is probably "no".
If you ask them where they will make a major investment, it will probably be
WNT (if they are telling the truth).

My personal feeling is that if the UNIX vendors desert the desktop,
that UNIX is doomed.  Most ISVs will not buy a server system to do software
development, and users will not enjoy the use of UNIX tools on their
invisible server.  Without application development on UNIX workstations,
applications will disappear, and UNIX servers will be doomed.  Java will
help, but I think it will be a long time before CAD applications and databases
are written in Java.

I think that part of the reason for UNIX's troubles these days is reflected by
going to a university and seeing Microsoft and Apple products on the desks of
99% of the undergraduate student population. Linux can help to stem that tide,
and put UNIX literate graduates back into the marketplace.

Your point that Digital has sold "most of its UNIX products" was not based
on "UNIX or not UNIX", but on the fact that Digital decided that we were into
too many products.  We could not stretch our resources to continue to do
desktop application development (DECwrite, DECchart, DECpresent) and others.
We sold those products off for a variety of both logistical and financial
reasons.

I was not privy to our sale of our middleware, but I think that part of it
was probably spawned by the inability of the middleware people to sell to
Digital's competitors a product that was absolutely controlled by Digital.
In order to be successful, middleware must be across EVERY platform, and
as long as it was controlled by Digital, people would not trust us to support
all platforms equally.

Digital UNIX is not dying.  Our sales increase quarter over quarter
faster than the general UNIX market (which means that Digital is getting
a larger share of the UNIX market).  Do we ship as many units as Sun?  No,
but if this were the criteria for continuing the UNIX business, we would have
stopped 15 years ago.

The reason why Digital "allows the Linux project" has been stated over and
over again:

	o it gives visibility of Digital to new markets
	o it helps grow interest in the UNIX market
	o it provides a prototyping arena for new functionality
	o it gives a low-cost platform for education/research

and many more reasons.  It is a very small investment (by Digital standards)
for potentially a high payback.

I do not want to start this mailing list to be a discussion of Digital's
business practices, but I felt that I had to answer you.  While you may
"work as a consdultant for Digital", you are not privy to all of the internal
workings and thought processes of our company.  Neither am I, but I tend to
think that I have a better handle on the tempo.

Sincerely,

md
-- 
=============================================================================
Jon "maddog" Hall                     Internet: maddog@zk3.dec.com
Senior Leader, UNIX Software Group    Executive Director, Linux International

Digital Equipment Corporation         Linux International
Mailstop ZK03-2/U15                   80 Amherst St.
110 Spit Brook Rd.                    Amherst, N.H. 03031-3032 U.S.A.
Nashua, N.H. 03062-2698 U.S.A.
WWW: http://www.unix.digital.com      WWW: http://www.li.org
Voice: +1.603.881.1341                Voice: +1.603.672.4557
FAX: +1.603.881.6424

List:       axp-redhat
Subject:    Maddog's talk in Dallas OR It is interesting to see what I said :-)
From:       "Jon 'maddog' Hall, USG Senior Leader" <hall () zk3 ! dec ! com>
Date:       1997-05-05 13:36:51

Hi,

I would like to clarify some of the things that I said in the *2.5 hour* :-)
talk in Dallas:

cbbrowne@knuth.brownes.org said:
>a) Alpha Linux sells more *new* boxes with Digital Unix than it does with 
>Linux.

Currently Digital has no system that ships without a license.  We ship our
system boxes with a license for either WNT, Digital UNIX or OpenVMS.  The UDB
was a special case.  In fact, the reason its name was changed from "Multia"
to "UDB" was a reflection of that decision, since "Multia" described both
the hardware and the software that originally shipped with it.

Our single-board computers which are made by Digital Semiconductor, and are
used in a series of OEM systems, are sold without an operating system.

Therefore it is impossible for us to determine how many Alpha Linux systems
are out there, since any system could be used for Linux.

What I thought I said (and what I think that Christopher meant to say) was that
having Linux on the Alpha has actually caused MORE sales of Digital UNIX than
we would have without Linux being on the Alpha. Despite the fact that
some people who *might* have bought Digital UNIX bought Linux instead, I have
had MORE people "discover" Alpha through Linux, then find out that (for one
reason or the other) Digital UNIX was a better choice *for* *their*
*particular* *need*.

By the way, the fact that Digital currently sells no systems without a license
is constantly in debate, and may change in the future.

cbbrowne@knuth.brownes.org said:
> b) The ideal target (from the Digital perspective) for Linux is in 
> the  aftermarket, allowing "obsolete" hardware to stay in use.  
> Systems that get  redeployed using Linux are more valuable than 
> systems that get crushed, which  *reduces the cost of ownership* of 
> the Digital UNIX systems.  Best of all,  this takes place without 
> Digital losing any sales dollars or profit margin.   Making the 
> hardware cheaper to the customers without cutting prices is "a good  
> thing." 

In my talk, I pointed out that the cost of ownership of a system is based upon
several things, not the least of which is the resale value of older equipment.
Since the port of Linux, the older Alphas have had a new lease on life, and
this helps customers either get more money for their older systems, or extend
the useful life of the Alpha systems inside their own organization.  Likewise
from time to time we have overruns, and the Linux market is a place where
people might like to take advantage of this situation.

However, to say that it is "the ideal target (from the Digital perspective)"
is not exactly what I wished to convey.  I think that I stressed in this
talk (as I have in others) that I believe turnkey systems (kiosks, reservation
systems, Inter/Intranet services, doctor/dentist/lawyer systems) as well
as education/research are the major markets while the number of desktop
applications builds.  Digital is trying to help with the latter, and that
is why we did the em86 emulator, and why we are glad to see the Alpha
port of the Java language and system proceeding.

cbbrowne@knuth.brownes.org said:
> c) People have come to Digital having heard that Alpha was a good 
> place to run  Linux, and come away concluding they really want DU.  
> That's what you call  "unsolicited sales," that are worth a mint.

Absolutely, but I also said that I am just as happy to have sold an Alpha
running Linux as I am to have sold an Alpha running Digital UNIX.

maurice@ellpspace.math.ualberta.ca said:
> The truly unfortunate thing that I deduce from this is that therefore 
> Digital is absolutely not interested in things like improved 
> libraries for Linux to be used for research purposes. As many who buy 
> Alpha systems and run Linux want to do exactly that (research) it is 
> a bit frustrating. 

The "truly unfortunate thing" is that Digital is made up of many independent
profit and loss centers that own various intellectual properties.  Not all of
them have bought into the Linux marketplace and (quite frankly) may never do so
since there is no profit *for them* in the marketplace for giving away their
technologies.

If, for example, the libraries that Maurice refers to were shifted under the
auspices of Digital Semiconductor (who makes and sells the Alpha chips), we
might find that the libraries would be made more available.

On the other hand, one of the reasons why I am interested in Linux is the
definition and production of libraries to do just what Maurice suggests, and
I would encourage the Linux community to develop libraries even better than
the ones that ship with Digital UNIX.

bheadley@interaccess.com said:
> Okay, but that's weird. When our company bought our Alpha's, we found 
> out that virtually no-one bought them with Digital UNIX; NT was the 
> platform of choice. If they want to say they sold more Digital UNIXs 
> than unit Linux sales, that's fine, but I am suspicious of all 
> second-hand numbers, and face it, DEC doesn't sell Linux. 

We sell a significant number of Digital UNIX systems.  While company policy
states that I can not tell you numbers, I can tell you that we sell more
Digital UNIX Alpha systems than we do either OpenVMS or WNT.  And before you
sneer at the number of OpenVMS systems we sell, remember that we built the
company on OpenVMS.  We still have a huge and loyal customer base for that
product.

khansen@njcc.com said:
> but DEC was never interested in selling CD-ROMs (or systems) with an 
> OS they didn't write. When the UDB/Multia was new DEC advertised it 
> as a low-cost Linux/AXP platfor

In my talk I stated that the current plans for Alpha Linux do not have Digital
selling Linux on their systems for two main reasons:

	o which Linux distribution would we choose?
	o when would we be able to put it out?

The first topic (although readers of this list might say "Red Hat, of course)
is one that I have dealt with since the beginning.  I did not want to
discourage any distribution by having Digital ship one particular distribution.

Or worse yet, bring out our own distribution, which touches on point two.
First of all, if we brought out our own distribution, we would probably be
bringing out kernel 1.2.13 when everyone else was on kernel 2.5.23  This is
not a hit at our engineering staff, but is just the facts of (large) corporate
life.  Secondly if we brought out our own distribution, it might discourage
others from doing their distribution, and from my observations, Linux people
purchase Linux because they LIKE a PARTICULAR distribution.  Slackware
people defend Slackware to the hilt.  Debian people keep criticizing me because
Digital does not mention Debian enough (they will probably find Digital
mentioning it more when there is an Alpha Debian distribution ready and
stable).

At the present time, I prefer our current strategy (which I helped form):

	o help the Linux community to do the port
	o rely on the existing distributions to port to Alpha
	o rely on resellers/sales people to match the distribution to the
	  customers' needs

M.P.Radford@exeter.ac.uk said:
> I think that DU, as well as a lot of (most?) other "proprietary, 
> license-encumbered UNIX ports" have features which Linux doesn't 
> have, and won't have for a least a few years.  Certainly this will be 
> less and less of an issue as time goes by, but at the moment I can't 
> see any major Unix OS producer ditching their product for Linux. 

Technical issues aside, DU has over 5,000 commercial applications that have
been ported to it.  People need these applications, and until they are
ported to Alpha Linux, they will need (and buy) Digital UNIX.

And before you say that Digital should "encourage" these vendors to port, I
will tell you that 15 years of commercial operating system development and
deployment shows that only works in a very few cases.  The only thing that
works (and works well) is that the application vendor CUSTOMERS demand a
Linux port, then support that demand with buying Linux copies.  Other methods
of doing this lead do disaster over time.  Trust me, I have the whip marks
on my back over this one.

eric@brouhaha.com said:
> > Outside of solaris, most unices are fairly stagnant.

> And inside Solaris, it's too dark to see.  :-)

> Software being stagnant is sometimes good.  New releases of Solaris 
> break far too many things.  At my last job we had lots of machines 
> running various versions of Solaris, and it was a pain supporting 
> them all.  I would really have loved to upgrade them all to SunOS 
> 4.1.3, or better yet to NetBSD. 

Well, I might ask whether "most unices are stagnant" as opposed to WHAT?

There are some operating systems that are busy trying to catch up to what
UNIX systems have today, but are doing it in a proprietary way (so you
know that I am not talking about Linux).  Do you call these systems
"dynamic" or simply lacking in features?

In another talk I give, I point out that the Whirlwind project was the
first computer to have "real-time", timesharing, light-pens and a series of
other features that most people consider "modern", and it was developed in
the late 1940s to early 1950s (Jay Forrester and Ken Olsen at MIT).  And I also
point a series of interfaces being added to UNIX under the auspices of POSIX
1003.1b that older "proprietary" systems had years ago, but which UNIX now
enjoys with standard interfaces.

Yes, Linux is moving ahead very rapidly, but the really *interesting* part will
be those *completely new* (and useful) OS functionalities which appear first
in the Linux community.....

Warmest regards to all,

md
-- 
=============================================================================
Jon "maddog" Hall                     Internet: maddog@zk3.dec.com
Senior Leader, UNIX Software Group    Executive Director, Linux International

Digital Equipment Corporation         Linux International
Mailstop ZK03-2/U15                   80 Amherst St.
110 Spit Brook Rd.                    Amherst, N.H. 03031-3032 U.S.A.
Nashua, N.H. 03062-2698 U.S.A.
WWW: http://www.unix.digital.com      WWW: http://www.li.org
Voice: +1.603.881.1341                Voice: +1.603.672.4557
FAX: +1.603.881.6424