Path: bga.com!news.sprintlink.net!hookup!news.moneng.mei.com!
howland.reston.ans.net!europa.eng.gtefsd.com!MathWorks.Com!news.duke.edu!
news-feed-1.peachnet.edu!umn.edu!mayonews.mayo.edu!usenet
From: r...@mayo.edu (Ray Ghanbari)
Newsgroups: comp.sys.next.advocacy
Subject: Future of UnixWare
Date: 24 May 1994 14:58:32 GMT
Organization: Mayo Foundation
Lines: 24
Distribution: world
Message-ID: <2rt4mo$a55@fermat.mayo.edu>
Reply-To: r...@mayo.edu
NNTP-Posting-Host: augustus.mayo.edu
Not directly related to NeXT advocacy, but I thought this might be
interesting none the less. The latest InfoWorld has a short article
entitled "Novell shift steers Unix off desktop" (page 5, 5/23/94).
If true, it is disturbing that a big player is throwing in the towel to
Microsoft. It only adds to the perception that Unix is dead, and that we
may as well start getting used to NT. Of course, I would be most curious
to hear other peoples' opinion on the matter (Mr. Vlcek?)
Juicy tidbits to prompt discussion (quoted without permission):
"Novell Inc. has virtually scuttled its plan to push UnixWare as a
high-volume desktop OS and Windows alternative, said sources close to the
company [...] 'Novell's unified desktop strategy for UnixWare is DOA right
now,' said one source who asked not to be identified. 'Top brass is now
deciding they have to stick with their knitting and have no business on
the desktop.' [...]"
--
Ray Ghanbari
Mayo Foundation
r...@mayo.edu
Newsgroups: comp.sys.next.advocacy,comp.unix.unixware
Path: bga.com!news.sprintlink.net!hookup!swrinde!pipex!uknet!EU.net!uunet!
molly!vlcek
From: uunet!molly!vlcek (Jim Vlcek)
Subject: Re: Future of UnixWare
Message-ID: <CqFzI7.DrM@molly.uucp>
Sender: vl...@molly.uucp (James Vlcek)
Reply-To: uunet!molly!vlcek (Jim Vlcek)
Organization: The Black Box of Lowertown
References: <2rt4mo$a55@fermat.mayo.edu>
Date: Fri, 27 May 1994 03:39:42 GMT
Lines: 39
Ray Ghanbari writes in comp.sys.next.advocacy (I've crossposted to
comp.unix.unixware):
>
> Not directly related to NeXT advocacy, but I thought this might be
> interesting none the less. The latest InfoWorld has a short article
> entitled "Novell shift steers Unix off desktop" (page 5, 5/23/94).
>
> If true, it is disturbing that a big player is throwing in the towel to
> Microsoft. It only adds to the perception that Unix is dead, and that we
> may as well start getting used to NT. Of course, I would be most curious
> to hear other peoples' opinion on the matter (Mr. Vlcek?)
>
> Juicy tidbits to prompt discussion (quoted without permission):
>
> "Novell Inc. has virtually scuttled its plan to push UnixWare as a
> high-volume desktop OS and Windows alternative, said sources close to the
> company [...] 'Novell's unified desktop strategy for UnixWare is DOA right
> now,' said one source who asked not to be identified. 'Top brass is now
> deciding they have to stick with their knitting and have no business on
> the desktop.' [...]"
Jim Vlcek's opinion:
Novell has never made clear its strategy with regard to Unix. One must
wonder if they ever had one. Certainly they never knew what to do with it,
and by now they must be wondering why they bought it in the first place.
I wish I could say otherwise, but I'm no longer optimistic about any future
for UnixWare.
My apologies to the hard-working and dedicated Novell employees who do their
best on the net to support this unloved orphan of a product. May your labors
not be in vain, and may I someday have to eat these words.
--
Jim Vlcek Elements of the information superhighway:
uunet!molly!vlcek UNIX: the concrete
molly!vl...@uunet.uu.net TCP/IP: the road signs
Beautiful downtown St. Paul Windows: the fast-food joints
Path: bga.com!news.sprintlink.net!hookup!usc!elroy.jpl.nasa.gov!swrinde!
ihnp4.ucsd.edu!network.ucsd.edu!mbk
From: m...@inls1.ucsd.edu (Matt Kennel)
Newsgroups: comp.sys.next.advocacy,comp.unix.unixware
Subject: Re: Future of UnixWare
Followup-To: comp.sys.next.advocacy,comp.unix.unixware
Date: 27 May 1994 21:19:11 GMT
Organization: Institute For Nonlinear Science, UCSD
Lines: 48
Message-ID: <2s5o4f$6vc@network.ucsd.edu>
References: <CqFzI7.DrM@molly.uucp>
NNTP-Posting-Host: lyapunov.ucsd.edu
X-Newsreader: TIN [version 1.1 PL8]
Jim Vlcek (uunet!molly!vlcek) wrote:
: > "Novell Inc. has virtually scuttled its plan to push UnixWare as a
: > high-volume desktop OS and Windows alternative, said sources close to the
: > company [...] 'Novell's unified desktop strategy for UnixWare is DOA right
: > now,' said one source who asked not to be identified. 'Top brass is now
: > deciding they have to stick with their knitting and have no business on
: > the desktop.' [...]"
: Jim Vlcek's opinion:
: Novell has never made clear its strategy with regard to Unix. One must
: wonder if they ever had one. Certainly they never knew what to do with it,
: and by now they must be wondering why they bought it in the first place.
: I wish I could say otherwise, but I'm no longer optimistic about any future
: for UnixWare.
: My apologies to the hard-working and dedicated Novell employees who do their
: best on the net to support this unloved orphan of a product. May your labors
: not be in vain, and may I someday have to eat these words.
It seems that Novell is even more clueless than AT&T was. At least let
somebody who *HAS A STAKE IN IT* own it for a change. Back in the BSD glory
days the hackers had a personal interest in Unix and it showed. Microsoft
certainly has a big financial interest in their own products, and it shows.
I mean, what has USL/Novell done with it besides jam NetWare into an already
huge kernel??
The OSF is another loser. What Unix needs an OSF that actually writes real,
important, *products* (not committees or documents) for UNIX and lets
everybody use it without paying extra, just as with NFS and TCP/IP but with
higher quality. Hell, Richard Stallman and the FSF is alot more important
to most Unix users than OSF or Novell ever will be.
Pathetic. It's not the programmers fault but the mis-organization.
: --
: Jim Vlcek Elements of the information superhighway:
: uunet!molly!vlcek UNIX: the concrete
: molly!vl...@uunet.uu.net TCP/IP: the road signs
: Beautiful downtown St. Paul Windows: the fast-food joints
--
-Matt Kennel m...@inls1.ucsd.edu
-Institute for Nonlinear Science, University of California, San Diego
-*** AD: Archive for nonlinear dynamics papers & programs: FTP to
-*** lyapunov.ucsd.edu, username "anonymous".
Path: bga.com!news.sprintlink.net!news.charm.net!news.charm.net!not-for-mail
From: cnor...@charm.net (Craig Nordin)
Newsgroups: comp.sys.next.advocacy,comp.unix.unixware
Subject: Re: Future of UnixWare
Date: 28 May 1994 00:39:14 -0400
Organization: Charm.Net : Baltimore Local Internet Access, Hon
Lines: 34
Message-ID: <2s6hti$qsr@sowebo.charm.net>
References: <CqFzI7.DrM@molly.uucp> <2s5o4f$6vc@network.ucsd.edu>
NNTP-Posting-Host: sowebo.charm.net
m...@inls1.ucsd.edu (Matt Kennel) writes:
>I mean, what has USL/Novell done with it besides jam NetWare into an already
>huge kernel??
OK Bro, now that you've torn one down, put something up.
I wanna know what Unix is better for the developer and sysadmin
and user than UnixWare. I'm probably all wet behind the ears,
so school me instead of just griping. Where do I go for better Unix?
It could be that you are saying UW is a loser but everything
else sux worst. Then that is fine, because I figure I'm onto
the best one around.
Nope, it ain't NCR/AT&T GIS.
Nope, it ain't SCO
Nope, it ain't Interactive
Nope, it ain't Microport
Nope, it ain't Dell
Nope, it ain't Linux
Nope, it ain't Coherent
Is it one of the BSD Clan?
Gawd, it ain't Mach, is it?
Gawd, it ain't AIX, is it?
Gawd, it ain't HP/UX, is it?
Gawd, it ain't SGI, is it?
What the heck is it?!!!!!!!!!
--
See the Emerald on the Matrix? Baltimore, Maryland Access to the Internet
That's Charm.Net Hon! E-Mail: i...@charm.net Voice:(410) 558.3900
Path: bga.com!news.sprintlink.net!hookup!news.moneng.mei.com!
howland.reston.ans.net!cs.utexas.edu!swrinde!news.dell.com!tadpole.com!
uunet!psinntp!heimdall!axpo5.sdrc.com!crbalsn
From: bal...@world.std.com (Jim Balson)
Newsgroups: comp.unix.unixware
Subject: Re: The future of Unixware
Message-ID: <8449@heimdall.sdrc.com>
Date: 31 May 94 22:15:24 GMT
References: <CqL1ty.Bq7@amiserv.chi.il.us>
Sender: n...@heimdall.sdrc.com
Reply-To: bal...@world.std.com (Jim Balson)
Organization: OrgFreeware
Lines: 63
Well, heres my $0.02 cents worth on the future of Unixware.
***** Competing with other Unix's on the desktop:
Despite Novell's desktop non-strategy, I think as far as Unix OS's
on the dektop, Unixware will come out the winner becasue of the pricing. But,
even though I think the pricing is great, it's still wrong. Novell should
provide a pricing structure such that a complete Unix runtime system is
available for roughly $100. Then the price of the SDK should increase to the
200-300 range. Thats my opinion. With a complete Unix system for $100, I think
you may see more trying it out at that price point.
***** Competing with other OS's on the desktop:
Unixware still needs some work. Namely in the area of graphical system
administration. Until a user can do absolutely everything from the desktop
and require no intricate knowledge of the Unix OS itself, it has no hope
of ever competing head to head against OS2, NT, Apple or "Chicago".
Now if Novell were to put such a GUi front end on the Sys Admin area and
make absolutely everything dooable from the desktop, I think Unixware can
compete. They may not get the market share of MS or Apple, but they will
be a viable alternative.
Now if Novell really has no plans or desire to compete on the desktop,
then they certainly don't have any plans on putting any GUI front end on top
of the sys admin. And if they don't want to do that, I have to wonder how much
desire they have to make the desktop they have now any better than it is.
***** A comment on Apple:
From what I've heard about the Apple OS, it's ugly. If that is indeed
true, look at how far Apple took that OS. I have not ever used a Apple PC. But
I feel that if this is indeed true, I feel Unix can do the same. Actually, I
feel that if Unixware had a complete UI such as the MAC, it would blow away
the MAC in market share.
***** Conclusion
Sigh. I guess Novell is not up to the task of taking Unixware to
the desktop. Which leads me to believe that the desktop will languish where
it is now without being enhanced and upgraded the way all good UI/Desktops
should. I'm a supporter of Unixware because of the desktop and the pricing.
But should Unixware lose the edge on the desktop and pricing, I'll jump ship
in a heartbeat. Like say perhaps Sun lowers it pricing structure to somewhere
in the ballpark of Unixware.
Just my current thoughts.
Jim
bal...@world.std.com
Newsgroups: comp.unix.unixware
Path: bga.com!news.sprintlink.net!news.onramp.net!convex!cs.utexas.edu!
swrinde!news.dell.com!tadpole.com!uunet!sparky!dsndata!backbone!backbone!wayne
From: wa...@backbone.uucp (Wayne Schlitt)
Subject: Re: The future of Unixware
In-Reply-To: balson@world.std.com's message of 31 May 94 22: 15:24 GMT
Message-ID: <WAYNE.94Jun2090948@backbone.uucp>
Sender: wayne@backbone (Wayne Schlitt)
Reply-To: wa...@cse.unl.edu
Organization: The Backbone Cabal
References: <CqL1ty.Bq7@amiserv.chi.il.us> <8449@heimdall.sdrc.com>
Date: Thu, 2 Jun 1994 15:09:48 GMT
Lines: 80
In article <8...@heimdall.sdrc.com> bal...@world.std.com (Jim Balson) writes:
>
> Well, heres my $0.02 cents worth on the future of Unixware.
>
>
> ***** Competing with other Unix's on the desktop:
>
> Despite Novell's desktop non-strategy, I think as far as Unix OS's
> on the dektop, Unixware will come out the winner becasue of the pricing. But,
> even though I think the pricing is great, it's still wrong. Novell should
> provide a pricing structure such that a complete Unix runtime system is
> available for roughly $100. Then the price of the SDK should increase to the
> 200-300 range. Thats my opinion. With a complete Unix system for $100, I think
> you may see more trying it out at that price point.
Of course, with Linux you can get a complete Unix system for $40, with
an hourly support fee. This includes a SDK, networking, etc. For
that $40, you can install it on as many computers as you want, let
employees take it home, give it to your friends, etc.
Oh, sure, Linux doesn't do everything that UW does. Lots of stuff is
still basically beta software or incomplete. The support given by
Yggdrasil or Trans-Ameritech is not up to the quality of Novells.
Linux doesn't run very many commercial packages, and the ones that it
does run are mostly DOS programs under its primitive DOS Emulator.
That stuff is a given.
On the other hand, there are probably _more_ Linux users than all the
SVR4 users combined. Heck, there may well be more people running
Linux than any other version of Unix out there, including Xenix,
SunOS, SCO, etc. Yggdrasil alone sold more CD's than Novell sold
copies UW over the last 6 months, and you can bet that a vast majority
of Linux users didn't get their copy from Yggdrasil.
Linux _is_ usable for a large number of things. The number of things
that Linux is able to do well is increasing fairly rapidly. The iBCS2
support is coming along fairly well, and you can run things like
Oracle and WP for SCO. There are improvements coming in the DOS
Emulator. The installation procedures and quality of integration is
improving rapidly.
Linux has broken the chicken and egg problem of Unix. It has gotten a
large number of installations without any significant commercial
support. If it ever starts to get wide commercial support than the
number of installations will really take off.
The thing that UW now has to compete with is "Does the added
functionality of UW over Linux worth $200+ to me?" For a large number
of people, the answer is clearly "No". Novell can either greatly
increase the functionality of UW, or greatly cut the price. Neither
of these is going to be easy to do.
Oh, I guess I should mention the obligatory "Yeah, and Linux comes with
complete source too". For many people this is irrelevant, but for
many people this is worth thousands of dollars.
Any Unix business plan that doesn't include the effects of Linux on
the Unix market is at best incomplete. Novell may well be looking at
using Linux to its advantage. Let people run Linux for the masses on
the desktops, let UW run the main servers.
-wayne
--
The human brain is a complex organ with the wonderful power of
enabling man to find reasons for continuing to believe whatever it is
that he wants to believe. -Voltaire
Path: bga.com!news.sprintlink.net!hookup!swrinde!howland.reston.ans.net!
europa.eng.gtefsd.com!MathWorks.Com!news2.near.net!yale!yale.edu!xlink.net!
zib-berlin.de!zrz.TU-Berlin.DE!netmbx.de!Germany.EU.net!EU.net!uunet!
psinntp!heimdall!axpo5.sdrc.com!crbalsn
From: bal...@world.std.com (Jim Balson)
Newsgroups: comp.sys.next.advocacy,comp.unix.unixware
Subject: Re: Future of UnixWare
Message-ID: <8472@heimdall.sdrc.com>
Date: 2 Jun 94 15:53:05 GMT
References: <CqFzI7.DrM@molly.uucp> <2s5o4f$6vc@network.ucsd.edu>
<2s6hti$qsr@sowebo.charm.net> <2sfugu$8o5@hamblin.math.byu.edu>
Sender: n...@heimdall.sdrc.com
Reply-To: bal...@world.std.com (Jim Balson)
Lines: 24
: I wanna know what Unix is better for the developer and sysadmin
: and user than UnixWare. I'm probably all wet behind the ears,
: so school me instead of just griping. Where do I go for better Unix?
: It could be that you are saying UW is a loser but everything
: else sux worst. Then that is fine, because I figure I'm onto
: the best one around.
Lots of discussion about the future of Unixware, but IMHO, as far as
Unix on the desktop/Intel, Unixware will most likely have the LIONS share
of it in a year or two. It's a really nice system to use. Unfortunately, the
resources it requires are to demanding for a mainstream desktop OS. Never
mind the fact that Novell will not try to push it onto the desktop, with it's
current pricing at $350.00 for a complete runtime + development system, it
will more than likely sell itself onto the desktop.
Jim
bal...@world.std.com
Path: nntp.gmd.de!Germany.EU.net!EU.net!uunet!psinntp!heimdall!axpo5.sdrc.com!
crbalsn
From: crba...@axpo5.sdrc.com (Jim Balson)
Newsgroups: comp.unix.unixware,comp.os.linux,comp.os.linux.misc
Subject: Re: future of Unixware
Message-ID: <8565@heimdall.sdrc.com>
Date: 6 Jun 94 22:45:18 GMT
References: <CquoKE.7no@rahul.net> <8561@heimdall.sdrc.com>
Sender: n...@heimdall.sdrc.com
Reply-To: crba...@axpo5.sdrc.com (Jim Balson)
Organization: OrgFreeware
Lines: 87
>In article <8...@heimdall.sdrc.com> bal...@world.std.com (Jim Balson) writes:
>>
>> Well, heres my $0.02 cents worth on the future of Unixware.
>>
>>
>> ***** Competing with other Unix's on the desktop:
>>
>> Despite Novell's desktop non-strategy, I think as far as Unix OS's
>> on the dektop, Unixware will come out the winner becasue of the pricing. But,
>> even though I think the pricing is great, it's still wrong. Novell should
>> provide a pricing structure such that a complete Unix runtime system is
>> available for roughly $100. Then the price of the SDK should increase to the
>> 200-300 range. Thats my opinion. With a complete Unix system for $100, I think
>> you may see more trying it out at that price point.
>
>Of course, with Linux you can get a complete Unix system for $40, with
>an hourly support fee. This includes a SDK, networking, etc. For
>that $40, you can install it on as many computers as you want, let
>employees take it home, give it to your friends, etc.
>
>
>Oh, sure, Linux doesn't do everything that UW does. Lots of stuff is
>still basically beta software or incomplete. The support given by
>Yggdrasil or Trans-Ameritech is not up to the quality of Novells.
>Linux doesn't run very many commercial packages, and the ones that it
>does run are mostly DOS programs under its primitive DOS Emulator.
>
>That stuff is a given.
[Nice sales pitch]
Don't waste your breath trying to sell me on Linux, I have already used
Esix, which offers everything Linux is striving for, but with an SVR4 kernel.
After seeing Unixware, I can never go back to using Esix or anything similar.
Besides, we were discussing Unixware, not Linux.
But, Wayne, Linux is not complete at $40.00. In case you haven't
noticed, Motif is missing. Linux System Labs will sell you Motif for, ah,
$175.00. plus the $40.00 you claim you can get Linux for , and thats a
whopping $215.00 for Linux (shipping not included). With Unixware at $280.00,
I know which I would rather have!
>The thing that UW now has to compete with is "Does the added
>functionality of UW over Linux worth $200+ to me?" For a large number
>of people, the answer is clearly "No". Novell can either greatly
>increase the functionality of UW, or greatly cut the price. Neither
>of these is going to be easy to do.
[Since my last post, I took the time to price Motif for Linux]
Ok, re-doing my super complex chart here, I see that:
Linux Unixware
===== ========
$ 40.00 OS $190.00 (OS)
$175.00 Motif (run + dev) $ 89.00 (SDK)
------- -------
$215.00 $277.00
So, with a price difference of exactly $62.00, I think the real
question that should be asked is "Does saving $62.00 really justify me
not using Unixware?". :-) Your last hope in arguing that Linux is more cost
effective than Unixware is somehow obtaining a royalty free Motif from the
OSF or writing your own Motif. Neither of these is going to be easy to do. :-)
>
>On the other hand, there are probably _more_ Linux users than all the
>SVR4 users combined. Heck, there may well be more people running
>Linux than any other version of Unix out there, including Xenix,
>SunOS, SCO, etc. Yggdrasil alone sold more CD's than Novell sold
>copies UW over the last 6 months, and you can bet that a vast majority
>of Linux users didn't get their copy from Yggdrasil.
I prefer to argue about facts I can back up one way or thge other.
Don't make claims such as this without proof.
Jim
bal...@world.std.com
Newsgroups: comp.unix.unixware,comp.os.linux,comp.os.linux.misc
Path: nntp.gmd.de!Germany.EU.net!EU.net!uknet!cf-cm!cybaswan!iiitac
From: iii...@uk.ac.swan.pyr (Alan Cox)
Subject: Re: future of Unixware
Message-ID: <1994Jun7.170859.23563@uk.ac.swan.pyr>
Organization: Swansea University College
References: <CquoKE.7no@rahul.net> <8561@heimdall.sdrc.com>
<8565@heimdall.sdrc.com>
Date: Tue, 7 Jun 1994 17:08:59 GMT
Lines: 40
In article <8...@heimdall.sdrc.com> crba...@axpo5.sdrc.com (Jim Balson) writes:
>>Oh, sure, Linux doesn't do everything that UW does. Lots of stuff is
>>still basically beta software or incomplete. The support given by
>>Yggdrasil or Trans-Ameritech is not up to the quality of Novells.
>>Linux doesn't run very many commercial packages, and the ones that it
>>does run are mostly DOS programs under its primitive DOS Emulator.
>>
Like Oracle 6, or wordperfect for X windows - both work under Linux +
the iBCS2 module.
> Don't waste your breath trying to sell me on Linux, I have already used
>Esix, which offers everything Linux is striving for, but with an SVR4 kernel.
>After seeing Unixware, I can never go back to using Esix or anything similar.
>Besides, we were discussing Unixware, not Linux.
Excuse me I've used Esix and I take that as an insult.
> But, Wayne, Linux is not complete at $40.00. In case you haven't
>noticed, Motif is missing. Linux System Labs will sell you Motif for, ah,
Sorry Motif ?? - oh yes how to waste gigabytes of memory I nearly forgot some
people use it. I've yet to find a motif application that uses shared libraries
and isnt static linked so Im not upset. Anyway the only useful motif application
is Mosaic !
>not using Unixware?". :-) Your last hope in arguing that Linux is more cost
>effective than Unixware is somehow obtaining a royalty free Motif from the
>OSF or writing your own Motif. Neither of these is going to be easy to do. :-)
There is a project underway to write a free Motif clone. There are loads of
alternative Motif libraries, and because motif itself is such a pain many
people use other kits like Tcl/Tk, or Object builder. I can get both of those
for Linux.
There are also projects underway to write a free windows emulator, to write
a free Novell server for Linux and a free Novell client for Linux.
Alan
Path: nntp.gmd.de!Germany.EU.net!EU.net!howland.reston.ans.net!
vixen.cso.uiuc.edu!newsrelay.iastate.edu!hobbes.physics.uiowa.edu!
newsfeed.ksu.ksu.edu!moe.ksu.ksu.edu!crcnis1.unl.edu!backbone!backbone!wayne
From: wa...@backbone.uucp (Wayne Schlitt)
Newsgroups: comp.unix.unixware,comp.os.linux,comp.os.linux.misc
Subject: Re: future of Unixware
Date: Wed, 8 Jun 1994 00:48:07 GMT
Organization: The Backbone Cabal
Lines: 36
Sender: wayne@backbone (Wayne Schlitt)
Message-ID: <WAYNE.94Jun7184807@backbone.uucp>
References: <CquoKE.7no@rahul.net> <8561@heimdall.sdrc.com>
<8565@heimdall.sdrc.com>
<1994Jun7.170859.23563@uk.ac.swan.pyr>
Reply-To: wa...@cse.unl.edu
NNTP-Posting-Host: cse.unl.edu
In-Reply-To: iiitac@uk.ac.swan.pyr's message of Tue, 7 Jun 1994 17: 08:59 GMT
In article <1994Jun7.170859.23...@uk.ac.swan.pyr> iii...@uk.ac.swan.pyr
(Alan Cox) writes:
> In article <8...@heimdall.sdrc.com> crba...@axpo5.sdrc.com (Jim Balson)
writes:
> >>Linux doesn't run very many commercial packages, and the ones that it
> >>does run are mostly DOS programs under its primitive DOS Emulator.
>
> Like Oracle 6, or wordperfect for X windows - both work under Linux +
> the iBCS2 module.
Be careful there... I have been watching the iBCS2 mailing list for a
fair while now and I am _very_ impressed with the work those people
have done. _However_, there are a lot of Ifs, Ands and Buts with
those reports of commercial packages running. Most of them have only
been lightly tested. Many of them require you to configure them just
right or they will crash. Many of them require you to copy libraries
from otherwise unused copies of SCO or SVR4. Until things like
libnsl.so are written, you will have to buy a copy of SVR4 in order to
run stuff under Linux.
Right now, I would be very careful with claims about what Linux can
run under the iBCS2 system.
OTOH, I am very confident that many of these problems will be worked
out in the coming months. The people involved seem to be very sharp
and very knowledgable, and I really don't think they will have come
this far to let a few more "minor" details stop them. In another 4-6
months, we should be able to make much grander claims, but for now,
iBCS2 is still very beta code.
-wayne
--
The human brain is a complex organ with the wonderful power of
enabling man to find reasons for continuing to believe whatever it is
that he wants to believe. -Voltaire
Path: nntp.gmd.de!Germany.EU.net!EU.net!uunet!sun.cais.com!ericy
From: er...@cais.cais.com (Eric Youngdale)
Newsgroups: comp.unix.unixware,comp.os.linux.misc
Subject: Re: future of Unixware
Date: 8 Jun 1994 12:46:43 GMT
Organization: Capital Area Internet Service
Lines: 34
Message-ID: <2t4ejj$t2j@sun.cais.com>
References: <CquoKE.7no@rahul.net> <8565@heimdall.sdrc.com>
<1994Jun7.170859.23563@uk.ac.swan.pyr> <WAYNE.94Jun7184807@backbone.uucp>
NNTP-Posting-Host: cais.com
In article <WAYNE.94Jun7184...@backbone.uucp>,
Wayne Schlitt <wa...@cse.unl.edu> wrote:
>Be careful there... I have been watching the iBCS2 mailing list for a
>fair while now and I am _very_ impressed with the work those people
>have done. _However_, there are a lot of Ifs, Ands and Buts with
>those reports of commercial packages running. Most of them have only
>been lightly tested. Many of them require you to configure them just
>right or they will crash. Many of them require you to copy libraries
>from otherwise unused copies of SCO or SVR4. Until things like
>libnsl.so are written, you will have to buy a copy of SVR4 in order to
>run stuff under Linux.
No, no, no. A copy of libnsl from SVr4 would be worse than
useless since linux does not support TLI. The shared libraries that are
prepared for iBCS2 contain a small libnsl.so/libsocket.so which define
socket() and related functions, and it is possible to run socket based
applications using these libraries. There is no need to copy anything
from a SVr4 machine. Once the linux kernel supports streams, then TLI
support can probably be added, and libnsl.so can be filled out.
You are probably thinking of the situation with SCO, where you
currently need a copy of the libc_s and libnsl_s shared libraries to run
some applications. Many SCO applications are static linked anyways
(such as WP), so you do not even need to go this far. There are a few
technical glitches in the linux version of libc_s, and once these are
solved, you can probably get away without copying anything. Technically
Oracle requires libnsl_s, but since linux does not support TLI, a dummy
file may suffice here for the time being - don't know, haven't tried it.
-Eric
--
"The woods are lovely, dark and deep. But I have promises to keep,
And lines to code before I sleep, And lines to code before I sleep."
Path: nntp.gmd.de!newsserver.jvnc.net!howland.reston.ans.net!
math.ohio-state.edu!hobbes.physics.uiowa.edu!newsfeed.ksu.ksu.edu!
moe.ksu.ksu.edu!crcnis1.unl.edu!backbone!backbone!wayne
From: wa...@backbone.uucp (Wayne Schlitt)
Newsgroups: comp.unix.unixware,comp.os.linux,comp.os.linux.misc
Subject: Re: future of Unixware
Date: Wed, 8 Jun 1994 04:29:52 GMT
Organization: The Backbone Cabal
Lines: 110
Sender: wayne@backbone (Wayne Schlitt)
Message-ID: <WAYNE.94Jun7222952@backbone.uucp>
References: <CquoKE.7no@rahul.net> <8561@heimdall.sdrc.com>
<8565@heimdall.sdrc.com>
Reply-To: wa...@cse.unl.edu
NNTP-Posting-Host: cse.unl.edu
In-Reply-To: crbalsn@axpo5.sdrc.com's message of 6 Jun 94 22: 45:18 GMT
In article <8...@heimdall.sdrc.com> crba...@axpo5.sdrc.com (Jim Balson)
writes:
> >On the other hand, there are probably _more_ Linux users than all the
> >SVR4 users combined. Heck, there may well be more people running
> >Linux than any other version of Unix out there, including Xenix,
> >SunOS, SCO, etc. Yggdrasil alone sold more CD's than Novell sold
> >copies UW over the last 6 months, and you can bet that a vast majority
> >of Linux users didn't get their copy from Yggdrasil.
>
> I prefer to argue about facts I can back up one way or thge other.
> Don't make claims such as this without proof.
>
Well, you are right, I have no proof that more people run Linux that
SVR4. On the other hand, I doubt that you can prove me wrong. I
didn't, however, just make a random claim. I do have some basis for
it. A lot of it is circumstantial, but here it is anyway.
Right now, most of the Linux news groups have readership numbers in
the 100,000-150,000 range. Now there is some room for doubt in the
accuracy of these numbers, and not everyone who reads these groups
necessarily runs Linux. OTOH, not everyone who runs Linux reads these
groups. So, there it is fairly safe to say that at _least_
100,000-200,000 people run Linux.
The Yggdrasil people have shipped well over 15,000 CD's since last
fall, and they are claiming that the Summer release will out sell UW.
Another CD company ran a special and got around 3,000 responses within
a couple of weeks. There is currently about 4 different companies
that you can get Linux on a CD from.
Last winter the German magazine iX ran a survey and they estimated
that around 50,000 people ran Linux in Germany alone.
When discussions like this have come up before in the Linux
newsgroups, it has been clear that there is a _large_ number of people
who run Linux that have never heard of Usenet. There are lots of
BBS's out there that have copies of Linux, and a lot of people get it
by copying it from other people who already have Linux. Multiplying
the number of readers of the Linux newsgroups by a factor of 2-10 is
probably in order.
So, it seems very likely to me that there are 300,000-600,000 Linux
users, and it is possible for there to be close to a million of them
out there.
What are the numbers for SVR4? Well, Sun probably sells the most with
Solaris on the Sparc. Lets say 200,000 per year for the last 2 years.
(We will ignore all those people who have "upgraded" from Solaris to
SunOS 4.0.x, and those that have replaced previous boxes...). Add in
UnixWares, and all the other little SVR4 companies, and you probably
get 500,000 to a million.
This total probably is more than any other version of Unix, including
SCO Unix or SCO Xenix. So, my claim that Linux might be the most used
version of Unix is really about the same as my claim that there are
more Linux users than SVR4 users.
Ok, these numbers are only ball park figures, and even then they are
probably too close to call.
But, I _meant_ the claim to be provocative. I _want_ Novell and us
SVR4 people to think about this seriously. SVR4 is having a hard
enough time catching up to its predecessor (SVR3) in sales, but here
is this newcomer that is rolling along quite nicely.
Even if Linux _doesn't_ have more users today, unless Novell does
something quick, it will have within a year or two. Novell only has a
year or so to do something to make UW start selling like hot cakes,
or Linux is going to do the same thing to UW that GNU Emacs did to
Unipress Emacs.
Yeah, you remember. About 8-10 years ago there was a big debate about
whether this new, freely available version of Emacs from GNU could
_possibly_ replace Unipress Emacs. There were all the arguments about
how Unipress had a larger install base. How Unipress had professional
programmers to support the software. What would happen to all those
GNU Emacs people if Richard Stallman decided to stop working on Emacs.
How, without a sales, marketing and support department, GNU Emacs
would get nowhere. How when you consider the cost of a typical
programmer, the added cost of the editor shouldn't even be consider
and that no real business would ever use a "free" editor.
I don't know if Unipress is still selling Emacs or not. Even if they
are, they probably sell less than Lucid does of there modified and
commercialized version of GNU emacs. In any case, Unipress is now
just a bit player in the Emacs world.
I really hope that in 8-10 years, that there will be other companies
selling versions of Unix other than the ones that are based on Linux
sources. I can see it now. Warnings from people about not buying
stuff from SVR6 because it is not compatible with Linux. 1/2 :->
-wayne
--
The human brain is a complex organ with the wonderful power of
enabling man to find reasons for continuing to believe whatever it is
that he wants to believe. -Voltaire
Newsgroups: comp.unix.unixware,comp.os.linux,comp.os.linux.misc
Path: nntp.gmd.de!xlink.net!howland.reston.ans.net!gatech!
newsxfer.itd.umich.edu!zip.eecs.umich.edu!umn.edu!newsdist.tc.umn.edu!
uum1!gw.digibd.com!digibd!rick
From: r...@digibd.digibd.com (Rick Richardson)
Subject: Re: future of Unixware
Message-ID: <rick.771073998@digibd>
Sender: n...@gw.digibd.com (USENET News)
Nntp-Posting-Host: digibd.digibd.com
Organization: Digi International, Incorporated, Eden Prairie,MN
References: <CquoKE.7no@rahul.net> <8561@heimdall.sdrc.com>
<8565@heimdall.sdrc.com> <WAYNE.94Jun7222952@backbone.uucp>
Date: Wed, 8 Jun 1994 11:13:18 GMT
Lines: 36
wa...@backbone.uucp (Wayne Schlitt) writes:
>So, it seems very likely to me that there are 300,000-600,000 Linux
>users, and it is possible for there to be close to a million of them
>out there.
You missed one source of users for SVR4. Using information which
is publically available, I estimate the size of the PC multiport
serial market at $200M, or about 2 million ports/year. 70% of
that is Unix (all flavors), or 1.4 million ports/year. I don't
know what the breakdown is after that, but lets be conservative:
-SCO 80% 1,112,000 ports/year
-SVR4 20% 280,000 ports/year
-Linux 0% 0 ports/year
So there's another half million SVR4 users (over two years) out there
that you didn't count.
Also, I believe that if the Linux numbers being bandied about were
actually real installations for commercial purposes, then the multiport
serial vendors would be getting beat up for not having Linux drivers.
We do get inquiries about Linux from time to time, but not enough
at this point to cause the Sales droids to beat up on Marketing
who would then beat up Engineering. Right now, I suspect the
hottest source of interest for Linux drivers in this company
would be from our own engineering hacker types. I think this is
really indicative of the real world use of Linux today.
-Rick
--
Rick Richardson Senior Staff Engineer Visit my WWW home page:
DigiBoard APD Email: r...@digibd.com http://www.digibd.com/people/rick
6801 Shady Oak Rd. Fax: (612) 947-1129
Eden Prarie, MN 55344 Tel: (612) 947-1111 <standard disclaimer>
Path: nntp.gmd.de!xlink.net!howland.reston.ans.net!newsserver.jvnc.net!
netnews.summit.novell.com!netnews.summit.novell.com!sfwksh!pend
From: p...@usl.com (Steve Pendergrast)
Newsgroups: comp.unix.unixware,comp.os.linux,comp.os.linux.misc
Subject: Re: future of Unixware
Followup-To: comp.unix.unixware,comp.os.linux,comp.os.linux.misc
Date: 8 Jun 1994 18:04:09 GMT
Organization: Novell
Lines: 108
Message-ID: <2t516p$6vl@bird.summit.novell.com>
References: <WAYNE.94Jun7222952@backbone.uucp>
NNTP-Posting-Host: sfwksh.summit.novell.com
X-Newsreader: TIN [version 1.1 PL9]
Sorry about the last accidental repost of Wayne's post. Here is
my intended post.
Wayne Schlitt (wa...@backbone.uucp) wrote:
: The Yggdrasil people have shipped well over 15,000 CD's since last
: fall, and they are claiming that the Summer release will out sell UW.
Oh, that's a very unbiased way to assess things, ask the people who
sell a product whether it will be successful!
In 1990 most analysts claimed OSF/1 would overtake and bury SRV4 by 1991.
: Another CD company ran a special and got around 3,000 responses within
: a couple of weeks. ...
Responses or sales? Marketing theory says that for every 100 responses
you get to an advertisement you might make 5 sales if you're lucky.
: ... There is currently about 4 different companies
: that you can get Linux on a CD from.
And I'd bet all 4 versions are different. What does the number of
companies have to do with it? There's only one company making MS/Windows
CD's so I guess there must be more Linux than MS/Windows...nice logic.
: Last winter the German magazine iX ran a survey and they estimated
: that around 50,000 people ran Linux in Germany alone.
That's something like one out of every 350 adult Germans running Linux.
Sounds fishy to me, what was their method of extrapolating those
figures? I doubt magazines use much in the way of scientific
polling methods. What kind of magazine is iX anyway? Is it highly
speculative? Is it run by free software proponents, or is it unbiased?
I can find some magazine somewhere on earth to say just about anything
I want... (Bet I can find one somewhere that gives the number of
space aliens using Linux).
: <rampant mathematical speculation snipped>
: So, it seems very likely to me that there are 300,000-600,000 Linux
: users, and it is possible for there to be close to a million of them
: out there.
Seems unlikely to me, but keep reading and you'll see you're still
way off for other reasons...
: What are the numbers for SVR4? Well, Sun probably sells the most with
: Solaris on the Sparc. Lets say 200,000 per year for the last 2 years.
: (We will ignore all those people who have "upgraded" from Solaris to
: SunOS 4.0.x, and those that have replaced previous boxes...). Add in
: UnixWares, and all the other little SVR4 companies, and you probably
: get 500,000 to a million.
I find it fascinating that you take into account "upgrades" in the case
of SVR4, while above you assume that every CD that gets sold by those
4 companies is a new sale of Linux. Seems to me that an operating system
that gets rev'ed once a month would have a very high percentage of
"upgrades" too, no? In fact, it could be that most of the Linux CD's
are upgrades, couldn't it?
The actual number of active SVR4 installations (yes, it takes into
account retired machines and upgrades) is at the high end of your
estimate (according to IDC research), a little over 1 million.
However, the average number of users per system is something like 5
in the case of SVR4, making the number of users on the order of 5
million. I'd bet 95% of the Linux installations are single user, but
I admit I have no figures to back that up.
: This total probably is more than any other version of Unix, including
: SCO Unix or SCO Xenix. So, my claim that Linux might be the most used
Actually, SCO Unix is a mighty big fraction of the total unix market right
now. But SCO isn't SVR4 anyway so it's irrelavant to the current discussion.
: version of Unix is really about the same as my claim that there are
: more Linux users than SVR4 users.
Well, I don't know about hackers or academia, but a recent IDC survey
(that was conducted using a reasonably scientific methodology, by the
way) found that only 3% of all corporations were running Linux in 1993,
and there was virtually no future planned introduction
of Linux into others in 1994 (I think it was a percent or so). By contrast,
UnixWare was already in 5% of the corporate comp centers (even though
that was only the first year of the product), and that was expected to double
in 1994. Solaris on SPARC was in something like 75% of the comp centers,
and that was expected to grow a few percent in 1994. So, you're looking
at something like 25 to 1 in favor of SVR4 over Linux in the corporation in
1993, with the balance actually widening in favor of SVR4 in 1994.
: Ok, these numbers are only ball park figures, and even then they are
: probably too close to call.
NOT! All you can claim is that there might be a few hundred thousand
people who might be playing with Linux, but no hard figures
are available. Figures are available
for corporate users (the ones who pay money for products), and in that
arena it's a slam dunk for SVR4 and apparently will continue to be so.
: <snip>
: -wayne
//===================================================================\\
|| J. Stephen Pendergrast, Jr. Software Engineer Consultant ||
|| p...@summit.novell.com Transaction Processing Systems ||
|| AppWare Systems Group Novell, Inc. ||
|| ||
|| OPINIONS ARE MINE ALONE. ||
\\===================================================================//
Path: nntp.gmd.de!xlink.net!howland.reston.ans.net!math.ohio-state.edu!
hobbes.physics.uiowa.edu!newsfeed.ksu.ksu.edu!moe.ksu.ksu.edu!
crcnis1.unl.edu!backbone!backbone!wayne
From: wa...@backbone.uucp (Wayne Schlitt)
Newsgroups: comp.unix.unixware,comp.os.linux,comp.os.linux.misc
Subject: Re: future of Unixware
Date: Thu, 9 Jun 1994 14:20:56 GMT
Organization: The Backbone Cabal
Lines: 51
Sender: wayne@backbone (Wayne Schlitt)
Message-ID: <WAYNE.94Jun9082056@backbone.uucp>
References: <CquoKE.7no@rahul.net> <8561@heimdall.sdrc.com> <8565@heimdall.sdrc.com>
<WAYNE.94Jun7222952@backbone.uucp> <rick.771073998@digibd>
Reply-To: wa...@cse.unl.edu
NNTP-Posting-Host: cse.unl.edu
In-Reply-To: rick@digibd.digibd.com's message of Wed, 8 Jun 1994 11: 13:18 GMT
In article <rick.771073998@digibd> r...@digibd.digibd.com (Rick Richardson) writes:
> wa...@backbone.uucp (Wayne Schlitt) writes:
>
> >So, it seems very likely to me that there are 300,000-600,000 Linux
> >users, and it is possible for there to be close to a million of them
> >out there.
>
> You missed one source of users for SVR4. Using information which
> is publically available, I estimate the size of the PC multiport
> serial market at $200M, or about 2 million ports/year. 70% of
> that is Unix (all flavors), or 1.4 million ports/year. I don't
> know what the breakdown is after that, but lets be conservative:
>
> -SCO 80% 1,112,000 ports/year
> -SVR4 20% 280,000 ports/year
> -Linux 0% 0 ports/year
Hmmm.... I am sure you have better info on this market than I do,
seeing as you are involved in it, but....
* I had no idea that the _intelegent_ multi-port market was so
_large_. Obviously, you must be refering to only the cards that
require special device drivers since you gave Linux 0%. I know that
people are buying Digiboards to run put in Linux boxes, it is being
discussed right now in the Linux newsgroups. They just they use
the 16550 versions...
* I had no idea that the intelegent multi-port market was so
_expensive_. I bought a 4-port 16550 based multi port card for
around $100, or $25 per port. Your above calculations say that
they cost around $100 per port.
* I had no idea that people use multi-port cards for just adding
terminals. I thought they added modems for uucp and faxes,
printers, plotters, etc. Strange as it may seem, I have a higher
percentages of terminals (25%) hooked up to my home machines rs232
ports, than where I work. At work, there are _lots_ of rs232 ports
being used, but only rarely is it to add terminals or modems used by
people.
-wayne
--
The human brain is a complex organ with the wonderful power of
enabling man to find reasons for continuing to believe whatever it is
that he wants to believe. -Voltaire
Newsgroups: comp.unix.unixware,comp.os.linux,comp.os.linux.misc
Path: nntp.gmd.de!Germany.EU.net!EU.net!uunet!gatech!newsxfer.itd.umich.edu!
zip.eecs.umich.edu!umn.edu!newsdist.tc.umn.edu!uum1!gw.digibd.com!digibd!rick
From: r...@digibd.digibd.com (Rick Richardson)
Subject: Re: future of Unixware
Message-ID: <rick.771219594@digibd>
Sender: n...@gw.digibd.com (USENET News)
Nntp-Posting-Host: digibd.digibd.com
Organization: Digi International, Incorporated, Eden Prairie,MN
References: <CquoKE.7no@rahul.net> <8561@heimdall.sdrc.com> <8565@heimdall.sdrc.com>
<WAYNE.94Jun7222952@backbone.uucp> <rick.771073998@digibd>
<WAYNE.94Jun9082056@backbone.uucp>
Date: Fri, 10 Jun 1994 03:39:54 GMT
Lines: 31
[ My back-of-the-envelope calculation of SVR4 users served by serial ports
has been deleted. -Rick
]
wa...@backbone.uucp (Wayne Schlitt) writes:
>* I had no idea that the intelegent multi-port market was so
> _expensive_. I bought a 4-port 16550 based multi port card for
> around $100, or $25 per port. Your above calculations say that
> they cost around $100 per port.
I actually purposely picked a figure out of the hat that I knew was too
high, in order to make the number of ports lower, giving the benefit of the
doubt in Linux's favor. In fact, every number I picked was designed to
make Linux look as good as possible.
Checking an old Unix review ad from a reseller, I find, for example, our
very popular PC/16em discounted to $1095, or about $69/port street price.
A pretty fair deal for what you get - *true* 115kbaud performance on 16
ports expandable to 64. I think it uses about 12% of a 486/25 to run 16
ports at 115kbaud output. If I remember correctly, the excellent FAS
driver on 16550's uses about the same amount of CPU to run 1 or maybe 2
ports on a 486/25, and I don't even want to talk about the hideous host
loading with *some* nameless operating systems stock drivers on 16550's.
-Rick
--
Rick Richardson Senior Staff Engineer Visit my WWW home page:
DigiBoard APD Email: r...@digibd.com http://www.digibd.com/people/rick
6801 Shady Oak Rd. Fax: (612) 947-1129
Eden Prarie, MN 55344 Tel: (612) 947-1111 <standard disclaimer>
Newsgroups: comp.unix.unixware,comp.os.linux,comp.os.linux.misc
Path: nntp.gmd.de!xlink.net!howland.reston.ans.net!news.moneng.mei.com!
hookup!reptiles.org!gts!robohack!telly!evan
From: e...@telly.on.ca (Evan Leibovitch)
Subject: Re: future of Unixware
Message-ID: <Cr5zEF.Ltx@telly.on.ca>
Date: Fri, 10 Jun 1994 04:34:59 GMT
References: <WAYNE.94Jun7222952@backbone.uucp> <2t516p$6vl@bird.summit.novell.com>
Organization: Somewhere just far enough out of Toronto
Lines: 40
In article <2t516p$...@bird.summit.novell.com>
p...@usl.com (Steve Pendergrast) writes:
>The actual number of active SVR4 installations (yes, it takes into
>account retired machines and upgrades) is at the high end of your
>estimate (according to IDC research), a little over 1 million.
>However, the average number of users per system is something like 5
>in the case of SVR4, making the number of users on the order of 5
>million. I'd bet 95% of the Linux installations are single user, but
>I admit I have no figures to back that up.
While we don't have hard and fast data, Rick made the point earlier
about the availability of serial cards for commercial Unix and Linux;
There appear to be no members of the current crop of high capacity
( that is, > 16 ports) intelligent serial cards which support Linux.
Most of the Linux supported stuff is 4-port 16550 cards, few of which
come with more than four ports.
Hooking up >96 ports to a commercial Unix box is trivial; doing it with
Linux is impossible without a terminal server, which adds in its own
series of headaches, especially if you just want a stand-alone box with
no networking.
>NOT! All you can claim is that there might be a few hundred thousand
>people who might be playing with Linux, but no hard figures
>are available. Figures are available
>for corporate users (the ones who pay money for products), and in that
>arena it's a slam dunk for SVR4 and apparently will continue to be so.
Also, as Rick suggested, there hasn't even been much call at DigiBoard
for a Linux driver.
My guess is that while there may be plenty of Linux installations out
there, once one counts the number of *seats* using R4 as opposed to
Linux, it's no contest.
--
Evan Leibovitch, Sound Software Ltd., located in beautiful Brampton, Ontario
e...@telly.on.ca / uunet!utzoo!telly!evan / (905) 452-0504
"We're just not good committee people. Some of us don't have that long to live"
Newsgroups: comp.unix.unixware,comp.os.linux,comp.os.linux.misc
Path: nntp.gmd.de!Germany.EU.net!EU.net!uknet!cf-cm!cybaswan!iiitac
From: iii...@uk.ac.swan.pyr (Alan Cox)
Subject: Re: future of Unixware
Message-ID: <1994Jun10.122856.10401@uk.ac.swan.pyr>
Organization: Swansea University College
References: <rick.771073998@digibd> <WAYNE.94Jun9082056@backbone.uucp>
<rick.771219594@digibd>
Date: Fri, 10 Jun 1994 12:28:56 GMT
Lines: 14
In article <rick.771219594@digibd> r...@digibd.digibd.com (Rick Richardson)
writes:
>A pretty fair deal for what you get - *true* 115kbaud performance on 16
>ports expandable to 64. I think it uses about 12% of a 486/25 to run 16
>ports at 115kbaud output. If I remember correctly, the excellent FAS
>driver on 16550's uses about the same amount of CPU to run 1 or maybe 2
>ports on a 486/25, and I don't even want to talk about the hideous host
Its about 1% loading with the new tty drivers on my Linux system + the loading
from user input itself. I remember what the smart tty controllers on an
NCR tower 16 could do to your machine load however so I'm still a fan of
smart serial boards for some uses.
Alan
Newsgroups: comp.unix.unixware,comp.os.linux,comp.os.linux.misc
Path: nntp.gmd.de!Germany.EU.net!EU.net!uknet!cf-cm!cybaswan!iiitac
From: iii...@uk.ac.swan.pyr (Alan Cox)
Subject: Re: future of Unixware
Message-ID: <1994Jun10.123214.10674@uk.ac.swan.pyr>
Organization: Swansea University College
References: <WAYNE.94Jun7222952@backbone.uucp> <2t516p$6vl@bird.summit.novell.com>
<Cr5zEF.Ltx@telly.on.ca>
Date: Fri, 10 Jun 1994 12:32:14 GMT
Lines: 15
In article <Cr5zEF....@telly.on.ca> e...@telly.on.ca (Evan Leibovitch) writes:
>My guess is that while there may be plenty of Linux installations out
>there, once one counts the number of *seats* using R4 as opposed to
>Linux, it's no contest.
Its undoubtedly the case that Linux and probably stuff like Solaris have
a lower users/license count than say SCO. Which makes it even more ironic
that these people charge for the per user licenses.
The average number of users/linux box according to some of the Linux
survey attempts is about 3-4 but thats mostly made up of 1-2 user machines
and machines with 10+ simultaneous logins.
Alan
Newsgroups: comp.unix.unixware,comp.os.linux.misc
Path: nntp.gmd.de!xlink.net!howland.reston.ans.net!math.ohio-state.edu!
jussieu.fr!univ-lyon1.fr!frmug.fr.net!cpio1!cpio1.frmug.fr.net!bernard
From: bern...@cpio1.frmug.fr.net (Bernard Fouche)
Subject: Re: Multiport Bored and Linux (Was: future of Unixware)
Message-ID: <1994Jun13.134758.13245@frmug.fr.net>
Sender: n...@frmug.fr.net (Net News Admin)
Organization: CPIO S.A.
References: <CquoKE.7no@rahul.net> <8565@heimdall.sdrc.com>
<WAYNE.94Jun7222952@backbone.uucp> <rick.771073998@digibd>
<2t626o$2b8@angmar.dataflux.bc.ca>
Date: Mon, 13 Jun 1994 13:47:58 GMT
Lines: 23
In article <rick.771073998@digibd>,
Rick Richardson <r...@digibd.digibd.com> wrote:
>[bits deleted]
>I estimate the size of the PC multiport
>serial market at $200M, or about 2 million ports/year. 70% of
>that is Unix (all flavors), or 1.4 million ports/year. I don't
>know what the breakdown is after that, but lets be conservative:
>[some conservative figures deleted]
> -Linux 0% 0 ports/year
My company bought about 10 boards from digiboard to use them over ISC
and Dell Unix. ISC has been bought by Sun, Dell Unix has been dropped
by Dell, we do not want to develop under SCO, etc. There is now only
Unixware that seems serious enough. But for how long ?
Linux seems to be very near to be stable enough to run business
applications on it. The only reason that we do not plan to quickly
move to Linux is the lack of support from hardware manufacturers (like
Digiboard) and major sofware house (that sells RDBMS).
--------------------------------------------------------------------------
Email : bern...@cpio1.frmug.fr.net
Postal Mail : CPIO S.A., 4 Rue Beaubourg, 75004 Paris, France
Path: nntp.gmd.de!Germany.EU.net!EU.net!uunet!sangam!iitb!
bhairav.ee.iitb.ernet.in!annapurna!vinod
From: vinod@annapurna (Vinod.G.Kulkarni)
Newsgroups: comp.unix.unixware,comp.os.linux,comp.os.linux.misc
Subject: Re: future of Unixware
Followup-To: comp.unix.unixware,comp.os.linux,comp.os.linux.misc
Date: 14 Jun 1994 04:37:04 GMT
Organization: Dept. of Electrical Engg., IIT, Bombay INDIA.
Lines: 26
Message-ID: <2tjc5g$4ma@bhairav.ee.iitb.ernet.in>
NNTP-Posting-Host: annapurna.ee.iitb.ernet.in
X-Newsreader: TIN [version 1.2 PL2]
Any product, to be successful, requires
-- A very good marketing support
-- A good amount of advertising
-- After sales support
-- support from other peripheral vendors etc.
Taking into consideration any big organisation like Novell, SCO, SUN
these are all available by default.
For linux, NONE of these are available. It is perhaps IMPOSSIBLE to
advertise for linux since the benefits of advertising neednot necessarily
go to advertiser. (At least not at the scale expected.) After sales
support is not possible easily - no one company can set up its offices
all over to support linux.
In spite of so many deficiencies, you find so much of discussion going on
about linux! Does it convey anything?
I believe it is just matter of time before we see linux as the major
base platform for unix! (At least for we Indians, linux is going to
save much of foreign exchange by removing need for other unix and even
novell platforms.)
Vinod
Dept.of CSE
IIT Bombay INDIA.
Newsgroups: comp.unix.unixware,comp.os.linux,comp.os.linux.misc
Path: nntp.gmd.de!Germany.EU.net!EU.net!uknet!cf-cm!cybaswan!iiitac
From: iii...@uk.ac.swan.pyr (Alan Cox)
Subject: Re: future of Unixware
Message-ID: <1994Jun14.114741.9058@uk.ac.swan.pyr>
Organization: Swansea University College
References: <2tjc5g$4ma@bhairav.ee.iitb.ernet.in>
Date: Tue, 14 Jun 1994 11:47:41 GMT
Lines: 22
In article <2tjc5g$...@bhairav.ee.iitb.ernet.in> vinod@annapurna
(Vinod.G.Kulkarni) writes:
Not after sales support. Not in my experience. I've had great marketing demos
(doesn't windows NT run well on 35mm slides 8)) seen the adverts but try phoning
a bug report in.
>
>For linux, NONE of these are available. It is perhaps IMPOSSIBLE to
Lots of people sell Linux support. Yggdrasil for one do both personal and
corporate support. You pay for your support though - thats a big part of the
GNU concept.
>I believe it is just matter of time before we see linux as the major
>base platform for unix! (At least for we Indians, linux is going to
>save much of foreign exchange by removing need for other unix and even
>novell platforms.)
I can't speak for India 8). Most of what I see is a different market. It really
is a 'personal unix system' to many people. Software for slip client sides,
pop clients for Unix and stuff are much in the 'wanted' area, and the normal
pop servers and stuff are of much less interest. A lot of it is a different
target area.
Alan
Newsgroups: comp.unix.unixware,comp.os.linux.misc
Path: nntp.gmd.de!xlink.net!howland.reston.ans.net!spool.mu.edu!sgiblab!
a2i!truffula!cls
From: c...@truffula.sj.ca.us (Cameron L. Spitzer)
Subject: Only 7000 Linux boxes, Re: Multiport Bored ...
Keywords: market share, Arbitron, device driver support, demographics
References: <CquoKE.7no@rahul.net> <rick.771073998@digibd>
<2t626o$2b8@angmar.dataflux.bc.ca> <1994Jun13.134758.13245@frmug.fr.net>
Followup-To: comp.os.linux.misc
Organization: Me and my cat.
Date: Tue, 14 Jun 1994 03:55:15 GMT
Message-ID: <1994Jun14.035515.3041@truffula.sj.ca.us>
Summary: why hardware co.s don't support Linux more
Lines: 24
In article <1994Jun13.134758.13...@frmug.fr.net> bern...@cpio1.frmug.fr.net
(Bernard Fouche) writes:
>
>Linux seems to be very near to be stable enough to run business
>applications on it. The only reason that we do not plan to quickly
>move to Linux is the lack of support from hardware manufacturers (like
>Digiboard) and major sofware house (that sells RDBMS).
Argh! I work for (but don't speak for!) a major manufacturer of
boards for PCs.
As far as I can tell, the *ONLY* reason we have not already developed,
tested, documented, released, and field-supported performance-optimized
Linux drivers is that the best data we can get say there are less than
7000 actual Linux systems in use! That data is from the Linux Counter
project which is announced now and then in comp.os.linux.announce.
We can't spend tens of thousands of US dollars and divert critical
personnel based on a survey in a German magazine, or newsgroup Arbitron
ratings, which suggest the numbers might be higher.
(Or at least that's what the product managers tell me.)
Have **YOU** registered **YOUR** Linux system?? I've registered mine.
Cameron in San Jos\'e
(who will be off the net on vacation till the 20th.)
Not speaking for *anyone* but myself!!!
Newsgroups: comp.unix.unixware,comp.os.linux.misc
Path: nntp.gmd.de!Germany.EU.net!EU.net!uunet!cs.utexas.edu!utnut!utzoo!
telly!evan
From: e...@telly.on.ca (Evan Leibovitch)
Subject: Re: Multiport Bored and Linux (Was: future of Unixware)
Message-ID: <CrGLxI.Gux@telly.on.ca>
Date: Wed, 15 Jun 1994 22:17:38 GMT
References: <2t626o$2b8@angmar.dataflux.bc.ca>
<1994Jun13.134758.13245@frmug.fr.net> <1994Jun13.221751.4801@kf8nh.wariat.org>
Organization: Somewhere just far enough out of Toronto
Lines: 42
In article <1994Jun13.221751.4...@kf8nh.wariat.org>
b...@kf8nh.wariat.org (Brandon S. Allbery) writes:
>Joe Portman has Oracle 6 running under iBCS2 emulation; I have Unify 2000
>running; various others have Informix-SE running, and Informix-Online should
>work once Linux IPC has been rewritten to fix various problems with the
>message queue interface (among others). Ironically, the biggest failure to
>date is dBASE IV... which should also be fixed when the IPC code is fixed.
>Native support seems almost redundant. It also seems unlikely, especially
>from the "first tier" RDBMS vendors.
But that's the whole point. The applications drive the OS, not the other
way around, in commercial environments.
It is a totally useless point that Linux runs this-or-that database vendor,
though it may make people feel good. What's important is that the database
vendor supports Linux. And so far, of the major database vendors are.
IMO, the reason is that they figure someone who's about to spend >$10K
on their database engine and 4GL, isn't going to piss around arguing
that the OS cost $1500. When you add training, peripherals, wiring,
networking, system design, customization, administration and maintenance,
the difference between the cost of installing (unsupported by the DB
vendor) Linux or (supported) UnixWare is insignificant.
UnixWare, to those who use it (at least the ones I know), is a tool, a
means to an end, an appliance. It's neither a project unto itself nor
a substantial part of the effort that goes into automating a business.
Nor should it be, which is perhaps why happy UnixWare users can't get as
worked up about their OS as happy Linux users.
Linux, as an inexpensive entry to Unix or a research project, is
perfect. But let's not pump it up into the best solution for everything.
Both Linux and the commercial systems have their rightful places, and
both can (and should) learn from each other. Saying that one system is
better than the other in all cases is nothing more than religion.
--
Evan Leibovitch, Sound Software Ltd., located in beautiful Brampton, Ontario
e...@telly.on.ca / uunet!utzoo!telly!evan / (905) 452-0504
"We're just not good committee people. Some of us don't have that long to live"
Path: nntp.gmd.de!Germany.EU.net!EU.net!howland.reston.ans.net!
europa.eng.gtefsd.com!newsxfer.itd.umich.edu!newsrelay.iastate.edu!
hobbes.physics.uiowa.edu!newsfeed.ksu.ksu.edu!moe.ksu.ksu.edu!news.mid.net!
crcnis1.unl.edu!backbone!backbone!wayne
From: wa...@backbone.uucp (Wayne Schlitt)
Newsgroups: comp.unix.unixware,comp.os.linux.misc
Subject: Re: Multiport Bored and Linux (Was: future of Unixware)
Date: Thu, 16 Jun 1994 06:52:05 GMT
Organization: The Backbone Cabal
Lines: 93
Sender: wayne@backbone (Wayne Schlitt)
Message-ID: <WAYNE.94Jun16005205@backbone.uucp>
References: <2t626o$2b8@angmar.dataflux.bc.ca> <1994Jun13.134758.13245@frmug.fr.net>
<1994Jun13.221751.4801@kf8nh.wariat.org> <CrGLxI.Gux@telly.on.ca>
Reply-To: wa...@cse.unl.edu
NNTP-Posting-Host: cse.unl.edu
In-Reply-To: evan@telly.on.ca's message of Wed, 15 Jun 1994 22: 17:38 GMT
In article <CrGLxI....@telly.on.ca> e...@telly.on.ca (Evan Leibovitch) writes:
> In article <1994Jun13.221751.4...@kf8nh.wariat.org>
> b...@kf8nh.wariat.org (Brandon S. Allbery) writes:
> >Native support seems almost redundant. It also seems unlikely, especially
> >from the "first tier" RDBMS vendors.
>
> But that's the whole point. The applications drive the OS, not the other
> way around, in commercial environments.
>
> It is a totally useless point that Linux runs this-or-that database vendor,
> though it may make people feel good. What's important is that the database
> vendor supports Linux. And so far, of the major database vendors are.
Most vendors support only SCO, and if you try to run their product on
any other version of Unix and have problems, well, gee, upgrade to SCO
is the solution. Yeah, UW gets some support, but if you choose
Consensys, IF, Dell, UHC, etc, you are kind of running on the edge
anyway. Linux is certainly not going to be any worse off than say
Coherent or QNX.
> IMO, the reason is that they figure someone who's about to spend >$10K
> on their database engine and 4GL, isn't going to piss around arguing
> that the OS cost $1500. When you add training, peripherals, wiring,
> networking, system design, customization, administration and maintenance,
> the difference between the cost of installing (unsupported by the DB
> vendor) Linux or (supported) UnixWare is insignificant.
Yes, and this exactly the argument that was used to show that Unipress
would always dominate the emacs market over GNU emacs. Heck, this
argument was also used as to why PCs would never replace mainframes.
The flaw in this logic is that once you have brought something in,
whether it is Linux, OS/2 or UW, you have already made the really huge
investment in installation, training, etc. In order to _keep_ this
investment, you will fight to get support on what you have.
If a few programmers bring in Linux so that they can use it instead of
Desview/X, and then start spreading it around, it will eventually work
its way into more and more critical/important spots. The database
might be just a small project that isn't critical at first, but after
a few years it is grown to the point that you are running dozens of
users on it. The two person outfit that started on a shoe string can
turn into a thriving business that has never had time to replace that
Linux system with a "real" Unix.
Yeah, no one is going to sit down today and say "I am going to install
this $100,000 dollar system that supports 200 users and is critical to
the business and I am going to run it on Linux to save a few bucks."
It is just after 5 or 10 years, you look back and see that you have done
just that. You may have started small, but once you trust it a little
bit, you start doing more with it.
The 6.4 billion dollar question is: "How much will Linux creap into
businesses over the next few years?" If it starts to make a
substantial presence (5-10%) then in 5 to 10 years, you can rest
assured that people _will_ be sinking $100,000 (plus inflation) into
Linux systems.
In a lot of ways, it is much easier to "grow up" than it is to "grow
down". People _are_ replacing IBM mainframes with PCs. Considering
most PCs have more power than 10 year old mainframes, this isn't too
surprising. VAXs and AS/400s have grown up to replacing mainframes
too. But the PC/370 went no where, and even the best example of
"growing down" (the VAX) didn't make that much of an impact on the
low end market.
Actually, the "grow down" problem has been Unix's problem with
breaking into the desktop market for years. It has "grown up" to
replace mini and mainframe OSes, but it has had a real hard time
growing down.
Anyway, I doubt that too many people will even be playing with things
like Oracle on Linux, but word processors, spread sheets, "toy"
databases and other stuff that people buy for the desktop has a lot of
potential to be sold to Linux customers. The Linux market is _not_
going to look like the SCO market. Linux people, on the whole, are
going to be much more intersted in soundblasters than multiport serial
cards.
-wayne
--
The human brain is a complex organ with the wonderful power of
enabling man to find reasons for continuing to believe whatever it is
that he wants to believe. -Voltaire
Newsgroups: comp.os.linux.misc
Path: nntp.gmd.de!Germany.EU.net!EU.net!uknet!cf-cm!cybaswan!iiitac
From: iii...@uk.ac.swan.pyr (Alan Cox)
Subject: Re: Only 7000 Linux boxes, Re: Multiport Bored ...
Message-ID: <1994Jun16.140920.13263@uk.ac.swan.pyr>
Keywords: market share, Arbitron, device driver support, demographics
Organization: Swansea University College
References: <2t626o$2b8@angmar.dataflux.bc.ca>
<1994Jun13.134758.13245@frmug.fr.net> <1994Jun14.035515.3041@truffula.sj.ca.us>
Date: Thu, 16 Jun 1994 14:09:20 GMT
Lines: 17
In article <1994Jun14.035515.3...@truffula.sj.ca.us> c...@truffula.sj.ca.us
(Cameron L. Spitzer) writes:
>As far as I can tell, the *ONLY* reason we have not already developed,
>tested, documented, released, and field-supported performance-optimized
>Linux drivers is that the best data we can get say there are less than
>7000 actual Linux systems in use! That data is from the Linux Counter
>project which is announced now and then in comp.os.linux.announce.
>We can't spend tens of thousands of US dollars and divert critical
>personnel based on a survey in a German magazine, or newsgroup Arbitron
>ratings, which suggest the numbers might be higher.
>(Or at least that's what the product managers tell me.)
>
If you make sure the relevant specs are available, the boards sensibly priced
and the Linux people know the board specs are available and complete all you
should need to do is wait.
Alan
Newsgroups: comp.unix.unixware,comp.os.linux.misc
Path: nntp.gmd.de!xlink.net!howland.reston.ans.net!usc!hookup!reptiles.org!
gts!robohack!telly!evan
From: e...@telly.on.ca (Evan Leibovitch)
Subject: Re: Multiport Bored and Linux (Was: future of Unixware)
Message-ID: <CrnJs8.HIt@telly.on.ca>
Date: Sun, 19 Jun 1994 16:14:27 GMT
References: <1994Jun13.221751.4801@kf8nh.wariat.org> <CrGLxI.Gux@telly.on.ca>
<WAYNE.94Jun16005205@backbone.uucp>
Organization: Somewhere just far enough out of Toronto
Lines: 120
In article <WAYNE.94Jun16005...@backbone.uucp> wa...@cse.unl.edu writes:
>In article <CrGLxI....@telly.on.ca> e...@telly.on.ca (Evan Leibovitch) writes:
>> It is a totally useless point that Linux runs this-or-that database vendor,
>> though it may make people feel good. What's important is that the database
>> vendor supports Linux. And so far, of the major database vendors are.
>Most vendors support only SCO, and if you try to run their product on
>any other version of Unix and have problems, well, gee, upgrade to SCO
>is the solution.
As a dealer who faithfully sold Esix for years before picking up
UnixWare, I don't think you really know how that corner of the
marketplace worked.
No, neither Esix nor Dell nor UHC nor Microport nor Interactive(!) nor
AT&T individually had a huge piece of the Intel SVR4 market. However,
combined their user base was significant enough to evoke commercial
support. The products were similar enough that vendors like WordPerfect,
Lotus and Progress (to name three with which I had first-hand experience)
*did* produce either a specific binary for SVR4, or specific support for
these Unix vendors using a catch-all System V Intel binary.
I would not have installed Esix at customer sites had I not known that
the applications would be both runnable *and* supported on that platform.
Sure, there were a number of products that only supported SCO, including
a number of hardware items. But the point is, that the stuff I
*needed* to run was supported *officially* by the application vendor.
I could say to WordPerfect honestly that I was running Esix and they'd
still support me without passing the buck to the OS immediately.
>Yeah, UW gets some support, but if you choose
>Consensys, IF, Dell, UHC, etc, you are kind of running on the edge
>anyway.
Not as on the edge as you may think. Certainly not enough to matter.
>Linux is certainly not going to be any worse off than say
>Coherent or QNX.
Accurate, but hardly a raging endorsement. Note that QNX has a
non-trivial amount of third-party support.
>> IMO, the reason is that they figure someone who's about to spend >$10K
>> on their database engine and 4GL, isn't going to piss around arguing
>> that the OS cost $1500. When you add training, peripherals, wiring,
>> networking, system design, customization, administration and maintenance,
>> the difference between the cost of installing (unsupported by the DB
>> vendor) Linux or (supported) UnixWare is insignificant.
>Yes, and this exactly the argument that was used to show that Unipress
>would always dominate the emacs market over GNU emacs.
I can't believe you would use this silly and useless an analogy.
1) Every system needs an OS. Not every system needs a spare text editor,
especially when most commercial sites are getting a commercial word
processor anyway;
2) Emacs has a high visibility in the techie world but *none* in the
corporate world; how many organoizations have budgets for training
on text editors or their administration?
3) GNU emacs' penetration in Unix sites, especially commercial Unix
sites, is far smaller than you may think. It's probably likely that
there are more copies of Linux floating around than GNU emacs.
>Heck, this
>argument was also used as to why PCs would never replace mainframes.
Merely because an argument has not worked in a totally different
situation does not invalidate it here. If you can't refute it, please
don't resort to this kind of desperation.
>The flaw in this logic is that once you have brought something in,
>whether it is Linux, OS/2 or UW, you have already made the really huge
>investment in installation, training, etc. In order to _keep_ this
>investment, you will fight to get support on what you have.
Welcome to the real world. Applications drive the OS, not vice versa.
First you pick the applications/4GLs/necessary hardware and only then
pick the right OS engine, based on the intersection of what's supported
by the applications.
>If a few programmers bring in Linux so that they can use it instead of
>Desview/X, and then start spreading it around, it will eventually work
>its way into more and more critical/important spots. The database
>might be just a small project that isn't critical at first, but after
>a few years it is grown to the point that you are running dozens of
>users on it. The two person outfit that started on a shoe string can
>turn into a thriving business that has never had time to replace that
>Linux system with a "real" Unix.
Which is fine as long as they don't have to run "real" applications.
>Yeah, no one is going to sit down today and say "I am going to install
>this $100,000 dollar system that supports 200 users and is critical to
>the business and I am going to run it on Linux to save a few bucks."
And nobody is going to say that about $10,000 5-user systems either.
>It is just after 5 or 10 years, you look back and see that you have done
>just that. You may have started small, but once you trust it a little
>bit, you start doing more with it.
>The 6.4 billion dollar question is: "How much will Linux creap into
>businesses over the next few years?" If it starts to make a
>substantial presence (5-10%) then in 5 to 10 years, you can rest
>assured that people _will_ be sinking $100,000 (plus inflation) into
>Linux systems.
Time will tell. But for those looking for a commercially viable system
*now*, the choice is clear.
--
Evan Leibovitch, Sound Software Ltd., located in beautiful Brampton, Ontario
e...@telly.on.ca / uunet!utzoo!telly!evan / (905) 452-0504
"We're just not good committee people. Some of us don't have that long to live"
Newsgroups: comp.unix.unixware,comp.os.linux.misc
Path: nntp.gmd.de!Germany.EU.net!EU.net!uunet!sparky!dsndata!backbone!
backbone!wayne
From: wa...@backbone.uucp (Wayne Schlitt)
Subject: Re: Multiport Bored and Linux (Was: future of Unixware)
In-Reply-To: evan@telly.on.ca's message of Sun, 19 Jun 1994 16: 14:27 GMT
Message-ID: <WAYNE.94Jun20223757@backbone.uucp>
Sender: wayne@backbone (Wayne Schlitt)
Reply-To: wa...@cse.unl.edu
Organization: The Backbone Cabal
References: <1994Jun13.221751.4801@kf8nh.wariat.org> <CrGLxI.Gux@telly.on.ca>
<WAYNE.94Jun16005205@backbone.uucp> <CrnJs8.HIt@telly.on.ca>
Date: Tue, 21 Jun 1994 04:37:57 GMT
Lines: 150
In article <CrnJs8....@telly.on.ca> e...@telly.on.ca (Evan Leibovitch) writes:
> In article <WAYNE.94Jun16005...@backbone.uucp> wa...@cse.unl.edu writes:
> >In article <CrGLxI....@telly.on.ca> e...@telly.on.ca (Evan Leibovitch) writes:
> >> It is a totally useless point that Linux runs this-or-that database vendor,
> >> though it may make people feel good. What's important is that the database
> >> vendor supports Linux. And so far, of the major database vendors are.
>
> >Most vendors support only SCO, and if you try to run their product on
> >any other version of Unix and have problems, well, gee, upgrade to SCO
> >is the solution.
>
> As a dealer who faithfully sold Esix for years before picking up
> UnixWare, I don't think you really know how that corner of the
> marketplace worked.
>
> [ but then goes on to say there are companies that do just what I
> said they did and that he only cares about a few products... ]
All I can say is that I must have poked around a different corner of
the Unix marketplace than you...
> >Linux is certainly not going to be any worse off than say
> >Coherent or QNX.
>
> Accurate, but hardly a raging endorsement. Note that QNX has a
> non-trivial amount of third-party support.
Hmm... So you are saying that you think that it is possible for Linux
to have a non-trivial amount of third-party support?
> >> IMO, the reason is that they figure someone who's about to spend >$10K
> >> on their database engine and 4GL, isn't going to piss around arguing
> >> that the OS cost $1500. When you add training, peripherals, wiring,
> >> networking, system design, customization, administration and maintenance,
> >> the difference between the cost of installing (unsupported by the DB
> >> vendor) Linux or (supported) UnixWare is insignificant.
>
> >Yes, and this exactly the argument that was used to show that Unipress
> >would always dominate the emacs market over GNU emacs.
>
> I can't believe you would use this silly and useless an analogy.
Nice ad Hominem..
> >Heck, this
> >argument was also used as to why PCs would never replace mainframes.
>
> Merely because an argument has not worked in a totally different
> situation does not invalidate it here. If you can't refute it, please
> don't resort to this kind of desperation.
Followed by a nice ad Ignorantiam and another ad hominem.
In my opinion, these situations are not totally different and nothing
you have said so far has convinced me otherwise. If you can not see
the similarities, let me know and I will email you a longer
description.
You are the one who is saying that Linux will "never be used by
someone who is willing to spend >$10k on a system." It is up to you
to provide prove that this is true, I don't have to prove it false.
All I have to do is give evidence that your logic is incomplete or
faulty.
> >The flaw in this logic is that once you have brought something in,
> >whether it is Linux, OS/2 or UW, you have already made the really huge
> >investment in installation, training, etc. In order to _keep_ this
> >investment, you will fight to get support on what you have.
>
> Welcome to the real world. Applications drive the OS, not vice versa.
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> First you pick the applications/4GLs/necessary hardware and only then
> pick the right OS engine, based on the intersection of what's supported
> by the applications.
I don't know which world you live in, but in the world that I live in
but for the most part, _this_is_not_true_. In many ways I wish what
you said was true, but for a vast majority of the time it isn't.
This is often true for first time buyers, or where the cost and the
benefits of the application can justify buying a complete new system.
Even then the customers often have (IMHO) wrong ideas about what will
be best and often won't change. Things like "Well, DOS systems _must_
be cheaper", or "I only want Open Systems", or "If it isn't from IBM
we wont buy it", or "PC hardware must be cheaper" or "Mainframe
hardware must be more reliable", etc.
For secondary purchase, most people look at only the applications that
already run on whatever OS and hardware that they have. Only if the
functionality that they really want can't be found for that
combination or if the functionality/price ratio isn't high enough do
they start looking at changing either the OS or the hardware. They
may also start looking to convince someone to do better on that
particular OS/hardware combination.
I don't know about you, but I don't want an Amiga on my desk because
it has the best paint program and a Mac because it has the best word
processor and a RISC box because it runs the database the best and a
PC because of the spread sheet. I will choose one (or maybe a few) of
those programs as the one I am really interested in and I will make
compromises on the rest. Actually, even this isn't true. I will
choose one of these combinations of OS and hardware at some point in
the past, and make compromises until I can find something that
justifies switching everything to a new system. Usually, this just
means switching hardware or the version of the OS.
This is where things like PCs, GNU Emacs, or Linux come in. Because
they are much cheaper than the alternatives, they are brought in as a
low-cost item for non-critical applications. The two areas that tend
to open new markets are when something has a much lower cost than
before, or when something has a new functionality that was never
available before.
Over time, people became comfortable with PCs and GNU Emacs and used
them for more critical applications. I think that this may also be
possible for Linux.
> >If a few programmers bring in Linux so that they can use it instead of
> >Desview/X, and then start spreading it around, it will eventually work
> >its way into more and more critical/important spots. The database
> >might be just a small project that isn't critical at first, but after
> >a few years it is grown to the point that you are running dozens of
> >users on it. The two person outfit that started on a shoe string can
> >turn into a thriving business that has never had time to replace that
> >Linux system with a "real" Unix.
>
> Which is fine as long as they don't have to run "real" applications.
Why not? People have been running stuff on unsupported systems for
years. If it works for what they want to do and it is cheaper, then
there is little reason why they won't use Linux to run "real"
applications. There are people today using Linux to run "real"
applications, and I think that that is fine.
-wayne
--
The human brain is a complex organ with the wonderful power of
enabling man to find reasons for continuing to believe whatever it is
that he wants to believe. -Voltaire
Newsgroups: comp.unix.unixware,comp.os.linux.misc
Path: bga.com!news.sprintlink.net!hookup!swrinde!cs.utexas.edu!utnut!
utzoo!telly!evan
From: e...@telly.on.ca (Evan Leibovitch)
Subject: Re: Multiport Bored and Linux (Was: future of Unixware)
Message-ID: <Crrup0.39E@telly.on.ca>
Date: Wed, 22 Jun 1994 00:00:32 GMT
References: <WAYNE.94Jun16005205@backbone.uucp> <CrnJs8.HIt@telly.on.ca>
<WAYNE.94Jun20223757@backbone.uucp>
Organization: Somewhere just far enough out of Toronto
Lines: 139
In article <WAYNE.94Jun20223...@backbone.uucp> wa...@cse.unl.edu writes:
>> >Linux is certainly not going to be any worse off than say
>> >Coherent or QNX.
>> Accurate, but hardly a raging endorsement. Note that QNX has a
>> non-trivial amount of third-party support.
>Hmm... So you are saying that you think that it is possible for Linux
>to have a non-trivial amount of third-party support?
Yes. Whether that support extends to the major DBMS vendors (none of
whome support QNX or Coherent) is for you to decide.
>> >> IMO, the reason is that they figure someone who's about to spend >$10K
>> >> on their database engine and 4GL, isn't going to piss around arguing
>> >> that the OS cost $1500. When you add training, peripherals, wiring,
>> >> networking, system design, customization, administration and maintenance,
>> >> the difference between the cost of installing (unsupported by the DB
>> >> vendor) Linux or (supported) UnixWare is insignificant.
>> >Yes, and this exactly the argument that was used to show that Unipress
>> >would always dominate the emacs market over GNU emacs.
>> I can't believe you would use this silly and useless an analogy.
>Nice ad Hominem..
\begin{mild_flame}
Check out the dictionary on what an ad hominum argument is, before
tossing the accusation around like a broken frisbee. Ad hominum
arguments are those which attack the speaker and not what is said.
I believed your analogy was silly and useless (and I still do), but
that says nothing of what I think of you. While you may not like my
tone, don't fling around insults unless you know what they mean. They
certainly don't help you establish a defense to the points. Either
defend the silly analogy, or dump it.
\end{mild_flame}
>> >Heck, this
>> >argument was also used as to why PCs would never replace mainframes.
>> Merely because an argument has not worked in a totally different
>> situation does not invalidate it here. If you can't refute it, please
>> don't resort to this kind of desperation.
>Followed by a nice ad Ignorantiam and another ad hominem.
The reasoning is desperate, not (necessarily) the reasoner.
>You are the one who is saying that Linux will "never be used by
>someone who is willing to spend >$10k on a system."
Not really. The lack of interest in Linux by major applications and
hardware vendors can do that better than I can, whether or not you
choose to ignore that fact of life. You want to argue with DigiBoard
and Oracle and Ataptec and Borland, go ahead.
I've tried to explain the lack of interest, not prove it.
And if I said "never", I apologise, I should have said "rarely".
There will always be freaks against the norm who will do something
merely to prove it can be done. That does not mean the practise is
common.
>> Welcome to the real world. Applications drive the OS, not vice versa.
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>> First you pick the applications/4GLs/necessary hardware and only then
>> pick the right OS engine, based on the intersection of what's supported
>> by the applications.
>I don't know which world you live in, but in the world that I live in
>but for the most part, _this_is_not_true_. In many ways I wish what
>you said was true, but for a vast majority of the time it isn't.
Whose fault is that? If it's your wish to impress this line of thinking
on the customer, then you measure your success by whether you have made
your strategy palatable to the customer, and you convince them to do it
right the first time.
>This is often true for first time buyers, or where the cost and the
>benefits of the application can justify buying a complete new system.
>Even then the customers often have (IMHO) wrong ideas about what will
>be best and often won't change. Things like "Well, DOS systems _must_
>be cheaper", or "I only want Open Systems", or "If it isn't from IBM
>we wont buy it", or "PC hardware must be cheaper" or "Mainframe
>hardware must be more reliable", etc.
Then the onus is on *you*, as the VAR, to make the other case. We have
walked out on clients who insisted on being pig-headed about their
approaches; on about half of those we've come in later on to clean up
the damages, with the customer paying almost double what they would have
paid had they done it right the first time.
How does that ad go? "You can pay me now, or pay me later..."
>This is where things like PCs, GNU Emacs, or Linux come in. Because
>they are much cheaper than the alternatives, they are brought in as a
>low-cost item for non-critical applications.
I disagree, the first item of the three you mentioned *is* often used
now for critical applications.
>Over time, people became comfortable with PCs and GNU Emacs and used
>them for more critical applications. I think that this may also be
>possible for Linux.
I still don't see the analogy with GNU Emacs, which may not have a
fraction of the installed base you might think.
Still, if you want to look at Linux as the Emac of operating systems, go
ahead. I don't know what point it'll help you make...
>> >The two person outfit that started on a shoe string can
>> >turn into a thriving business that has never had time to replace that
>> >Linux system with a "real" Unix.
>> Which is fine as long as they don't have to run "real" applications.
>Why not? People have been running stuff on unsupported systems for
>years.
People, individuals, maybe. Big companies who depend on their computers
to run the business, rarely.
>If it works for what they want to do and it is cheaper, then
>there is little reason why they won't use Linux to run "real"
>applications. There are people today using Linux to run "real"
>applications, and I think that that is fine.
How nice for them. Would they put their jobs on the line for it?
--
Evan Leibovitch, Sound Software Ltd., located in beautiful Brampton, Ontario
e...@telly.on.ca / uunet!utzoo!telly!evan / (905) 452-0504
"We're just not good committee people. Some of us don't have that long to live"
Newsgroups: comp.unix.unixware,comp.os.linux.misc
Path: bga.com!news.sprintlink.net!news.onramp.net!convex!cs.utexas.edu!
swrinde!pipex!uknet!cf-cm!cybaswan!iiitac
From: iii...@uk.ac.swan.pyr (Alan Cox)
Subject: Re: Multiport Bored and Linux (Was: future of Unixware)
Message-ID: <1994Jun22.091933.3557@uk.ac.swan.pyr>
Organization: Swansea University College
References: <CrnJs8.HIt@telly.on.ca> <WAYNE.94Jun20223757@backbone.uucp>
<Crrup0.39E@telly.on.ca>
Date: Wed, 22 Jun 1994 09:19:33 GMT
Lines: 52
In article <Crrup0....@telly.on.ca> e...@telly.on.ca (Evan Leibovitch) writes:
>Not really. The lack of interest in Linux by major applications and
>hardware vendors can do that better than I can, whether or not you
>choose to ignore that fact of life. You want to argue with DigiBoard
>and Oracle and Ataptec and Borland, go ahead.
I don't care about Digiboards attitude - there are other drivers in Alpha
and Digiboard can either make sure decent specs are available (and people
will write the drivers needed) or watch as the other vendors for whom
Alpha test drivers come out. Most vendors are very helpful. I've had quite
interesting conversations with several about 'Can I get XX sale or return to
try with Linux iBCS2. Normally not only is the reply yes but also let us know
how well it works'.
Oracle is looking promising under Linux. Adaptec provided good info for the
writing of the AHA SCSI drivers (and thats sold them at least 3 boards just
to friends of mine). As to Borland - no thanks I'll stick to gcc.
There are a few hardware people who have been unhelpful to say the least
including the well known diamond and Xfree86 matter and the Xircom network
adapters. Fine - they've got as much chance of selling me product as seagate
have disks after the sticky disk drive fiasco a few years back.
>How does that ad go? "You can pay me now, or pay me later..."
Or pay someone else. Which is part of what the GPL is about. Since after
90 days people like microsoft are now you've paid us once now pay us again
and again and again I don't see the difference.
>>Why not? People have been running stuff on unsupported systems for
>>years.
>People, individuals, maybe. Big companies who depend on their computers
>to run the business, rarely.
Big companies regularly. Oh it may claim to be supported but its cheaper
to buy a different product and quicker to tour australia than to actual get
an answer from their 'support' phones.
>>If it works for what they want to do and it is cheaper, then
>>there is little reason why they won't use Linux to run "real"
>>applications. There are people today using Linux to run "real"
>>applications, and I think that that is fine.
>
>How nice for them. Would they put their jobs on the line for it?
Linux is running hospitals, NASA projects, our mail hub, problem tracking
database, usenet and MIPS cross development. I have had mail from inside three
different 'national security' agencies and two of those I know for sure use
Linux internally.
Alan
Newsgroups: comp.unix.unixware,comp.os.linux.misc
Path: bga.com!news.sprintlink.net!hookup!yeshua.marcam.com!MathWorks.Com!
uhog.mit.edu!news.mtholyoke.edu!news.byu.edu!news.provo.novell.com!uel!msohnius
From: msohn...@novell.co.uk (Martin Sohnius)
Subject: Re: Multiport Bored and Linux (Was: future of Unixware)
Sender: n...@novell.co.uk
Message-ID: <Crt3pA.7Dr@novell.co.uk>
Date: Wed, 22 Jun 1994 16:12:46 GMT
References: <CrnJs8.HIt@telly.on.ca> <WAYNE.94Jun20223757@backbone.uucp>
<Crrup0.39E@telly.on.ca> <1994Jun22.091933.3557@uk.ac.swan.pyr>
Organization: Novell Europe
X-Newsreader: TIN [version 1.2 PL0]
Followup-To: comp.unix.unixware,comp.os.linux.misc
Lines: 23
Alan Cox (iii...@uk.ac.swan.pyr) wrote:
^^^^^^^^^^^^^^
<snip>
: Linux is running hospitals, NASA projects, our mail hub, problem tracking
^^^^^^^^
: database, usenet and MIPS cross development. I have had mail from inside three
: different 'national security' agencies and two of those I know for sure use
: Linux internally.
Thanks, but no!
BTW, since this subject is now "Multiport Bored and Linux", can we stop
cross-posting to comp.unix.unixware? This might cool down the
discussion, too.
--
+--------------------------------------------+
Martin Sohnius | "The Information Superhighway" |
Novell Labs Europe | voted "Word of the Year 1993" |
Bracknell, England | by the American Dialect Society. |
+44-344-724031 | - The Wall Street Journal - |
+--------------------------------------------+
(I speak for myself, not for Novell or anyone else.)
Newsgroups: comp.unix.unixware,comp.os.linux.misc
Path: bga.com!news.sprintlink.net!hookup!yeshua.marcam.com!zip.eecs.umich.edu!
newsxfer.itd.umich.edu!gatech!howland.reston.ans.net!EU.net!uknet!cf-cm!
cybaswan!iiitac
From: iii...@uk.ac.swan.pyr (Alan Cox)
Subject: Re: Multiport Bored and Linux (Was: future of Unixware)
Message-ID: <1994Jun22.190934.17299@uk.ac.swan.pyr>
Organization: Swansea University College
References: <Crrup0.39E@telly.on.ca> <1994Jun22.091933.3557@uk.ac.swan.pyr>
<Crt3pA.7Dr@novell.co.uk>
Date: Wed, 22 Jun 1994 19:09:34 GMT
Lines: 20
In article <Crt3pA....@novell.co.uk> msohn...@novell.co.uk (Martin Sohnius) writes:
>Alan Cox (iii...@uk.ac.swan.pyr) wrote:
> ^^^^^^^^^^^^^^
><snip>
>: Linux is running hospitals, NASA projects, our mail hub, problem tracking
> ^^^^^^^^
>Thanks, but no!
Who said that machine was our mailhub - pyr is a pyramid OS/x and the address
is the wrong way around because its an X.29 coloured book host.
>BTW, since this subject is now "Multiport Bored and Linux", can we stop
>cross-posting to comp.unix.unixware? This might cool down the
>discussion, too.
Sensible
Alan