PC Unix/Xenix vendors
Larry Snyder
>This is all the info I have currently:
>SCO Open Desktop/UNIX (liked it enough to build a business around it)
requires the largest initial investment and continues
to rake users with expensive support and upgrades -- look
at the upgrade from ODT 1.X to 2.X -- a loaded system would
cost just about as much as another vendors product which is
based on real standards, not self defined ones (ie: SCO's
long file names and symbolic links come to mind). Of course,
if your income is based on sales and percentages, then SCO
might make you the most money in the short term --
>Dell (makes no upgrades, little to no support)
800 number support, quick support turn-around on call-backs, and
email support -- for what one paid for the ODT 1.X - 2.0 upgrade
one could have almost purchased the complete Dell with unlimited
users, DOS under UNIX, NFS, TCP, X11R5 (SCO is still behind with
X11R4 last time I checked) and as all SVR4's, support for long file
names, symbolic links and > 65k inodes.
Dell also has an excellent upgrade from prior release (cheap!)
Just my 2 cents
--
Larry Snyder internet: la...@gator.use.com
keeper of the Gator uucp: uunet!gator!larry
Christer Johansen
: la...@news.rn.com (Larry Snyder) writes:
: : gry...@openage.openage.com (The Golden Gryphon) writes:
: :
: : >This is all the info I have currently:
: : >SCO Open Desktop/UNIX (liked it enough to build a business around it)
: :
: : requires the largest initial investment and continues
: : to rake users with expensive support and upgrades -- look
: : >Dell (makes no upgrades, little to no support)
: :
: : 800 number support, quick support turn-around on call-backs, and
: : email support -- for what one paid for the ODT 1.X - 2.0 upgrade
: Not the point. It doesn't matter if SCO costs more, has less support or
: anything else you want to complain about. What matters is that SCO
: *advertises* and *sells*.
: Grab a thousand x86 Unix/Xenix boxes and *most* of them will be SCO. In
: fact, I've seen ONE AIX box, a couple of Interactives, and zero Dell
: in all my experience.
: That's the reason my business is based on SCO.
: As an individual user, you can justify (to yourself, anyway) buying
: Dell. Personally, I wouldn't recommend anything but SCO to a business,
: because I am darn sure that SCO will be selling Unix 5 years from now, but
: I cannot be as sure for anyone else.
:
: Tony a...@world.std.com
I agree with Tony on this one. What matters for most commercial Unix end-users
is that they get a stable platform on wich to run their applications and that the
Unix they choose still exists years from now. It's of course very hard to look in
to the future, but SCO should have a fair chance.
We use SCO ODT 2.0 ourselves, and it provides an excellent user friendly platform.
Most of our users use X-terminals, and SCO's integration of X.desktop with their
own tools and the standard X-tools provides an excellent environment for novice
X-users. They also have their own character based menu (odtsh) for those who still
use dumb terminals.
I'll have to say I don't understand why Mr. Snyder keeps on complaining about
SCO all the time. Every time he gets the change he seems to pick on them.
This is in no way ment to be a defence speech for SCO, but they can't possibly be
as bad as Larry seems to think.
I know very little about Dell Unix, but one could be fooled into belive that
Mr. Snyder is a major stockholder in that company.
--
Christer Johansen, Tandberg Data A/S, PO Box 9, Korsvoll, N-0808 OSLO 8, NORWAY
Phone +47 (0)2 189090, dir +47 (0)2 189527 Fax +47 (0)2 189550
Internet: ch...@tdata.no X.400: C=no; A=telemax; P=tdata; O=tdd; S=chjo
Larry Snyder
>I agree with Tony on this one. What matters for most commercial Unix end-users
>is that they get a stable platform on wich to run their applications and that the
>Unix they choose still exists years from now. It's of course very hard to look in
>to the future, but SCO should have a fair chance.
SCO is no more stable than Dell 2.2. And Dell 2.2 works and works. We
use it for our internet connection running multiple feeds -- try
PPP with SCO. Someone here mentioned that PPP with SCO is based on old
standards, and doesn't like to talk to the outside world unless it is
another SCO box running PPP. SCO also (according to the net) won't allow
a port to run both SLIP and bidirectional communications -- only one or
the other.
SCO be around, maybe. But if they keep trimming back their staff, who
knows?
>I'll have to say I don't understand why Mr. Snyder keeps on complaining about
>SCO all the time. Every time he gets the change he seems to pick on them.
Because SCO is so modified -- it's not real UNIX. SCO took Unix 3.2 and
made it SCO. They symbolic links and long file names from SVR4 and added
support with their 3.2 product calling it release 4.0 to confuse the issue.
They also removed support for alternate protocols like "g","e" and "f", then
added them (mind you they were in the original code from AT&T) in one of
their updates -- thus enabling them to charge for what had been available under
other vendors products for the pervious 3 years.
For your information, Dell also sells $CO, and no, I don't own Dell stock.
Steve Sheldon
>Because SCO is so modified -- it's not real UNIX. SCO took Unix 3.2 and
>made it SCO. They symbolic links and long file names from SVR4 and added
>support with their 3.2 product calling it release 4.0 to confuse the issue.
Hmm, what exactly makes something a "real" UNIX? We're running SCO
ODT 2.0 here, along with our DECStation's running Ultrix 4.2a.
I don't know how much SVR4 differs from SVR3.2(.4 SCO), but IMHO, the
DECstation is running "real" UNIX, and the AT&T system is simply a
kludge.
Now some people would say that Ultrix is not "real" UNIX.
It sounds to me, that this is really just a matter of opinion, and relates
more to what you're used to working with.
Except for the poor and confusing documentation that came with SCO, I do
like the system. It works well for working in a production situation like
what we use it for.
(and from what I understand the documentation that comes with Dell is no
better than SCO's)
--
she...@iastate.edu Steve Sheldon
Project Vincent ICSS Resource Unit
SCO ODT, Arc/Info, Atlas GIS 2142 Agronomy Hall
Iowa State University
Evan Leibovitch
you who don't get them through the e-mail gateway) read the other
newsgroups which have SCO-relevant content. This message (as have most
of the others in this thread,) is being cross-posted to all of them.
Anyone who's followed these groups, especially comp.unix.sysv386, should
know by now where Larry's coming from. He likes bandwagons. Loves them.
The term used in the newsgroups is Dellevangelist, and it seems to stick
pretty well.
Don't take it personally, SCO lovers. Larry's just as quick to tear
apart *any* Intel UNIX vendor besides Dell. To him, Dell can do no wrong
and the others can do no right.
Having said that, he often has valid points to make, if you can bother
to read his stuff through the appropriate bias filter. Being rabidly
pro-SCO is no better or worse than being rabidly anti-SCO.
In article <C0tEw...@world.std.com>
a...@world.std.com (Anthony P Lawrence) writes:
>la...@news.rn.com (Larry Snyder) writes:
>: [SCO] requires the largest initial investment and continues
>: to rake users with expensive support and upgrades -- look
>Not the point. It doesn't matter if SCO costs more, has less support or
>anything else you want to complain about.
I think it should matter more to you. If your application can run on two
platforms, and both are stable and do what you want them to do, wouldn't
it make sense to at least consider the one that's less expensive?
If you treat the systems software like the hardware -- as a commodity --
then stability, functionality and price are the driving concerns.
>What matters is that SCO *advertises* and *sells*.
That SCO sells as well as it has is partially due to the fact that there
haven't been any competitors with a serious committment to UNIX *and*
pockets deep enough to do the necessary marketing. SCO has essentially
had the field to itself, at worst being a Gulliver in a market full of
Liliputians.
Understandably, it may be difficult to consider a UNIX which is poorly
marketed and has little public profile. In technical terms and price
terms, though, it *may* be superior.
As has been stated in the editorial pages of UNIX Review, it's been easy
so far for SCO to stare down the half-hearted UNIX marketing attempts by
Interactive, Dell, Microport, ESIX, and even AT&T's own shot at Intel UNIX.
This is about to change. Novell has entered the Intel UNIX field, through
(its now wholly-owned subsidiary) Univel. Sun is due in soon. These
companies have good track records, though of course both have had their
turkeys. More importantly, both have the UNIX commitment and the
resources to take on SCO in the marketing trenches.
Univel has already started taking out glossy ads in various magazines.
It's packaged UnixWare almost identically to NetWare. That capitalizes
on Novell's reputation in the PC industry as a whole, which (like it or
not) dwarfs SCO's.
>Grab a thousand x86 Unix/Xenix boxes and *most* of them will be SCO. In
>fact, I've seen ONE AIX box, a couple of Interactives, and zero Dell
>in all my experience.
>
>That's the reason my business is based on SCO.
There are still a lot of good reasons to choose SCO, as I do for people
when it's the right choice for the installation. The lemming approach,
however, is not one of those reasons. I (and you too, I would imagine)
have enough confidence in our UNIX skills to choose a product because
*we* like it, not because the rest of the world does. There are
countless examples, not just in the computer field, when the rest of the
world can be wrong.
If momentum, marketing savvy and installed base were all that mattered, we'd
all still be using SAA. Or MS-DOS.
>As an individual user, you can justify (to yourself, anyway) buying
>Dell. Personally, I wouldn't recommend anything but SCO to a business,
>because I am darn sure that SCO will be selling Unix 5 years from now, but
>I cannot be as sure for anyone else.
I don't like making *any* sure bets like that. I wouldn't even put money
on Microsoft's being around in five years. Things change too damned fast
in this industry.
Frankly, I'm looking forward to SCO getting a whole lot better over the
coming months. Univel and Solaris will be the first *real* competition
the company has ever had, and I certainly don't expect SCO to sit on its
hands and just let the others muscle in.
What I've least liked about SCO is its corporate attitude of arrogance,
its feeling that it makes standards, rather than participates in their
creation and follows them afterwards. I still blame SCO for sabotaging
the Intel ABI programme, which would have helped the entire Intel UNIX
industry but made SCO more compatible with its competitors.
Some real competition will knock that attitude down a notch or three,
which in itself is good news. For the first time ever, SCO may have to
worry about its products being compatible with other companies' UNIX.
I expect some sort of Open Look and ELF capabilities from SCO within the
year, just as they've recently announced support for NetWare protocols.
The next few years promise to be very interesting indeed in this field.
--
Evan Leibovitch, Sound Software Ltd., located in beautiful Brampton, Ontario
ev...@telly.on.ca / uunet!utzoo!telly!evan / (416) 452-0504
What's with all this multimedia stuff? Most vendors can't get *one* done right.
Shane Bouslough
: [ much excellent analysis of SCO, et al, deleted ]
:
: I don't like making *any* sure bets like that. I wouldn't even put money
: on Microsoft's being around in five years. Things change too damned fast
: in this industry.
I'll take that bet! :-)
: What I've least liked about SCO is its corporate attitude of arrogance,
: its feeling that it makes standards, rather than participates in their
: creation and follows them afterwards. I still blame SCO for sabotaging
: the Intel ABI programme, which would have helped the entire Intel UNIX
: industry but made SCO more compatible with its competitors.
:
: Some real competition will knock that attitude down a notch or three,
: which in itself is good news. For the first time ever, SCO may have to
: worry about its products being compatible with other companies' UNIX.
Damn well said.
: Evan Leibovitch, Sound Software Ltd., located in beautiful Brampton, Ontario
: ev...@telly.on.ca / uunet!utzoo!telly!evan / (416) 452-0504
-Shane
--
Shane Bouslough | #include <stddisc.h>
sh...@sbcs.sunysb.edu | #include <happy.rhodes>
Daniel R. Edelson
>In <C0uD4...@news.rn.com> la...@news.rn.com (Larry Snyder) writes:
>
>>Because SCO is so modified -- it's not real UNIX. SCO took Unix 3.2 and
>>made it SCO. They symbolic links and long file names from SVR4 and added
>>support with their 3.2 product calling it release 4.0 to confuse the issue.
> It sounds to me, that this is really just a matter of opinion, and relates
>more to what you're used to working with.
All that matters is that your applications run and that your platform
is stable.
If emacs runs, one hurdle is crossed :-).
I've heard that the libraries with SCO's X11R4 implementation
are actually only R3. If true, that would be a disadvantage.
Anthony P Lawrence
: In article <C0tEw...@world.std.com>:
: >Not the point. It doesn't matter if SCO costs more, has less support or
: >anything else you want to complain about.
:
: I think it should matter more to you. If your application can run on two
: platforms, and both are stable and do what you want them to do, wouldn't
: it make sense to at least consider the one that's less expensive?
Well, first of all, I don't sell either SCO or any applications, all I
do is support and troubleshooting and maybe a little programming.
The reason I choose to concentrate on SCO is that they have enough market
penetration that I can maintain my standard of living specializing in
their OS and their OS alone. If I supported Dell, I think it would
be a long time between paychecks...
SCO as SCO has no particular importance to me. Nor does V3.2 2 vs 3.2 4
or the type of filesystem they use, or whether or not they support
mmap() or anything else. All I care about is how many business customers
buy it, and that's it. I'm just a parasite; a flea on the dog, and
if this dog dies I'll learn to like the taste of some other canine.
SCO is my business, not my crusade.
Tony a...@world.std.com
Lawrence & Clark, Inc (617) 762-0707 (206) 323-2864
Xenix/Unix support,etc Boston Seattle
Kevin Clark is embarrassed by most of what I say.
Henry van Cleef
>ODT 2.0 here, along with our DECStation's running Ultrix 4.2a.
>
> I don't know how much SVR4 differs from SVR3.2(.4 SCO), but IMHO, the
>DECstation is running "real" UNIX, and the AT&T system is simply a
>kludge.
>
> Now some people would say that Ultrix is not "real" UNIX.
>
> It sounds to me, that this is really just a matter of opinion, and relates
>more to what you're used to working with.
>
> Except for the poor and confusing documentation that came with SCO, I do
>like the system. It works well for working in a production situation like
>what we use it for.
>
> (and from what I understand the documentation that comes with Dell is no
>better than SCO's)
Now hold the phone a moment, Steve. I spent a year at Iowa State with
the Vincent system and a lot more years before that with Ultrix. Before
you talk about comparing equals to equals, Iowa State has an Ultrix
kernel source license, a large staff who have put years into tuning up
that kernel for the Vincent system.
Ultrix is a port of BSD4.2. It took DEC up to version 4.1 to get a
halfway decent system running. And where DEC is, in the Unix market, is
anybody's question. The latest cry is OSF/1, and those MIPS boxes you
have may very well be dead-end orphans. What Iowa State is running on
Vincent isn't a Unix that anyone can buy.
I will say point blank that I could build a Vincent system on SCO Unix
and Intel 486 Eisa setups that will run right along side what you have,
for a lot less money than it would cost to do it on Suns or DEC
products.
If you are talking about trying to integrate SCO Unix into the present
Vincent setup, lotsa luck. There are tons of symbolic link spiderwebs
in Vincent that don't make much space for any System V setup. That
includes any of the 5.4 products.
When it comes to documentation, (and I do know how much paper
documentation on Ultrix Iowa State does not have---but then it really
doesn't exist for anyone to buy, either), I have always felt boggled by
the boxes of books that come with ODT. Nobody's Unix comes with the
truckload of books that one can get for VMS or the IBM Mainframe setups.
O'Reilly to the rescue, but how many vendors make do with O'Reilly and
nothing else?
When it comes to "real Unix" vs. "imitation Unix," it seems pretty clear
to me that System V is historically "real" and the SCO implementation is
and always was cleaner than the AT&T 3b2 version. BSD diverged from
AT&T Unix 12 or 13 years ago, and has gotten the lion's share of
university attention, including yours. I am sitting here in front of a
SCO 3.2.4.1 box running X11R5, and it runs right along with a DEC 5000
MIPS workstation on Vincent, with some plusses, such as 30 seconds to
boot up, and ten seconds to log in under xdm. And I don't have to lean
on gcc as the only compiler that will compile some people's idea of
creativity, also known as "spaghetti code."
You can thank your lucky stars that if you feel like the lone ranger
running SCO Unix in that world that it isn't the other way around. You
have a team of top people---some of the best I have ever worked with in
30-odd years in this business---to deal with (small) crates from DEC,
and you can be damned glad you don't have to confront the box that
simply won't boot the TK-50 install tape or the other box that is, by
damn, simply not going to read from the CD-ROM.
--
Hank van Cleef vanc...@netcom.com vanc...@tmn.com Andover, Mass.
Unix systems consultant---kernel, device drivers, networking, X11
SysV, BSD, Sun, Ultrix, AIX, SCO. Porting a specialty.
The Union Institute History of Science
Evan Leibovitch
a...@world.std.com (Anthony P Lawrence) writes:
>Well, first of all, I don't sell either SCO or any applications, all I
>do is support and troubleshooting and maybe a little programming.
>[...]
>SCO as SCO has no particular importance to me. Nor does V3.2 2 vs 3.2 4
>or the type of filesystem they use, or whether or not they support
>mmap() or anything else. All I care about is how many business customers
>buy it, and that's it. I'm just a parasite; a flea on the dog, and
>if this dog dies I'll learn to like the taste of some other canine.
If your support of SCO is based *only* on the fact that your work is
totally in the aftermarket, I see your point. If you only come in after
the purchase is made, your reasoning ("I work with SCO because that's
what's out there...") makes sense.
BUT...
That's not what you said in the article I was responding to.
In article <C0tEw...@world.std.com>, you said:
>As an individual user, you can justify (to yourself, anyway) buying
>Dell. Personally, I wouldn't recommend anything but SCO to a business,
There you implied that people trusted your recommendations for purchases
*before* the fact. If you recommend SCO only because that was the one UNIX
you were familiar with, then I don't think you're giving your recommendee
the best possible advice. It's the choice that will keep you working
after they buy, but it may not be the best choice for them.
>because I am darn sure that SCO will be selling Unix 5 years from now, but
>I cannot be as sure for anyone else.
Check out SCO's financial situation. According to UNIX Review, SCO
hasn't been more than a break-even company for all the years its had
the market to themselves. With the new competition, they'll be losing
market share and/or lowering their prices and/or adding new features,
all of which will hurt SCO's bottom line in the short-to-medium term.
And there are other factors to consider when boasting of SCO's
"Prudential rock" image. A lot of money and manpower went into ACE,
resources that can be now considered down the toilet. *AND*, at the worst
possible time, the company's guiding light (Michaels) has been forced to
quit because of a sexual harrasment scandal.
Yes, the company is more stable than most of its previous competition,
but that ain't saying much. If Univel and/or Solaris make significant
inroads into SCO's market share, the company could be in real trouble.
It has yet to be seen how SCO copes with big-league competition
following a history of being the industry's 200-pound bully.
For the health of the industry, I want SCO to remain vibrant and
competitive. I want Univel and SCO to keep each other on their toes.
That's good for all of us.
But I believe that the ability to do this will require significant
attitude shifts in SCO's management. Is it capable? I sure hope so.
>SCO is my business, not my crusade.
Sorry, but "Personally, I wouldn't recommend anything but SCO" are the
words of a crusader. There are situations where SCO is clearly the
best solution and some situations where clearly it is not. The best
approach is an open mind and a familiarity with all the choices.
Christer Johansen
: ch...@huldra.tdata.no (Christer Johansen) writes:
: >I agree with Tony on this one. What matters for most commercial Unix end-users
: >is that they get a stable platform on wich to run their applications and that the
: >Unix they choose still exists years from now. It's of course very hard to look in
: >to the future, but SCO should have a fair chance.
: SCO is no more stable than Dell 2.2. And Dell 2.2 works and works. We
: use it for our internet connection running multiple feeds -- try
: PPP with SCO. Someone here mentioned that PPP with SCO is based on old
: standards, and doesn't like to talk to the outside world unless it is
: another SCO box running PPP. SCO also (according to the net) won't allow
: a port to run both SLIP and bidirectional communications -- only one or
: the other.
I didn't say SCO is more stable than Dell and I really don't know either.
What matters to me is that the OS I choose is stable and well supported
and that software companies are porting their software for that OS.
SCO has buildt a worldwide sales/support organization, you should
try to get local support on Dell Unix here in Norway!
As goes for SLIP and PP I don't know if what you say is correct and
I don't care. I don't use them anyway. And thats one of my points,
different OS's for different needs.
: SCO be around, maybe. But if they keep trimming back their staff, who
: knows?
As I said before, it's VERY hard to look into the future.
: >I'll have to say I don't understand why Mr. Snyder keeps on complaining about
: >SCO all the time. Every time he gets the change he seems to pick on them.
: Because SCO is so modified -- it's not real UNIX. SCO took Unix 3.2 and
: made it SCO. They symbolic links and long file names from SVR4 and added
: support with their 3.2 product calling it release 4.0 to confuse the issue.
Oooops, you're on thin ice. What exactly is "real UNIX".
Dell of course, rigth?
: They also removed support for alternate protocols like "g","e" and "f", then
: added them (mind you they were in the original code from AT&T) in one of
: their updates -- thus enabling them to charge for what had been available under
: other vendors products for the pervious 3 years.
I didn't say SCO is perfect and I don't have such feelings for SCO as you
seems to have for Dell. Maybe in a few years from now we're using Dell
Unix instead of SCO. My point is that one should choose the OS that
best suits their needs. It doesn't matter if the package reads Dell, SCO,
Interactive, Windows NT or something else. But rigth now SCO seems to be
the best choise, at least for us.
: For your information, Dell also sells $CO, and no, I don't own Dell stock.
Thank you, I didn't know that. Maybe I should go to Dell the next time
I buy a ODT license......
Bill Campbell
....LOTS DELETED TO GET TO MY POINT :-).
:>What matters is that SCO *advertises* and *sells*.
This is precisely the reason that I made the decision to use SCO
back in 1987 when I had to choose Unix for an Intel system. At
that time the major choices were Microport and Interactive.
Interactive had the reputation amongst the net.hackers of being
the technically superior product, but I was (am) more concerned
about having choices in peripherals, 3rd party software, and a
solid system for my commercial clients. It helped SCO that I had
been using Xenix on 68000 platforms for about 5 years so it would
be easier for me to use their Xenix product (and every time I ran
AT&T Unix I pissed and moaned about the things they didn't have
that made life easier on a Xenix system).
It seems to me, five years later, that this was a good decision.
As for the future -- it remains to be seen what Novell's
committment is to Unix. My impressions of SunSoft/Solaris aren't
too good (a large dose of arrogance on their part).
NEWS FLASH:
My posting was just interrupted by my friend Gary who is in the
process of installing the latest and greatest Solaris/
Intaractive from SunSoft (he's considering reselling it). He
asked ``Doesn't this thing have soft links?''. He wanted to
use symbolic links to put X11 stuff in another file system.
Then he went on about what a pain in the ass it was because
now he had to start from scratch with the 56 floppy disks
(tape isn't available yet). Each floppy is a mountable file
system and takes about 20 minutes to process. Gary has been
working with Sun, Solbourne, and various other Real Unix
systems for years and wasn't too impressed :-).
:That SCO sells as well as it has is partially due to the fact that there
:haven't been any competitors with a serious committment to UNIX *and*
:pockets deep enough to do the necessary marketing. SCO has essentially
:had the field to itself, at worst being a Gulliver in a market full of
:Liliputians.
:Understandably, it may be difficult to consider a UNIX which is poorly
:marketed and has little public profile. In technical terms and price
:terms, though, it *may* be superior.
:As has been stated in the editorial pages of UNIX Review, it's been easy
:so far for SCO to stare down the half-hearted UNIX marketing attempts by
:Interactive, Dell, Microport, ESIX, and even AT&T's own shot at Intel UNIX.
Where are these companies now? I would much rather be an SCO
customer who paid a bit more for a product that continues to
be supported than have an orphaned product and have to start over
from scratch.
I want my suppliers to stay in business and they have to make a
profit to do this. Too many technoids (and liberals :-) don't
seem to understand that businesses have to make money in order to
pay their salaries, the rent, and other necessities like that.
Few seem to realize the real costs of running a business,
particularly when the compare it to dealing a few PCs out of
their garages where they have no overhead costs.
When I managed Radio Shack Computer Centers between 1980 and 1983
the standard markup was about 100% (54% cost of goods) and a 12%
net profit was required to qualify for the leaders club (or
whatever they called it :-). Very few RSCCs made that 12% because
they had to pay for customer service reps, classroom instructors,
and repair technicians. Of course nobody wanted to pay for
customer support so I had to pay the CSR out of the profit on
sales.....
Bill
--
INTERNET: bi...@Celestial.COM Bill Campbell; Celestial Software
UUCP: ...!thebes!camco!bill 6641 East Mercer Way
uunet!camco!bill Mercer Island, WA 98040; (206) 947-5591
SPEED COSTS MONEY -- HOW FAST DO YOU WANT TO GO?
Bill Campbell
:In article <C0uD4...@news.rn.com> la...@news.rn.com (Larry Snyder) writes:
:> Because SCO is so modified -- it's not real UNIX.
:"Real UNIX" is as hard to pin down as "true christian" at times. Perhaps
:you could explain WHY it's not REAL UNIX when my AT&T bog-standard SVR3.2
:UNIX *is*?
I'll take a guess. Isn't it Real Unix because SCO has fixed many
of the bugs? Or maybe it's because SCO left /usr/spool/mail
where it's always been so they wouldn't break all the old
programs (just one example of ``improvements'' in Unix over the
years).
Karl Denninger
>
>>I agree with Tony on this one. What matters for most commercial Unix end-users
>>is that they get a stable platform on wich to run their applications and that the
>>Unix they choose still exists years from now. It's of course very hard to look in
>>to the future, but SCO should have a fair chance.
>
>SCO is no more stable than Dell 2.2. And Dell 2.2 works and works. We
>use it for our internet connection running multiple feeds -- try
>PPP with SCO. Someone here mentioned that PPP with SCO is based on old
>standards, and doesn't like to talk to the outside world unless it is
>another SCO box running PPP. SCO also (according to the net) won't allow
>a port to run both SLIP and bidirectional communications -- only one or
>the other.
That is correct. Also, an example from work:
12:50pm up 63 days, 17:07, 8 users, load average: 0.03, 0.00, 0.00
portal:/users/auspex/kdenning>
That's 63 days since we had to turn it off because our power in the room was
about to go out :-)
This machine is running DELL. So is that one (portal). Both are fine
pieces of work. Mine runs with 2 SCSI adapters, 4 disks, and two tape
drives as well as a real network (IP) to several other machines.
Mine goes offline more frequently, but then again I perform more maintenance
and "tinkering" :-)
>SCO be around, maybe. But if they keep trimming back their staff, who
>knows?
Or if they keep having women sue their officers for harassment. ;-)
>>I'll have to say I don't understand why Mr. Snyder keeps on complaining about
>>SCO all the time. Every time he gets the change he seems to pick on them.
>
>Because SCO is so modified -- it's not real UNIX. SCO took Unix 3.2 and
>made it SCO. They symbolic links and long file names from SVR4 and added
>support with their 3.2 product calling it release 4.0 to confuse the issue.
>
>They also removed support for alternate protocols like "g","e" and "f", then
>added them (mind you they were in the original code from AT&T) in one of
>their updates -- thus enabling them to charge for what had been available under
>other vendors products for the pervious 3 years.
In UUCP, yes they did that.
My primary problem with them is arrogance. SCO, a few years ago, had a
really nice XENIX product. I sold the product, and had software which ran
under it on the commercial market. When the time came to move to a Unix
architecture, SCO wanted me to shell out over $3,000 to "upgrade" my
operating system. I had already paid, over the last couple of years, more
than $1,000 in upgrade fees.
Considering that the replacement I finally settled on cost me $600 COMPLETE,
I told SCO to go stuff it.
Then, to add insult to injury, SCO's new "Unix" broke my major revenue
product, despite their claims of "complete Xenix compatibility". Well, that
was true, as long as you didn't depend on the authentication stuff (like the
password file, being able to have accounts without a password, etc).
COMPLETE compatibility? No. They refused to fix it, and in fact did not
until recently when the "security" options became real options (ie: you
could disable instead of "relax" them).
Therefore, I no longer support ANY SCO products. Period. I also strongly
advocate that people not support, use, buy, write for or otherwise get
revenue to them.
--
Karl Denninger (ka...@ddsw1.MCS.COM, <well-connected>!ddsw1!karl)
Data Line: [+1 312 248-0900]
Roger B.A. Klorese
>A lot of money and manpower went into ACE,
>resources that can be now considered down the toilet.
Take it from those of us who followed ACE from outside SCO... they didn't put
anywhere near the enormous investment into ACE you may think.
--
ROGER B.A. KLORESE +1 415 ALL-ARFF
rog...@unpc.QueerNet.ORG {ames,decwrl,pyramid}!sgiblab!unpc!rogerk
"Normal is not something to aspire to, it's something to get away from."
-- J. Foster
Bill Pottenger
>I agree with Tony on this one. What matters for most commercial Unix end-users
>the Unix they choose still exists years from now. It's of course very hard to
Couldn't help but put a plug in for an excellent PPP product, Morningstar
PPP (slip and cslip too). I'm using it under Ultrix, but it does run
under SCO. See sa...@morningstar.com.
Bill
Larry Snyder
>resources that can be now considered down the toilet. *AND*, at the worst
>possible time, the company's guiding light (Michaels) has been forced to
>quit because of a sexual harrasment scandal.
I heard that as well -- was that the original Michaels (or the son?). That
is bad press for SCO. Many folks will take notice these days and it will
end up costing sales in the long run (even though he is no longer actively
involved at SCO).
There are more pieces to Unix pie than even, and SCO will have to keep
current to continue to draw users. I was called the other day to work
on a local 200 user system that was based on SCO until last year -- when
they switched to ESIX SVR4. The SA expressed a concern with SCO and their
failure to move with current technology (SVR4) and instead adding new
features to an old product. His analogy was the VW bug -- they kept
adding and adding -- and finally it died a slow death.
Larry Snyder
>Oooops, you're on thin ice. What exactly is "real UNIX".
You are right there, without a definition, what is real Unix?
That is like, "who is God?" or "what is life" to many of us.
I was making reference to adding symbolic links and long
file names without increasing inodes to a 3.2 based product
which is non-standard.
Larry Snyder
>Where are these companies now? I would much rather be an SCO
>customer who paid a bit more for a product that continues to
>be supported than have an orphaned product and have to start over
>from scratch.
There is a fine line to this. A couple of months ago, folks were
complaining about the cost to upgrade a complete ODT 1.x system to
ODT 2.x. The cost of this upgrade was almost as much as another
vendors release of Unix which would have offered many extra features.
The bottom line, with ODT 2.x, you still would not be running the
latest release of Unix (maybe the latest release of SCO) since
regardless how you look at it, SCO is still a 3.2 based product.
And don't give me the SVR4 isn't stable. That's hogwash. I can
provide a list of machines that I am aware of which are running SVR4
that are stable and solid.
Larry Snyder
>Take it from those of us who followed ACE from outside SCO... they didn't put
>anywhere near the enormous investment into ACE you may think.
They sure liked to "sell" the fact that they did.
Stuart Lynne
>ch...@huldra.tdata.no (Christer Johansen) writes:
>
>>Oooops, you're on thin ice. What exactly is "real UNIX".
>
>You are right there, without a definition, what is real Unix?
>
>That is like, "who is God?" or "what is life" to many of us.
>
>I was making reference to adding symbolic links and long
>file names without increasing inodes to a 3.2 based product
>which is non-standard.
It really gets down to whether you are talking dejure or defacto
standards. AT&T (supposedly) sets the dejure UNIX standard in it's
current product (now USL, now Novell). SCO merely sets the market
defacto standard by selling more than anyone else.
At least in the Intel PC UNIX environment. In that market if
doesn't say SCO, it ain't UNIX.
If you're really in favour of dejure standards I'm sure you also
use X.400, FTAM, etc for your networking :-)
--
Stuart Lynne <s...@wimsey.com> ......................... UNIX Facsimile Software
Wimsey Information Technologies ................... moderator biz.sco.binaries
uucp login:nuucp passwd:nuucp .................... ftp.wimsey.com:~ftp/ls-lR.Z
PD Software for SCO UNIX .................... ftp.wimsey.com:~ftp/pub/wimseypd
Bill Campbell
:ch...@huldra.tdata.no (Christer Johansen) writes:
:>Oooops, you're on thin ice. What exactly is "real UNIX".
:You are right there, without a definition, what is real Unix?
Real Unix was Version 7 when everything was still bundled. Care
to go back to that?
:That is like, "who is God?" or "what is life" to many of us.
:I was making reference to adding symbolic links and long
:file names without increasing inodes to a 3.2 based product
:which is non-standard.
So what? The ONLY time I've been concerned with the 64K inode
limit was with news spool partitions. With symbolic links I can
simply break /usr/spool/news across several file systems once and
forget about it. It isn't like this is an ongoing problem.
This reminds me of my hot-rod days as a kid. I just had to have
the latest and greatest for my flat-head Ford. It didn't matter
that the Mallory Magspark ignition didn't offer a bit of
improvement over the stock ignition -- it sure did look neat!
Don't forget the chrome air cleaners which didn't do nearly as
good a job as the stock one.
Now I want a car that works when I turn the key, and a Unix that
does the same. SCO fits the bill pretty well for that. What are
the alternatives?
Dell -- I've had pretty bad luck with their support. Dell is
still primarily a hardware company.
Solaris/Interactive -- get real.
Microport, ESIX,....?????
SCO -- just keeps on ticking.
Fred Rump from home
>>SCO is my business, not my crusade.
>Sorry, but "Personally, I wouldn't recommend anything but SCO" are the
>words of a crusader. There are situations where SCO is clearly the
>best solution and some situations where clearly it is not. The best
>approach is an open mind and a familiarity with all the choices.
Again I disagree with Evan. I would rather say that if the SCO solution is not
the right one, we tell the customer about and part ways. I do not think
recommending this or that solution out of a choice of many does anyone any
good except the guy who comes after us and is really an expert on what we
recommended. No one is best at everything.
The Ford dealer does not also sell any other car in his shop. He would need a
whole other support organization to support some other car. It is not much
different in the software world. The idea is always to be the best at what you
do and not recommend the product of the moment.
Besides, we don't run into situations where SCO does not provide the solution
the customer is looking for. Perhaps we simply have a unique set of customers
who simply want to run their business and not be bothered by all this mine is
bigger (or better) than yours business. The proof of the pudding is always in
the satisfied customer.
Fred
--
W. Fred Rump office: fr...@COMPU.COM "A man's library is a sort of
26 Warren St. home: f...@icdi10.compu.com harem" - Emerson (1860)
Beverly, NJ. 08010
609-386-6846 bang:uunet!cdin-1!icdi10!fr
Fred Rump from home
>Therefore, I no longer support ANY SCO products. Period. I also strongly
>advocate that people not support, use, buy, write for or otherwise get
>revenue to them.
A real professional never knocks his competition. He rather attempts to
respect their view of the world as at least as valid as his own. A know-it-all
has no such compunction.
Karl Denninger
>ev...@telly.on.ca (Evan Leibovitch) writes:
>
>>resources that can be now considered down the toilet. *AND*, at the worst
>>possible time, the company's guiding light (Michaels) has been forced to
>>quit because of a sexual harrasment scandal.
>
>I heard that as well -- was that the original Michaels (or the son?). That
>is bad press for SCO. Many folks will take notice these days and it will
>end up costing sales in the long run (even though he is no longer actively
>involved at SCO).
I believe it was the chairman, who stepped down. It was reported in the
trade press. This is quite bad news for them.
Given their arrogance in the last few years, however, perhaps this will wake
them up. I would consider that this is all very bad for their chances to
pull off a successful public stock offering (which they have said they have
to do to keep going).
Franky, if this guy is half as guilty as these women make him out to be, I
hope they win a few million and really put the screws to SCO in the
process. The stuff that has been rumored in the trades was quite damaging.
Larry Snyder
>Given their arrogance in the last few years, however, perhaps this will wake
>them up. I would consider that this is all very bad for their chances to
>pull off a successful public stock offering (which they have said they have
>to do to keep going).
>Franky, if this guy is half as guilty as these women make him out to be, I
>hope they win a few million and really put the screws to SCO in the
>process. The stuff that has been rumored in the trades was quite damaging.
And the saga will continue. Can SCO absorb a large judgement? This was
really stupid of Michaels. Now SCO will get what they have coming..
Now is the time for the other Unix vendors to make a push into the market -
as the judgement (and press) can't be measured. SCO's massive amount of money
spent on advertising will mean nothing if this press continues.
Stew Ellis
>ev...@telly.on.ca (Evan Leibovitch) writes:
>>>SCO is my business, not my crusade.
>>Sorry, but "Personally, I wouldn't recommend anything but SCO" are the
>>words of a crusader. There are situations where SCO is clearly the
>>best solution and some situations where clearly it is not. The best
>>approach is an open mind and a familiarity with all the choices.
>Again I disagree with Evan. I would rather say that if the SCO solution is not
>the right one, we tell the customer about and part ways. I do not think
>recommending this or that solution out of a choice of many does anyone any
>good except the guy who comes after us and is really an expert on what we
>recommended. No one is best at everything.
>The Ford dealer does not also sell any other car in his shop. He would need a
>whole other support organization to support some other car. It is not much
>different in the software world. The idea is always to be the best at what you
>do and not recommend the product of the moment.
But your Ford dealer and every other dealer I know of in the US does sell
anything he can make a profit on whether he can support it or not. Every
dealer I know of sells used cars without regard to make (and sometimes
without regard to condition if they look right (e.g. have a 5 and a 4 in
there somewhere)). Most dealers will now lease you any brand car you want.
Car dealers have not generally been about service (I had to go to the Buick
dealer to get my Pontiac's power steering fixed under partial warranty after
the Pontiac dealer I bought it from said there was no program (he was only
reimbursed real hours rather than flatrate manual hours by GM)). Most
dealers will work on any kind of car you drive in. My Toyota dealer (owned
by the same guy who owned the Buick dealer above) had so little work to
perform on Toyotas that they were able to keep a proper staff of good
mechanics by keeping them busy working mostly on other brands of cars.
If you are looking for anologies|metaphors for the computer VAR or
consultant's situation, I hardly think the retail car sales is appropriate
or desirable. If you want to talk about the marketroid tricks by the
vendors then that might be more appropriate. The editor of Road & Track or
Car & Driver took his parents shopping for a new car a few years ago before
front wheel drive was as dominant as today. When they asked which end of
the car had the drive wheels at several dealers, the answers were randomly
distributed against the truth, except where the sales droid asked them which
end they were looking for, then the drive wheels were definitely on that end
regardless of the truth. "What kind of UNIX were you looking for? SysVr4?
Yeah we got UNIX with a V and a 4 in it. Fix you right up."
>Besides, we don't run into situations where SCO does not provide the solution
>the customer is looking for. Perhaps we simply have a unique set of customers
>who simply want to run their business and not be bothered by all this mine is
>bigger (or better) than yours business. The proof of the pudding is always in
>the satisfied customer.
>Fred
>--
>W. Fred Rump office: fr...@COMPU.COM "A man's library is a sort of
>26 Warren St. home: f...@icdi10.compu.com harem" - Emerson (1860)
>Beverly, NJ. 08010
>609-386-6846 bang:uunet!cdin-1!icdi10!fr
Of course the time to market of all of the commercial-quality SVR4 products
has been so long and the prices have risen so dramatically and they are so
severely unbundled these days in most instances and the companies are
falling apart so bad. The last person I recommended Dell to could not find
anybody at Dell who knew anything about how to order a topline Dell computer
with the tapedrive and UNIX preloaded. Esix's prices have gone through the
roof so that SCO ODT with developer bundle begins to look halfway reasonable
again. UnixWare is SOOO expensive by the time you have networking and
development tools, that a Sun IP{C|X} or Classic or LX with GCC is not that
much more expensive than a PeeCee and complete UNIX. Your mileage may vary
if you are not subject to academic discounts, but there are some VARs
putting together very nicely priced Sun bundles for the public. Of course
anyone with hacker inclinations is going to gravitate toward Linix or
BSD386, but I think this thread started out concerning commercial UNIXes.
As lots of people keep saying, it is going to be very sad when the son of
VMS is finally unleashed, years late with numerous incompatibilities and
UNIX still will have blown it because of overpricing and lack of market
focus.
--
___________________
R.Stewart(Stew) Ellis, Assoc.Prof., (Off)313-762-9765 / _____ ______
Humanities & Social Science, GMI Eng.& Mgmt. Inst. / / / / / /
Flint, MI 48504 el...@nova.gmi.edu /________/ / / / /
Shane Bouslough
: ev...@telly.on.ca (Evan Leibovitch) writes:
:
: >>SCO is my business, not my crusade.
:
: >Sorry, but "Personally, I wouldn't recommend anything but SCO" are the
: >words of a crusader.
:
: Again I disagree with Evan. I would rather say that if the SCO solution is
: not the right one, we tell the customer about and part ways.
:
: Besides, we don't run into situations where SCO does not provide the solution
: the customer is looking for.
Well, which is it? Do you "part ways" sometimes, or is SCO "always the
solution"?
: W. Fred Rump office: fr...@COMPU.COM "A man's library is a sort of
: 26 Warren St. home: f...@icdi10.compu.com harem" - Emerson (1860)
: Beverly, NJ. 08010
: 609-386-6846 bang:uunet!cdin-1!icdi10!fr
-Shane
--
Shane Bouslough | #include <stddisc.h>
sh...@sbcs.sunysb.edu | SVR4 - UNIX as nature intended
Larry Snyder
>again. UnixWare is SOOO expensive by the time you have networking and
>development tools, that a Sun IP{C|X} or Classic or LX with GCC is not that
>much more expensive than a PeeCee and complete UNIX. Your mileage may vary
Yep. I was thinking the same thing. The Sparc Classic
I believe is $3900 with the education discounts -- and
includes something like a 200 meg fast SCSI drive, with
16 megs of RAM and a large monitor.
Of course, the OS is included, as is an ethernet port.
Anthony P Lawrence
:
: >SCO is my business, not my crusade.
: Sorry, but "Personally, I wouldn't recommend anything but SCO" are the
: words of a crusader. There are situations where SCO is clearly the
: best solution and some situations where clearly it is not. The best
: approach is an open mind and a familiarity with all the choices.
I disagree somewhat. I am intimately familar with the caoabilities and
limitations of SCO. If I recommend it, it is because I know it will do
the job and that the client will be happy with it. If the situation is
such that I cannot feel comfortable, then I recommend nothing, and advise
the client to seek help elsewhere.
I cannot be knowledgable of SCO, OS/2, SUN, VMS, et al. Hell, I can barely
keep up with one vendor's product!
That is not to say I am not familiar with other products: I have SUN and
OS/2 manuals right here on my bookshelf, along with books on everything from
CP/M through Pick and now Windows NT. But I couldn't begin to make
a recommendation of Pick, for example, because I don't work with it.
I do work with SCO. Again, not because it's the most wonderful product in
the world, but because it is the market leader. Someone disdainfully
referred to "lemmings" in this context, but let's get real: SCO got to
be number one because they gave the small business customer an inexpensive
solution that worked.
Well, maybe things are changing. It looks to me like most of the fuss in
the Unix world right now concerns big customers with lots of interconnected
machines, probably with Dos and Windows and OS/2 thrown in. Personally,
I don't work with clients of that size. First of all, they usually have
their own department, and probably don't need me. Secondly, if they
do need an outside firm, they probably are too damn big for me to handle.
Finally, (and probably most important), I just don't get along with big
stuffy corporate structures (and they never seem to be too happy with me and
my dirty jeans and uncombed hair, either). How many of the same reasons apply
to SCO? I don't know. It may be that SCO should stick to it's knitting and
concentrate on the small business market. Somebody sure should: it's
the market that put SCO where it is today.
Larry Snyder
>Well, which is it? Do you "part ways" sometimes, or is SCO "always the
>solution"?
Let's face it, if you work on a percentage, and OS X is more, you will make
more selling X over Y, so it's in your best interest to sell X.
Fred Rump from home
>I was called the other day to work
>on a local 200 user system that was based on SCO until last year -- when
>they switched to ESIX SVR4. The SA expressed a concern with SCO and their
>failure to move with current technology (SVR4) and instead adding new
>features to an old product. His analogy was the VW bug -- they kept
>adding and adding -- and finally it died a slow death.
Care to swap facts about your 200 user ESIX system and our 200 user SCO
system?
I'm really curious to see a 200 user ESIX box. If it exists, maybe I'm missing
something.
I'm also curious to know why a 200 user user would switch operating systems in
midstream. He obviously must have thought the grass was greener on the other
side of the street for some reason. What was it?
fred
--
Fred Rump from home
>The bottom line, with ODT 2.x, you still would not be running the
>latest release of Unix (maybe the latest release of SCO) since
>regardless how you look at it, SCO is still a 3.2 based product.
So?
The NEXT OS is a much finer product than 4.2, doesn't mean the world is
running to it, does it? People tend to go with what other people use. That
tends to be more of a standard than what someone simply calls a standard.
Karl Denninger
>el...@nova.gmi.edu (Stew Ellis) writes:
>
>>again. UnixWare is SOOO expensive by the time you have networking and
>>development tools, that a Sun IP{C|X} or Classic or LX with GCC is not that
>>much more expensive than a PeeCee and complete UNIX. Your mileage may vary
>
>Yep. I was thinking the same thing. The Sparc Classic
>I believe is $3900 with the education discounts -- and
>includes something like a 200 meg fast SCSI drive, with
>16 megs of RAM and a large monitor.
>
>Of course, the OS is included, as is an ethernet port.
The Classic is $4k RETAIL!
Including a 15" (blech) color monitor, SCSI, RAM, keyboard, etc.
UnixWare either gets reasonable about pricing or they are going to get blown
to hell in this market. If I want to run PeeCee applications on that
Classic I can buy a Puzzleboard '386 SBUS card for another thousand or so
and have that too.
Now, for under $5k, I get a system which has TWO processors and will blow
the PC folks out of the water. Oh yeah, it has built in Ethernet too.
Intel machines are still cheaper, but not if the OS costs $3,000.
I would predict that the Sparc Classic means serious trouble for both Univel
AND SCO. Both are seriously overpriced and are in for some real
competition.
Roger B.A. Klorese
>Intel machines are still cheaper, but not if the OS costs $3,000.
Don't forget that you'll pay for a development environment on the SPARCclassic
now too...
Roger B.A. Klorese
>In article <C10z...@queernet.org> rog...@queernet.org (Roger B.A. Klorese) writes:
>>In article <C10rn...@ddsw1.mcs.com> ka...@ddsw1.mcs.com (Karl Denninger) writes:
>>>Intel machines are still cheaper, but not if the OS costs $3,000.
>>
>>Don't forget that you'll pay for a development environment on the SPARCclassic
>>now too...
>
OK. Make that "you'll need to pay for a vendor-supported development
environment on Sun now, just as you would on Intel-based UNIX, if that's
important to you"...
--
ROGER B.A. KLORESE +1 415 ALL-ARFF
rog...@unpc.QueerNet.ORG {ames,decwrl,pyramid}!sgiblab!unpc!rogerk
"Sometimes you wake up. Sometimes the fall kills you. And sometimes, when
you fall, you fly." -- N. Gaiman
Larry Snyder
>Care to swap facts about your 200 user ESIX system and our 200 user SCO
>system?
It's a ESIA 486/50 with 32 megs of RAM and a Eqinox cluster controller
running ESIX SVR4 3.0. It's running an internal accounting system..
>I'm also curious to know why a 200 user user would switch operating systems in
>midstream. He obviously must have thought the grass was greener on the other
>side of the street for some reason. What was it?
They wanted to run the latest OS, and got tired of SCO's rates for upgrades
which added features which had been available in other operating systems for
years.
Actually, several of their other offices have since installed ESIX SVR4. I
by no means am a ESIX fan, but I would choose ESIX over SCO. In my opinion,
SCO is on the lowest end of the spectrum. I (like others have expressed)
hope they loose this legal case and it costs them millions. SCO was stupid
to tolerate Michaels behavior even if he was (is) the president.
william E Davidsen
| Franky, if this guy is half as guilty as these women make him out to be, I
| hope they win a few million and really put the screws to SCO in the
| process. The stuff that has been rumored in the trades was quite damaging.
Let's put his wife and children in prison, too, if you want to punish
innocent people. Why should alledged behavior on the part of any one
ex-employee result in problems for the company? Do you believe in guily
by association?
Rumors are cheap, as my grandmother used to say "the paper holds
still." What I've read is that much of what is supposed to have happened
was years ago, and that the claim was made just before the IPO because
they thought Michaels would settle to avoid publicity. If that's true I
think it's pretty close to blackmail as moral distance is measured. The
fact that the suit is being defended instead of settled indicates that
the man feels he should defend himself, and was willing to give up the
company he built from nothing to go to court to protect his name.
--
bill davidsen, GE Corp. R&D Center; Box 8; Schenectady NY 12345
Keyboard controller has been disabled, press F1 to continue.
william E Davidsen
| And the saga will continue. Can SCO absorb a large judgement? This was
| really stupid of Michaels. Now SCO will get what they have coming..
The suit is against the person, not the company.
| Now is the time for the other Unix vendors to make a push into the market -
| as the judgement (and press) can't be measured. SCO's massive amount of money
| spent on advertising will mean nothing if this press continues.
Most of the press is being pretty careful to distinguish between the
man and the company. Pity usenet readers aren't that fair.
Roger B.A. Klorese
>>Intel machines are still cheaper, but not if the OS costs $3,000.
>
>Don't forget that you'll pay for a development environment on the SPARCclassic
>now too...
Oh, and you'll probably want an operating system. That right-to-use license
is $795.
And you may want to install it... another $600 or so for a CD-ROM drive
should get you there...
--
ROGER B.A. KLORESE +1 415 ALL-ARFF
rog...@unpc.QueerNet.ORG {ames,decwrl,pyramid}!sgiblab!unpc!rogerk
william E Davidsen
| That is correct. Also, an example from work:
|
| 12:50pm up 63 days, 17:07, 8 users, load average: 0.03, 0.00, 0.00
| portal:/users/auspex/kdenning>
|
| That's 63 days since we had to turn it off because our power in the room was
| about to go out :-)
I've run both SCO and Dell to uptimes of 1+year (my public access
system went 368 days on the first boot after I got the UPS - SCO). I
think you can get good stability from either.
| My primary problem with them is arrogance. SCO, a few years ago, had a
| really nice XENIX product. I sold the product, and had software which ran
| under it on the commercial market. When the time came to move to a Unix
| architecture, SCO wanted me to shell out over $3,000 to "upgrade" my
| operating system. I had already paid, over the last couple of years, more
| than $1,000 in upgrade fees.
When a vendor forces you to change to a new product, you should look
around.
|
| Considering that the replacement I finally settled on cost me $600 COMPLETE,
| I told SCO to go stuff it.
I'm still on good terms with SCO, although I use less expensive brands
when they make sense.
| Then, to add insult to injury, SCO's new "Unix" broke my major revenue
| product, despite their claims of "complete Xenix compatibility". Well, that
| was true, as long as you didn't depend on the authentication stuff (like the
| password file, being able to have accounts without a password, etc).
| COMPLETE compatibility? No. They refused to fix it, and in fact did not
| until recently when the "security" options became real options (ie: you
| could disable instead of "relax" them).
You now have a choice of four levels of security, which hasn't
remotely solved the problems, I agree. I want to have some short login
names to be compatible with hundreds of existing installations. I also
need short passwords for the same reason. If I turn down the security
enough to allow this I lose shadow passwords. And if I want multiple
logins to have the same UID, I better be willing to hand edit the passwd
file myself, because it's a "feature" to protect me from doing that.
william E Davidsen
| NEWS FLASH:
| My posting was just interrupted by my friend Gary who is in the
| process of installing the latest and greatest Solaris/
| Intaractive from SunSoft (he's considering reselling it). He
| asked ``Doesn't this thing have soft links?''. He wanted to
| use symbolic links to put X11 stuff in another file system.
| Then he went on about what a pain in the ass it was because
| now he had to start from scratch with the 56 floppy disks
| (tape isn't available yet). Each floppy is a mountable file
| system and takes about 20 minutes to process. Gary has been
| working with Sun, Solbourne, and various other Real Unix
| systems for years and wasn't too impressed :-).
Solaris has slinks, ISC V.3 doesn't.
Keith Barrett
You know what would be nice? If someone did a comprehensive list of all the
pro and cons of the various unix offerings, including price, and all
the various "complaints" brought up in this topic thread, and posted the
summary here so others could make up their own minds.
As I look at the UNIX offerings on intel platforms right now, I'm disappointed.
SCO is not true SVR4 (let alone V4.2) and is expensive, DELL only wants
to talk DELL (and again, isn't SVR4.2), most of the others are UNIX look-
alikes or "bare" offerings. The one with the greatest potentional is USL;
only time and pricing will determine that. Perhaps given time, the 386BSD
product will over take them all.
It's sad that all these come close, but "miss the mark" over issues that
could be rectified. If DELL started officially supporting non-DELL hardware and
went onward to V4.2 -- they'd have it. If SCO dropped their prices below
everyone else -- they's have it. If all the popular products get ported to
386BSD -- they might have it. How long do I have to wait until a good
personal UNIX offering exists? I've been trying to decide "whose" to buy for
the last month or so, and I'm no closer now than when I started.
I'm leaning toward 386BSD just because it's the lowest cost, or USL because
of its potential.
Stuart Lynne
I've complained bitterly about short names since I got the first version
with the security software. Didn't take me more than five minutes to
get it working :-)
Anyway with ODT 2.0 I havn't noticed any problems with short names. I
don't use the new user scripts though. I just add lines to the
passwd file.
Evan Leibovitch
>>at the worst
>>possible time, [SCO's] guiding light (Michaels) has been forced to
>>quit because of a sexual harrasment scandal.
>I heard that as well -- was that the original Michaels (or the son?). That
>is bad press for SCO. Many folks will take notice these days and it will
>end up costing sales in the long run (even though he is no longer actively
>involved at SCO).
I don't want to blow this into more than it is. I'm almost sorry I
brought the subject up.
People have human failings, some of them severe. I don't believe that Mr.
Michaels' actions, if proven true, should reflect whit upon all the other
SCO employees, some of whom haven't even met him.
I for one don't believe that the personal life of one man is going to
make people think worse of SCO the company. Michaels' problems are between
him, his alleged victims, and his Creator. I have no interest in
judging him (let alone his company), and I hope most other people out
there would feel the same way.
The reason I even brought the issue up was for its effect *within* SCO,
not to the company's public image. SCO needs some courageous and
inventive leadership right now; losing Michaels at this time, for
whatever reason, may hurt the company's abilities to amke the tough
decisions needed to deal with the new competition.
These people do exist within SCO. Michael Tilson, who I believe is a VP at
SCO, is quite capable. He was president of HCR in Toronto, Canada's only
distributor of Interactive UNIX, at the time SCO purchased HCR and turned
it into SCO Canada. During his reign at HCR, Tilson was peresented with the
first-ever "UNIX person of the year" award by UniForum Canada. Hardly a
lightweight. And I'm sure there are many others within SCO.
Point is, let's not blow this out of proportion. This is Usenet, not
Hard Copy. Anyone who chooses to buy or not to buy SCO based on the
sexual conduct of its chairman, is an idiot.
--
Evan Leibovitch, Sound Software Ltd., located in beautiful Brampton, Ontario
ev...@telly.on.ca / uunet!utzoo!telly!evan / (416) 452-0504
What's with all this multimedia stuff? Most vendors can't get *one* done right.
Karl Denninger
>>Intel machines are still cheaper, but not if the OS costs $3,000.
>
>Don't forget that you'll pay for a development environment on the SPARCclassic
>now too...
No I won't. GCC is on the Catalyst CDROM for free....
--
Stew Ellis
>el...@nova.gmi.edu (Stew Ellis) writes:
>>again. UnixWare is SOOO expensive by the time you have networking and
>>development tools, that a Sun IP{C|X} or Classic or LX with GCC is not that
>>much more expensive than a PeeCee and complete UNIX. Your mileage may vary
>Yep. I was thinking the same thing. The Sparc Classic
>I believe is $3900 with the education discounts -- and
Actually it is closer to $3300 with 15" monitor.
>includes something like a 200 meg fast SCSI drive, with
>16 megs of RAM and a large monitor.
>Of course, the OS is included, as is an ethernet port.
>--
>Larry Snyder internet: la...@gator.use.com
>keeper of the Gator uucp: uunet!gator!larry
Stew Ellis
No, you can download the Cygnus version of GCC from uunet. Many people
report it yields smaller faster code than the Sun value-added compiler.
Many people who have both use GCC in preference.
Stew Ellis
I don't have it in front of me, but I believe the right to use is free, but
the media and docs are $795. This is a little bit unclear. The CDROM is
becoming less and less of a problem, with lots of people buying them. There
are VARS who will preload the OS for free on top of a reasonable price for
the system. There was a big discussion on misc.forsale a couple of weeks
ago.
Fred Rump from home
>If you are looking for anologies|metaphors for the computer VAR or
>consultant's situation, I hardly think the retail car sales is appropriate
>or desirable.
Well, my car dealer is different. Perhaps it is just like in computers, there
are those who are instant experts at everything if you wave a dollar bill in
front of them and there are those who worry first about keeping their existing
customers and only add new ones as they feel secure about supporting them
also.
My analogy about cars may not hold everywhere but I think you may get the
drift of what I'm saying.
>As lots of people keep saying, it is going to be very sad when the son of
>VMS is finally unleashed, years late with numerous incompatibilities and
>UNIX still will have blown it because of overpricing and lack of market
>focus.
Overpricing would only hold in a mass market. UNIX is still not there. It
still requires some handholding and assistance from computer folks. It can
only enter a pricing war once that is no longer required.
Market focus? Depends on where your head is. Lots of folks talk of UNIX as a
workstation product. SCO made its mark in the small systems multi-user world.
SCO ODT is an attempt to join the more fruitful UNIX on every box scenario.
But guess where the money still comes from? So should SCO stick to its old
market focus or should they become another SUN and sell hardware to make big
bucks and give away the software? It is a question of market focus and it
seems to be there. Stick to what you know and cautiously move into more
lucrative markets.
fred
Fred Rump from home
>Fred Rump from home (f...@compu.com) wrote:
>:
>: Again I disagree with Evan. I would rather say that if the SCO solution is
>: not the right one, we tell the customer about and part ways.
>:
>: Besides, we don't run into situations where SCO does not provide the solution
>: the customer is looking for.
>Well, which is it? Do you "part ways" sometimes, or is SCO "always the
>solution"?
Both.
We part ways sometimes when the solution is not appropriate or previously
preordained. This leaves us to our specialty: SCO. But as I said, it is
better for us not to have a customer who assumes he is the expert and attempts
to spec the system. For him SCO may not be the right solution as he already
has his mind made up for whatever reason. (He may have read an article on SYS
V 4.2) It is best he hire the local hacker to configure his dream system for
him and go from there. Sometimes they stick their tail between their legs and
come back anyway.
fred
--
Fred Rump from home
>If I recommend it, it is because I know it will do
>the job and that the client will be happy with it. If the situation is
>such that I cannot feel comfortable, then I recommend nothing, and advise
>the client to seek help elsewhere.
>I cannot be knowledgable of SCO, OS/2, SUN, VMS, et al. Hell, I can barely
>keep up with one vendor's product!
That also applies to larger companies with lots of staff. There various
experts can be allocated to subsystems and specialties, but still, all within
one vendor's line. Just all the hardware idiosyncrasies can drive a company
crazy.
If I were a customer instead of a vendor, I would feel very comfortable with
somebody like Lawrence who told me he doesn't know everything but what he does
know, he knows extremely well. It is the mark of a real professional. I'm sure
he can command appropriate fees for his expertise and nobody minds paying
them. Results count!
>to SCO? I don't know. It may be that SCO should stick to it's knitting and
>concentrate on the small business market. Somebody sure should: it's
>the market that put SCO where it is today.
Amen.
Fred
Fred Rump from home
>sh...@cs.sunysb.edu (Shane Bouslough) writes:
>>Well, which is it? Do you "part ways" sometimes, or is SCO "always the
>>solution"?
>Let's face it, if you work on a percentage, and OS X is more, you will make
>more selling X over Y, so it's in your best interest to sell X.
Look guys! If making a living depended on the revenue from selling an OS we'd
all be dead by now from starvation. Regardless of SCO's prices it is a drop in
the bucket as this is not a mass market. Each system includes lots of
expertise that provides future management and support - that is what the
custome buys. He doesn't give a hoot about spending $200 more today, he
worries much more about running his business reliably tomorrow and being able
to call someone if something fails.
fred
--
Karl Denninger
>
>| Franky, if this guy is half as guilty as these women make him out to be, I
>| hope they win a few million and really put the screws to SCO in the
>| process. The stuff that has been rumored in the trades was quite damaging.
>
> Let's put his wife and children in prison, too, if you want to punish
>innocent people. Why should alledged behavior on the part of any one
>ex-employee result in problems for the company? Do you believe in guily
>by association?
No, I believe in a company being spoken for, and its ethical tone being set
by, its top corporate officers. Which the accused was. About as top as you
can get in fact.
It can be argued that the "personal" representative of any corporation is
its top brass. They certainly are personally liable for serious mistakes
undertaken in the name of the firm; environmental problems, legal criminal
liability if someone is negligently killed (and they know about it), etc.
If anyone can be said to be "the company", as an "Inc'd" firm is a
fictional person, the corporate officers are.
Therefore, since he was in that position, he not only speaks for the firm,
but he <is>, in no small part, the firm. Therefore, what Michaels does in
an official capacity is inseparable from the name "SCO".
> Rumors are cheap, as my grandmother used to say "the paper holds
>still." What I've read is that much of what is supposed to have happened
>was years ago, and that the claim was made just before the IPO because
>they thought Michaels would settle to avoid publicity. If that's true I
>think it's pretty close to blackmail as moral distance is measured. The
>fact that the suit is being defended instead of settled indicates that
>the man feels he should defend himself, and was willing to give up the
>company he built from nothing to go to court to protect his name.
We'll see, and the courts will decide. Note that above I said <IF> he is
proven to be guilty. Until that time I'd not hang them out to dry -- but
if it is shown that he did indeed commit these acts then the reputation of
the company is tarnished IMHO.
That he has chosen to fight is not an issue. If he were to capitulate then
he would be admitting guilt, even though the "legal" definition wouldn't
work out that way. His <only> method of defense is to fight. That much
I'll give you.
This is, in my mind, much more serious than, say, dealing drugs out of his
home would have been. The reason is that it took place (if true) in an
official capacity, and as such is part of the firm's business AND its
statement on ethics from a corporate perspective.
Karl Denninger
>In article <C10z...@queernet.org> rog...@queernet.org (Roger B.A. Klorese) writes:
>>In article <C10rn...@ddsw1.mcs.com> ka...@ddsw1.mcs.com (Karl Denninger) writes:
>>>Intel machines are still cheaper, but not if the OS costs $3,000.
>>
>>Don't forget that you'll pay for a development environment on the SPARCclassic
>>now too...
>
>Oh, and you'll probably want an operating system. That right-to-use license
>is $795.
I believe the Classics come with a limited right-to-use (will have to check
again).
And you need only ONE CDROM drive for 100 classics; once you have the first
one booted, you can use the CDROM on it to load the rest. Just as with
DELL, or most other reasonable operating systems, you need a tape drive --
on ONE machine.
The beauty of networks.
Larry Snyder
>No, you can download the Cygnus version of GCC from uunet. Many people
>report it yields smaller faster code than the Sun value-added compiler.
>Many people who have both use GCC in preference.
So for $3300 one can obtain a Sparc based Sun and after obtaining gcc,
they will have a fully functional system that can compete head on with
the Intel based hardware and operating systems, correct? Gosh, last
time I checked, for one to buy a copy of everything included in the base
Dell package from SCO the cost was right around $3300 (discounted price
from Tech Data in Clearwater, Florida).
The Sparc Classic will run all Sparc software?
william E Davidsen
| I don't have it in front of me, but I believe the right to use is free, but
| the media and docs are $795. This is a little bit unclear. The CDROM is
| becoming less and less of a problem, with lots of people buying them. There
| are VARS who will preload the OS for free on top of a reasonable price for
| the system. There was a big discussion on misc.forsale a couple of weeks
| ago.
You're right, and I do have it in front of me. In the column labeled
"Soliris license included" is "Solaris 2.1, SunOS, ONC, OpenWindows V.3,
DeskSet and OPEN LOOK." As nearly as I can type the bizarre
caPitAlizatIOn.
Also: 6k cache, 8bit audio, 16MB RAM, 15" monitor.
william E Davidsen
| We'll see, and the courts will decide. Note that above I said <IF> he is
| proven to be guilty. Until that time I'd not hang them out to dry -- but
| if it is shown that he did indeed commit these acts then the reputation of
| the company is tarnished IMHO.
The courts are immaterial now, even if they decide the whole thing was
overblown, his reputation is gone, and he's lost the company he spent
decades building.
Roger B.A. Klorese
>No, you can download the Cygnus version of GCC from uunet. Many people
>report it yields smaller faster code than the Sun value-added compiler.
>Many people who have both use GCC in preference.
I'm sure it does. And in an academic or research environment, that
judgment makes sense. In a commercial product build environment,
the decision of cost-of-license vs. cost-of-staff-to-maintain-tools
is an important one, unless one purchases outside support for free
tools.
--
ROGER B.A. KLORESE +1 415 ALL-ARFF
rog...@unpc.QueerNet.ORG {ames,decwrl,pyramid}!sgiblab!unpc!rogerk
Roger B.A. Klorese
>I don't have it in front of me, but I believe the right to use is free, but
>the media and docs are $795. This is a little bit unclear.
You're probably right.
>The CDROM is
>becoming less and less of a problem, with lots of people buying them.
That doesn't mean they don't have a cost, only that people are more
inclined to buy them.
>There
>are VARS who will preload the OS for free on top of a reasonable price for
>the system. There was a big discussion on misc.forsale a couple of weeks
>ago.
True. Don't forget to add it to the system cost, however.
David Mason
And besides, what does this have to do with UNIX. The discussions in these
groups are supposed to be technically inclined, not given to malicious,
unproven rumours. According to these people, he's guilty before proven
anything and SCO is about to go down in flames because of an employee rather
than the end product.
There must be a raging discussion going on about this issue in some newsgroup
or another, this isn't the right place.
Now if someone wants to attack SCO on their percieved technical backwardness,
that's fine. But this kind of thing is really pathetic.
--
v...@zooid.guild.org -->> ZOOiD Public Access Usenet 416 322-7876
David Mason
>everyone else -- they's have it. If all the popular products get ported to
>386BSD -- they might have it. How long do I have to wait until a good
>personal UNIX offering exists? I've been trying to decide "whose" to buy for
>the last month or so, and I'm no closer now than when I started.
There's no third party support for 386BSD. Are there any intelligent multiport
serial ports that support it? That's crippling.
>
>I'm leaning toward 386BSD just because it's the lowest cost, or USL because
>of its potential.
I hate to say it, but Consensys is currently selling v4.2. They have a pretty
bad track record in the past but they appear to be making an effort to get
better. I used to call and get an irate sounding support person right away,
now I get a receptionist, then a first level support person, then a call back
from a techie. They actually called back too (that's rare in this city). Their
v4.2 is pretty complete, but they were first on the market so it's pretty bare
bones (no value added stuff like extra drivers) and a few bugs, though I (and
my guru friend - I'm not guru-level by any means) are pretty impressed with
SVR4.2 overall.
I wondered if maybe USL gave them a special price so they could be the "guina
pigs", since USL is releasing a few months after Consensys has been on the
market, testing out the waters. Not totally unplausible I think..
Anyways I really can't say with 100% certainty that Consensys is ok to deal
with now, since I haven't spoken to them that much, but they do seem to be
making an effort to improve, and it seems to be working.
Karl Denninger
>el...@nova.gmi.edu (Stew Ellis) writes:
>
>>No, you can download the Cygnus version of GCC from uunet. Many people
>>report it yields smaller faster code than the Sun value-added compiler.
>>Many people who have both use GCC in preference.
>
>So for $3300 one can obtain a Sparc based Sun and after obtaining gcc,
>they will have a fully functional system that can compete head on with
>the Intel based hardware and operating systems, correct? Gosh, last
>time I checked, for one to buy a copy of everything included in the base
>Dell package from SCO the cost was right around $3300 (discounted price
>from Tech Data in Clearwater, Florida).
$4200 for the machine, plus 795 for OS media and docs (includes everything,
you don't have to buy this if you have access to media, as the right-to-use
is included -- just not media and docs) plus $600 or so for the CDROM drive
to load it if you don't. If you already have a network with one or more
Suns on it you need neither of the above -- you can boot from the network,
adn load the OS over it as well.
>The Sparc Classic will run all Sparc software?
Correct.
Therefore, you need to be able to get the Intel machine, plus hi-res
graphics and monitor, for $1700.00. This is not impossible, but its not
easy either.
Don't forget that the Sparc includes Ethernet and sound ports in that base
price, color graphics, and a color monitor. The Intel machine includes none
of these in the base price. Further, the Sparc has the SCSI on the
motherboard, and two serial ports which work.
SCO has a serious value problem.
Karl Denninger
>Regardless of SCO's prices it is a drop in
>the bucket as this is not a mass market. Each system includes lots of
>expertise that provides future management and support - that is what the
>custome buys. He doesn't give a hoot about spending $200 more today, he
>worries much more about running his business reliably tomorrow and being able
>to call someone if something fails.
Correct. Nonetheless, $100 saved in direct costs you have to pass through
is $100 your customer doesn't pay.
If you are going up against me in a deal, I'm $2200 cheaper, and the customer
believes both of us offer equal value added, you're going to lose.
If the only reason you are $2200 more expensive is that you made a decision
to use an operating system which is $2200 more expensive, and it doesn't
bring tangible benefits worth that $2200, you deserve to lose.
Every time.
My job as an independant is to evaluate the available materials and come up
with the best possible solution. That evaluation not only includes costs,
but product stability and life-cycle costs.
BTW, the smart shopper considers the value of CPU boards and the like to be
near zero the minute you pay the bill. Same with Operating Systems. Even
SCO has proven this (they wanted twice what Xenix cost for me to upgrade
Xenix to Unix a few years ago, and I could have bought it over from a
discount house for the "upgrade" price)
Therefore, the issue is what you get from the system and OS <now>. On that
basis, SCO loses. With Sun entering this market, they're going to lose
badly.
Fred Rump from home
>>Care to swap facts about your 200 user ESIX system and our 200 user SCO
>>system?
>It's a ESIA 486/50 with 32 megs of RAM and a Eqinox cluster controller
>running ESIX SVR4 3.0. It's running an internal accounting system..
So they have 200 users doing accounting? Must be one hell of a large company.
GM maybe? But I doubt if even they have 200 accountants pounding away at
their keyboards.
And how does this Equinox cluster controller look? How did they hook 200
users onto their system. I mean physically.
>>I'm also curious to know why a 200 user user would switch operating systems
in
>>midstream. He obviously must have thought the grass was greener on the
other
>>side of the street for some reason. What was it?
>They wanted to run the latest OS, and got tired of SCO's rates for upgrades
>which added features which had been available in other operating systems for
>years.
So this huge corporation with 200 accountants didn't want to spend 200 bucks
for an upgrade to the latest SCO OS? They probably spent more than that on
copy machine paper that day. There's obviously another agenda here we are not
aware of.
>I (like others have expressed)
>hope they loose this legal case and it costs them millions. SCO was stupid
>to tolerate Michaels behavior even if he was (is) the president.
I keep wondering what this has to do with their operating system. Shall we
talk about the weather in Montana too?
Fred
Larry Snyder
>So they have 200 users doing accounting? Must be one hell of a large company.
>GM maybe? But I doubt if even they have 200 accountants pounding away at
>their keyboards.
They are keeping track of judgements and court activity on behalf of the
county (they are in the county courthouse)
>And how does this Equinox cluster controller look? How did they hook 200
>users onto their system. I mean physically.
boxes bolted on large pieces of plywook with a nest of wires (tie wrapped
of course)..
Larry Snyder
>SCO has a serious value problem.
I ditto this. Considering what you get with a loaded SCO system,
one can get a Sun Sparc Classic with 200 meg drive and complete OS
which will run circles around SCO. Likewise, take that 4100
retail and you can get the more product (and value) from the SVR4
vendors (Dell and the new ESIX product comes to mind) - basically
at 1/3 the cost of SCO. These SVR4 products are just as reliable
and support the latest industry standards -- and are not based on
an OS that is 5 years old. I would bet that if you were to place
packages side-by-side you would find that you still get more value
with Dell (for example) than SCO even if you spend the 4K retail
on SCO.
Robert Withrow
rog...@queernet.org (Roger B.A. Klorese) writes:
| And in an academic or research environment, that
| judgment makes sense. In a commercial product build environment,
| the decision of cost-of-license vs. cost-of-staff-to-maintain-tools
| is an important one, unless one purchases outside support for free
| tools.
One must factor into this the quality of the tool in question, and
the quality of the ``support'' you are getting. If you
assume that a tool you buy is perfect and the tool you don't buy is
imperfect the decision is obvious. This is *never* the case. What *is*
the case is that the tool you buy will have flaws, and your staff will
be spending their time on the phone (generally on hold), and sitting
on their thumb, waiting for the vendor to 1) understand what you are
talking about, 2) denigh that it is their problem, 3) blame someone else,
4) Admit that there is a problem, but it will be fixed in the next
release, 5) agree to send you the fix ahead of time (P=0.01), and 6)
go through 7 revs of the fix. This is what I call the ``Myth of support''.
Sure, it will cost you a day or two of programmer time to fix the problem
in your ``unsupported'' tool, but then your staff will be back to work
instead of floundering around waiting for that promised ``support''.
And, in the best scenario, if you have the sources you can attack the problem
both ways (using, say, cygnus *and* your own staff), and will be assured
of an even faster resolution of the problem.
--
Robert Withrow, Tel: +1 617 598 4480, Fax: +1 617 598 4430, Net: wi...@rwwa.COM
R.W. Withrow Associates, 21 Railroad Ave, Swampscott MA 01907-1821 USA
william E Davidsen
| of these in the base price. Further, the Sparc has the SCSI on the
| motherboard, and two serial ports which work.
Let's say my experience has not been all that great with the Sun
serial ports, which may explain why they sell an enhanced driver package
and enhanced serial card, which I believe are recommended if the
aggregate serial datarate is >9600 baud.
We were going to run a 56k line for slip and Sun suggested using a
terminal concentrator to come in with TCP.
The Sun ports are not any better than 16550s (Sun uses Zylog 8350s)
and even their extra price driver is no better than FAS. I don't see
much to recommend buying a Sun for the serial ports, and I surely can't
buy 4-port S-bus cards for <$100!
william E Davidsen
| BTW, the smart shopper considers the value of CPU boards and the like to be
| near zero the minute you pay the bill. Same with Operating Systems. Even
| SCO has proven this (they wanted twice what Xenix cost for me to upgrade
| Xenix to Unix a few years ago, and I could have bought it over from a
| discount house for the "upgrade" price)
And you wind up with a legal copy of Xenix which has some resale
value.
| Therefore, the issue is what you get from the system and OS <now>. On that
| basis, SCO loses. With Sun entering this market, they're going to lose
| badly.
Maybe. It depends on the price points. Having Solaris available for PC
will make a lot of corporate buyers consider PC based UNIX. When then
see the price of Solaris they might well decide to go with SCO, because
SCO has spent the bucks gaining name recognition.
SCO might lose market share while increasing revenues, because of
increases in the market size. The old "a rising tide lifts all boats"
theory.
What would hurt all current vendors is the possibility of cheap V.4.2
from USL. Imagine $99 runtime, or $200 complete systems, selling in vast
quantities, pushed by Novell's marketing and name recognition. USL has
aimed for fairly high income per customer, but under new management a
mass market UNIX could happen, making profit from volume.
Of course Novell coulld raise prices and make a great unit profit
until everyone switches to NT. I admit it, I grabbed a copy of NT for
$69, because at that price people will try things.
Karl Denninger
>In article <ellis.727413624@nova> el...@nova.gmi.edu (Stew Ellis) writes:
>>No, you can download the Cygnus version of GCC from uunet. Many people
>>report it yields smaller faster code than the Sun value-added compiler.
>>Many people who have both use GCC in preference.
>
>I'm sure it does. And in an academic or research environment, that
>judgment makes sense. In a commercial product build environment,
>the decision of cost-of-license vs. cost-of-staff-to-maintain-tools
>is an important one, unless one purchases outside support for free
>tools.
When was the last time any of you folks out there actually got a compiler
bug fixed in, oh, say, SCO's compilers? I remember bitching about some
problems for over two years in their Xenix compiler -- while on support --
and <never> receiving a fix. So much for "official support". Trust me,
their lack of producing a fix won't stop them from charging you the annual
or upgrade maintenance fee.
My life has been spent working around compiler problems in people's
products. Compilers that produce bad code. Optimizers which mangle code
and cause core dumps. Linkers that won't.
Usually there is a work-around. Not always, but usually.
"CC" lately on the V.4 machines has actually gotten good enough to use
seriously. Gcc is rumored to be of the same general quality, and to (in
some cases) produce better optimized objects.
My basic premise in the compiler area is that it either is good enough to
produce production code or it is not. If I have to work around a few
problems so be it. If I have <source> to the compiler, which with GCC I do,
then I can at least fix the problems myself.
That puts me in a position of being ahead of the "officially supported"
crowd.
At least if I pay for someone's compiler its unbundled so they know
<exactly> what the support money is supposed to buy. If they then do not
fix the problems I can likely sue to recover that support money (and
possibly the cost of the product) and/or switch to someone else's.
Shane Bouslough
: Yeah, really huge cost. Just get the Solaris sampler CD, which
: includes a Solaris version (complete) of the GNU compiler. Costs you
: all of $0.
Please tell me where I can get a CDROM drive for $0 so I can put my
Solaris sampler CD for $0 in it. Thanks.
: Fred....@dseg.ti.com - I don't speak for others and they don't speak for me.
-Shane
--
Shane Bouslough | #include <stddisc.h>
sh...@sbcs.sunysb.edu | "Follow your bliss" - Joseph Campbell
Roger B.A. Klorese
>In article <C149A...@queernet.org>,
> rog...@queernet.org (Roger B.A. Klorese) writes:
>
>| And in an academic or research environment, that
>| judgment makes sense. In a commercial product build environment,
>| the decision of cost-of-license vs. cost-of-staff-to-maintain-tools
>| is an important one, unless one purchases outside support for free
>| tools.
>
>One must factor into this the quality of the tool in question, and
>the quality of the ``support'' you are getting. If you
>assume that a tool you buy is perfect and the tool you don't buy is
>imperfect the decision is obvious. This is *never* the case. What *is*
>the case is that the tool you buy will have flaws, and your staff will
>be spending their time on the phone (generally on hold), and sitting
>on their thumb, waiting for the vendor to 1) understand what you are
>talking about, 2) denigh that it is their problem, 3) blame someone else,
>4) Admit that there is a problem, but it will be fixed in the next
>release, 5) agree to send you the fix ahead of time (P=0.01), and 6)
>go through 7 revs of the fix. This is what I call the ``Myth of support''.
Mostly accurate.
>Sure, it will cost you a day or two of programmer time to fix the problem
>in your ``unsupported'' tool, but then your staff will be back to work
>instead of floundering around waiting for that promised ``support''.
>
>And, in the best scenario, if you have the sources you can attack the problem
>both ways (using, say, cygnus *and* your own staff), and will be assured
>of an even faster resolution of the problem.
Again, absolutely true.
However: many of us don't have toolsmiths. For instance: my company is
a source licensee of SVR4 flavors from USL (we're a technology provider).
When we find problems in the scope of our expertise, the kernel, we
fix it, and send the fixes to USL, when possible. But when we encounter
a compiler bug, it's different. Even though we have a compiler tools
group, their development platform is not SVR4-on-Intel, and they don't
have the familiarity with the particular compiler, or the time to divert
from their projects, to hack on the SVR4-for-Intel compiler. So what
does our OS group do? Turn off optimization, or try workarounds... the
fact is that not many companies have toolsmiths who are capable of
doing compiler work, *plus* OS work, *plus* utilities work... plus the
products they are there to deliver.
So, again: given the possibility of assistance, it's a good model. But
given the (fairly common) situation of no internal resources to work on
tools... well, it's a good thing there's a Cygnus, and the rest of the free
software community, out there to help.
Stew Ellis
>fred j mccall 575-3539 (mcc...@mksol.dseg.ti.com) wrote:
>: Yeah, really huge cost. Just get the Solaris sampler CD, which
>: includes a Solaris version (complete) of the GNU compiler. Costs you
>: all of $0.
>Please tell me where I can get a CDROM drive for $0 so I can put my
>Solaris sampler CD for $0 in it. Thanks.
>: Fred....@dseg.ti.com - I don't speak for others and they don't speak for me.
>-Shane
The exact selfsame distribution is in ftp.uu.net:/vendor/cygnus, both binary
and source tar files.
Evan Leibovitch
>I can't help but think that when a hacker is reduced to pushing a
>commercial UNIX -- ANY commercial UNIX -- instead of Linux or 386BSD,
>she may win the battle, but she will have lost the war.
This particular "war" is not worth winning.
--
Evan Leibovitch, Sound Software Ltd., located in beautiful Brampton, Ontario
ev...@telly.on.ca / uunet!utzoo!telly!evan / (416) 452-0504
What's with all this multimedia stuff? Most vendors can't get *one* done right.
Martin Cracauer
>The Sparc Classic will run all Sparc software?
SPARC Classis can only run Solaris 2. That means, that most of
the old Software will not run for a time (freeware espcially).
And note, that a SPARC 2 or SPARC classic is much slower than a
66MHz-486 on interger-operations.
--
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Martin Cracauer <crac...@wavehh.hanse.de>, WAVEDATA, Norderstedt, Germany
Robert Withrow
sh...@cs.sunysb.edu (Shane Bouslough) writes:
| fred j mccall 575-3539 (mcc...@mksol.dseg.ti.com) wrote:
|
| : Yeah, really huge cost. Just get the Solaris sampler CD, which
| : includes a Solaris version (complete) of the GNU compiler. Costs you
| : all of $0.
|
| Please tell me where I can get a CDROM drive for $0 so I can put my
| Solaris sampler CD for $0 in it. Thanks.
I will, just as soon as you can tell me where to get a free computer
to put the free drive into.
And what about a free house with free electricity to put the free computer
into?
Cheesh.
There are enough free or nearly free CDROMS with stuff on them that
the cost of a drive pays for itself in reduced network costs in a
very short time.
(Of course, I know that those of you in the .edu domain believe or
are taught to believe that everything is either free or should be free.
So I suggest you ask your professor for the drive.)
Warren Tucker
: davi...@crd.ge.com (bill davidsen) writes:
: >Rumors are cheap......
: ... this kind of thing is really pathetic.
Absolutely. The speculation makes a soap opera out
of a serious matter. The spew belongs in the sewer.
--
------------------------------------------------------------------------------
Warren Tucker (404)587-5766 n4hgf!wht or w...@n4hgf.Mt-Park.GA.US
All readers cannot be leaders, but all leaders must be readers. - Harry Truman
Karl Denninger
>In article <C14Kv...@ddsw1.mcs.com>, ka...@ddsw1.mcs.com (Karl Denninger) writes:
>
>| BTW, the smart shopper considers the value of CPU boards and the like to be
>| near zero the minute you pay the bill. Same with Operating Systems. Even
>| SCO has proven this (they wanted twice what Xenix cost for me to upgrade
>| Xenix to Unix a few years ago, and I could have bought it over from a
>| discount house for the "upgrade" price)
>
> And you wind up with a legal copy of Xenix which has some resale
>value.
But today, not much.
>| Therefore, the issue is what you get from the system and OS <now>. On that
>| basis, SCO loses. With Sun entering this market, they're going to lose
>| badly.
>
> Maybe. It depends on the price points. Having Solaris available for PC
>will make a lot of corporate buyers consider PC based UNIX. When then
>see the price of Solaris they might well decide to go with SCO, because
>SCO has spent the bucks gaining name recognition.
No, you're missing the point.
I suspect that the Intel Unix systems may be doomed. The introduction of
low-cost Sparc-based systems could eat Intel's lunch. The current offerings
(and not just from Sun) are nice, fast, <cheap> and expandable through the
network.
The other interesting thing is that "board level" products are now available
for the "build your own clone" group. The prices on these are coming down
<fast>. Again, they're much faster than the 486 systems of today.
Then there is the application base. Sparc systems have a reasonably good
one. PC Unix systems have a horrible one.
> What would hurt all current vendors is the possibility of cheap V.4.2
>from USL. Imagine $99 runtime, or $200 complete systems, selling in vast
>quantities, pushed by Novell's marketing and name recognition. USL has
>aimed for fairly high income per customer, but under new management a
>mass market UNIX could happen, making profit from volume.
That would be WONDERFUL. It would also hurt NT <IF> Novell and USL do it
right and put a good GUI-based front end on this thing for installation and
administration. They would have to make it a "product for the masses", but
this isn't THAT hard to do with the underlying products of today.
> Of course Novell coulld raise prices and make a great unit profit
>until everyone switches to NT. I admit it, I grabbed a copy of NT for
>$69, because at that price people will try things.
Yep.
Jim Shirreffs
me to port our application to NT. I love Unix, but....
jim shirreffs
Jim Shirreffs
>Correct.
Are you sure?
>Therefore, you need to be able to get the Intel machine, plus hi-res
>graphics and monitor, for $1700.00. This is not impossible, but its not
>easy either.
>Don't forget that the Sparc includes Ethernet and sound ports in that base
>price, color graphics, and a color monitor. The Intel machine includes none
>of these in the base price. Further, the Sparc has the SCSI on the
>motherboard, and two serial ports which work.
>SCO has a serious value problem.
One good thing about the Intel platform is it ability to boot, DOS, OS2, NT or
Unix. And software is cheap.
jim shirreffs
Fred Rump from home
>In article <1993Jan18....@compu.com> f...@compu.com (Fred Rump from home) writes:
>>custome buys. He doesn't give a hoot about spending $200 more today, he
>>worries much more about running his business reliably tomorrow and being able
>>to call someone if something fails.
>Correct. Nonetheless, $100 saved in direct costs you have to pass through
>is $100 your customer doesn't pay.
I believe the subject was upgrade cost. It was deemed to be more expensive
than that of other vendors. Therefore the $200 which I mentioned.
>If the only reason you are $2200 more expensive is that you made a decision
>to use an operating system which is $2200 more expensive, and it doesn't
>bring tangible benefits worth that $2200, you deserve to lose.
>Every time.
Why not pick $10,000 or any other arbitrary number. NONE of our customers EVER
buys a development system. They simply buy the basic run-time and I have no
idea where a $2200 DIFFERENCE can be established for me to lose anything.
EVER. Users and developers are different animals. Users simply wish to get
their own work out. Developers worry about how that can be done most
efficiently and then sell their expertise to users for a price. That's the way
the market has worked since the first hunter sold his extra rabbit for some
veggies his neighbor had.
It is called a system of resource maximization.
>My job as an independant is to evaluate the available materials and come up
>with the best possible solution. That evaluation not only includes costs,
>but product stability and life-cycle costs.
I thought you worked for the videocart company doing whatever they do?
Fred Rump from home
>When was the last time any of you folks out there actually got a compiler
>bug fixed in, oh, say, SCO's compilers? I remember bitching about some
>problems for over two years in their Xenix compiler -- while on support --
>and <never> receiving a fix. So much for "official support". Trust me,
>their lack of producing a fix won't stop them from charging you the annual
>or upgrade maintenance fee.
>My life has been spent working around compiler problems in people's
>products. Compilers that produce bad code. Optimizers which mangle code
>and cause core dumps. Linkers that won't.
For once we agree. The problem with the SCO (read MS) compiler is that SCO has
little control over it. It least that was the story from the folks who worked
in that area. The long lists of 'things' just kept getting longer. GNU was the
answer to the problem. I'm not sure where the real fault lies here or why MS
can't do a better compiler. I mean the have the horses and resources.
In any case, I'm not so quick to throw this problem into SCO's lap. As far as
I know they were and are more of an integrator that an original developer. I
don't think it was in their scope to write YAC (yet another compiler).
Fred Rump from home
>They are keeping track of judgements and court activity on behalf of the
>county (they are in the county courthouse)
The state is Indiana. And can I check with 'any' county? Each county has a 200
ESIX box doing child support work?
>>And how does this Equinox cluster controller look? How did they hook 200
>>users onto their system. I mean physically.
>boxes bolted on large pieces of plywook with a nest of wires (tie wrapped
>of course)..
Sounds like a real rube-goldberg setup and not very professional.
Fred Rump from home
>ka...@ddsw1.mcs.com (Karl Denninger) writes:
>>SCO has a serious value problem.
>I ditto this.
>at 1/3 the cost of SCO. These SVR4 products are just as reliable
I hate to keep bringing up reality here. Where do these numbers come from all
the time? Who makes them up?
SCO UNIX costs $1295. Where do I buy the same thing for $431 retail? That is
1/3 the cost isn't it?
>packages side-by-side you would find that you still get more value
>with Dell (for example) than SCO even if you spend the 4K retail
>on SCO.
Again, what are you buying that needs a $4000 expenditure and why?
Shane Bouslough
: In article <1993Jan20.1...@sbcs.sunysb.edu>,
: sh...@cs.sunysb.edu (Shane Bouslough) writes:
: | fred j mccall 575-3539 (mcc...@mksol.dseg.ti.com) wrote:
: |
: | : Yeah, really huge cost. Just get the Solaris sampler CD, which
: | : includes a Solaris version (complete) of the GNU compiler. Costs you
: | : all of $0.
: |
: | Please tell me where I can get a CDROM drive for $0 so I can put my
: | Solaris sampler CD for $0 in it. Thanks.
:
: I will, just as soon as you can tell me where to get a free computer
: to put the free drive into.
:
: And what about a free house with free electricity to put the free computer
: into?
:
: Cheesh.
You missed the point by several parsecs. One topic in this thread is
trying to compare UNIX systems by listing features and cost for the
platforms of interest. I felt the post I replied to made a big assumption
that all sites have a CDROM drive. There was an implied smilie in my
reply, I guess you missed it.
: There are enough free or nearly free CDROMS with stuff on them that
: the cost of a drive pays for itself in reduced network costs in a
: very short time.
Totally subjective. If this is so clear, why are so many people raising
the issue of development package costs? Are they just dumb? *YOU* try
and tell my site admin that CDROM drives are "free".
: (Of course, I know that those of you in the .edu domain believe or
: are taught to believe that everything is either free or should be free.
: So I suggest you ask your professor for the drive.)
Well, I can tell the previous paragraph is content free. BTW, I was in
the .com domain for a decade before I went back to school, just for the
record. Anyway, I think you will find that of the systems discussed so
far, a CDROM drive was not an assummed component and I still believe
development package costs are a valid concern when comparing systems.
: Robert Withrow, Tel: +1 617 598 4480, Fax: +1 617 598 4430, Net: wi...@rwwa.COM
: R.W. Withrow Associates, 21 Railroad Ave, Swampscott MA 01907-1821 USA
-Shane
Larry Snyder
>>at 1/3 the cost of SCO. These SVR4 products are just as reliable
>I hate to keep bringing up reality here. Where do these numbers come from all
>the time? Who makes them up?
Buy SCO Unix unlimited users, add TCP/IP, NFS, Dos Merge, complete development,
X11R5, text processing, network support tools. Now what is the price? Take
that total and compare it against Dell's complete price which includes all of
the above for $1295. So you can spend 1295 and get an OS run time (SCO), or
get everything mentioned above.
Chip Salzenberg
>I suspect that the Intel Unix systems may be doomed.
Not a chance. The NT, OS/2, and MS-DOS markets will continue to push
the PC clone architecture QUICKLY toward better bang/buck ratios. Sun
can't amortize SPARC development costs like the PC clone makers can.
PC UNIX rides that wave of good and cheap hardware, and there's no
reason to think it won't continue to do so.
--
Chip Salzenberg at Teltronics/TCT <ch...@tct.com>, <7371...@compuserve.com>
"you make me want to break the laws of time and space / you make me
want to eat pork / you make me want to staple bagels to my face /
and remove them with a pitchfork" -- Weird Al Yankovic, "You Make Me"
Bruno Hall
[ bunch of stuff deleted -BH ]
>(Of course, I know that those of you in the .edu domain believe or
>are taught to believe that everything is either free or should be free.
>So I suggest you ask your professor for the drive.)
Do tell us, Robert, how much did *you* (as a non-.edu domain person)
contribute to MIT since you're using X? How much money did you send
the FSF when you installed gcc? How about when you installed emacs?
Foo.
Bruno
--
/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/
Bruno Hall | VE2HUM | br...@mcrcim.mcgill.edu
McGill Research Centre for Intelligent Machines - Controls Group
New systems generate new problems -- Join the Flat Earth Society.
william E Davidsen
| I can't help but think that when a hacker is reduced to pushing a
| commercial UNIX -- ANY commercial UNIX -- instead of Linux or 386BSD,
| she may win the battle, but she will have lost the war.
Time to realize that there is room for both commercial and free
software, and not try to use one solution to all problems. And I think
you know that and are just agitating...
Some people like to hack the o/s and some like to do applications.
After a while you want a change, and want to DO something with your
system (I know people who's favorite app is kdb, they're sick).
Personally I've been on a team which wrote an o/s, designed and written
a multitask multiproc o/s by myself, and while I run Linux for fun, I
keep it as a porting target, not a way of life. I develop apps on a
commercial o/s, both for fun and professionally.
william E Davidsen
| > And you wind up with a legal copy of Xenix which has some resale
| >value.
|
| But today, not much.
Actually, Xenix is much better to run than UNIX in a laptop, unless you
really NEED the features. You can run Xenix in about 2MB, although 4 is
nicer. Can't do that with UNIX. (To slightly change the topic) I like
Dell for that, too, because I can put runtime, networking, C, and swap
in 70MB and have room for a few apps.
| I suspect that the Intel Unix systems may be doomed. The introduction of
| low-cost Sparc-based systems could eat Intel's lunch. The current offerings
| (and not just from Sun) are nice, fast, <cheap> and expandable through the
| network.
I don't argue the point, but while the 8086 and 286 systems are gone,
Intel continues. And the 386 is vanishing as the 486 gets down to $300
or some more. If Intel ever chooses to ship the Pentium it should be
faster than the bottom of the line SPARC, etc, and the ability to run
DOS software is still important to some people. If someone were to ship
a RISC platform with an intact V.4 (instead of BSD) I'd be more
inclined. I like V.4 sysadmin and layout better than BSD, a rarity (and
I admin Suns and Ultrix at work).
|
| The other interesting thing is that "board level" products are now available
| for the "build your own clone" group. The prices on these are coming down
| <fast>. Again, they're much faster than the 486 systems of today.
Another good point. The system board and frame buffer are the only
cost differentials these days, since memory, disks, and monitors are
essentially the same price for all platforms. We're using Sun 1GB SCSI
drives in PCs these days, because we got a good price on quantity.
|
| Then there is the application base. Sparc systems have a reasonably good
| one. PC Unix systems have a horrible one.
True, but only for shrink wrapped apps, and lots of PC stuff is now
available for SCO (and runs on V.4, in my experience). While compiling
net stuff was really bad for Xenix, and is still a hassle for SCO (due
to header files), most V.4 will handle net source with only minor
changes at the makefile level. Not all, I agree, but when I got my
first V.4 I ported about 60 apps in a weekend, and most of the fixes
were genuinely non-portable code which was found with the compiler, not
the debugger.
There's more software for Xenix/SCO UNIX than SPARC, and priced
better. As the vendors see a small system market, I believe some will
drop prices to gain market share.
william E Davidsen
| >>And how does this Equinox cluster controller look? How did they hook 200
| >>users onto their system. I mean physically.
|
| >boxes bolted on large pieces of plywook with a nest of wires (tie wrapped
| >of course)..
|
| Sounds like a real rube-goldberg setup and not very professional.
Sounds state of the art to me. Look in a wire cabinet of any telephone
system, and you'll see that they have a maze of wires tie wrapped. If
you use a corolary or something the wire come into a box on J45 (or some
such) and out on a cable, but inside the black box is still a maze of
wires, probably with wire ties.
With modem control you have eight wires per user, so there are just a
lot of wires to be found. You can go into terminal servers and through
a net, but trying to do modem control signals over a net is
challenging.
The objective is to be mechanicaly stable for reliability, and
accessable for maintenence. What larry describes is not "esthetically
pleasing" (to use a State of NY contract term), but it sounds as though
it could be solid and reliable. I'll settle for that.
william E Davidsen
| Don't be too quick to let GCC's reputation cloud your judgement.
| Where I work, we find Sun's new compiler to be about the same as GCC
| a lot of the time. And on my Dell SVR4 system, the stock compiler
| is as good if not better than GCC most of the time, esp. in speed of
| compile.
Agreed, Dell rel 2.2 has gcc 1.4, gcc 2.2 and cc, and there is no
clear "best" in terms of performance of the generated code. Moreover,
the versions of gcc don't require the same options to generate optimal
code, so it's an adventure if you need best code. Most of my stuff won't
justify an hour of testing to save 5% or so, and for almost all programs
that's the magnitude of the advantage.
william E Davidsen
| Again, you can install with the "traditional" security defaults, and
| then lower the minimum login-id length and the password-length
| restrictions. You can even remove the password-checking stuff
| entirely. Try poking around under Accounts->User->Default in
| sysadmsh, as well as /etc/default/authsh and /etc/default/passwd.
Several helpful people from SCO contacted me, and they suggested that
I might want to install with lower security and then convert the
password file to shadow with pwcnv. Since I only have to create multiple
users with the same UID once, I can hack the passwd file before
converting, then set the authsh and passwd defaults low. It's not
exactly convenient, but it's more supported than what I'm doing now when
I must, which is a script which rebuilds the protection stuff after I
have passwd and shadow.
I believe that with higher security you have a minimum password length
forced somewhere else, but until I do another install I won't really
know.
william E Davidsen
| For once we agree. The problem with the SCO (read MS) compiler is that SCO has
| little control over it. It least that was the story from the folks who worked
| in that area. The long lists of 'things' just kept getting longer. GNU was the
| answer to the problem. I'm not sure where the real fault lies here or why MS
| can't do a better compiler. I mean the have the horses and resources.
The 3.2.4 compiler (ODT 2.0) seems to produce pretty good code
compared to gcc. I run a Dell box at home for pleasure, and an SCO box
to do cross compiles. I have occasionally compared the five available
flavors of C (SCO cc and rcc, Dell cc, gcc-1 and gcc-2) on compute
intensive things, and find that the SCO and V.4 cc do better than gcc
(1.4 or 2.2) in many cases. Not a lot better, but some.
I can compile with SCO, shared libraries, and execute on the Dell box,
so all executions are measured on the same system, with the same shared
libs. This is as close as I can come to equalizing conditions to look
only at the result of the compiler.
My compiler of choice on Dell is cc, oddly.
Adams
> According to ka...@ddsw1.mcs.com (Karl Denninger):
> >I suspect that the Intel Unix systems may be doomed.
> Not a chance. The NT, OS/2, and MS-DOS markets will continue to push
> the PC clone architecture QUICKLY toward better bang/buck ratios. Sun
> can't amortize SPARC development costs like the PC clone makers can.
Why? Chip development was done by Cypress, Logitec, TI, NEC or Fujitsu
[unsure] etc., but outside of SUN. SUN told them just secifications.
Period.
Indeed SPARC processors do not tend to be more expensive than 486DXs,
price driven by monopoly of Intel. Have a look at the
microSPARC by TI, which has a CPU, FPU and all glue logic,
including DRAM controller and SBUS interface, put on _ONE_ chip.
> PC UNIX rides that wave of good and cheap hardware,
Current cheap PC hardware does not seem to match current challenge
very well. Why do you think do Local Bus etc occur concurrently
to GUIs?
> and there's no
> reason to think it won't continue to do so.
At least here in Europe, one can prove, that workstations based
on PC will be tremedously more expensive than one based on SPARC.
Please glance through ct12/92, ct1/93, ct2/93 and iX1/93
all published by Heisse Verlag, Hannover, FRG.
Do not forget, that clone makers start to focus on SPARC meanwhile.
Window NT is targeted to processors like Mips R3000/R4000 and
DEC/alpha. Next versions of OS/2 will be based on a mciro kernel,
and IBM sees OS/2 just as one step into distributed computing (DCE)
and propose their PowerRisc therefor.
All arguments above need not be taken as arguments, though
they are too many and too strong hints to ignore.
best, adams
Larry Snyder
> Agreed, Dell rel 2.2 has gcc 1.4, gcc 2.2 and cc, and there is no
>clear "best" in terms of performance of the generated code. Moreover,
The bottom line is with Dell it truely is plug and play. You install
the OS, and the option packages. You can start compiling using gcc
and configuring elm, cnews, nntp and X11R5 while the installation
tape continues to install the man pages and assorted other optional
packages. With the majority of other products, one has to find the
software, then compile and install it. Sure, someone with the right
files and experience can do this in a couple of hours -- but others
might not have everything they need, nor access to the internet to get
it. With Dell that isn't needed since it truely is plug and play.
How many folks have we seen complained about having SCO and looking
for gcc or x11r5 since they don't want to be raked over the coals for
more products at SCO prices? How many folks on the net have placed the same ads
who run dell?
Packaged pricing is the key. Just because you buy a package which
includes everything doesn't mean you need to install or use everything.
But just in case, it's there for you -- when you add that extra hard
disk, extra memory, large video monitor or high res graphics card.
Shane Bouslough
: >>And how does this Equinox cluster controller look? How did they hook 200
: >>users onto their system. I mean physically.
:
: >boxes bolted on large pieces of plywook with a nest of wires (tie wrapped
: >of course)..
:
: Sounds like a real rube-goldberg setup and not very professional.
"Oh yeah? Well, your mother uses goto's!" Sheesh I wondered when this
was going to start sounding like a cat fight. Here it comes.
: W. Fred Rump office: fr...@COMPU.COM "A man's library is a sort of
: 26 Warren St. home: f...@icdi10.compu.com harem" - Emerson (1860)
: Beverly, NJ. 08010
: 609-386-6846 bang:uunet!cdin-1!icdi10!fr
-Shane
Chip Salzenberg
> Time to realize that there is room for both commercial and free
>software, and not try to use one solution to all problems.
But a source-available kernel isn't one solution; it's all the
solutions you and thousands of others can imagine and code.
My motto: If you don't have source code, it's not software.
>And I think you know that and are just agitating...
Reading my mind now?
James Deibele
>SCO has a serious value problem.
They're not the only ones. NextStep 3.0 looks like a wonderful
development environment. Price announced for the '486: $995 runtime,
$2495 for the developers version.
I could understand the $2495 for the developers if Next were going to
ship the runtime for $99.95 --- the programmers are going to need a lot
of support and you can use a price as a way of sorting out the serious
programmers. (Somewhere in a filing cabinet I have a write-up on the
experiences of a company selling a programmer's tool: their $100 version
generated 80% of their support calls. So they dropped it and only sold
their $500 deluxe package.) If you can't pay $2495, you might not have
the money to finish the package.
But at $995 for the OS, the product has to be of mind-boggling value and
a very high price. "Here, our WhizBangWriter is only $195 --- but you
have to pay $995 for the operating system to run it." That'd go over
real big against Word for Windows. The Next app would have to write the
letter for you ... of course, you could grab some Shakespeare off the
disk for a touch of class.
Of course Steve Jobs used to work with Guy Kawasaki, who wrote in his
book about _The Macintosh Way_ about how he would show the Mac to
developers, then tell them how Apple was going to charge them for
documentation, and support, and they'd have to develop on the Lisa since
the Mac itself didn't have development tools yet. And some people, the
true believers, would still go for it. But it's 1993, not 1983. It's
not CP/M or MS-DOS that's the enemy, it's X/Windows, MS Windows, OS/2, and
the Mac itself.
To come back to SCO: I would bet a large sum of money that most of the
people running SCO have no idea what operating system they're using.
They just do their accounting or whatever other vertical market software
they need. End-users don't buy operating systems, they buy solutions.
And for that market, the cost of the operating system is usually trivial
compared to support and custom programming.
(So why won't NextStep do well? Because they didn't get big enough fast
enough. If Next gets past all the PC competitors, there's SUN with
Solaris. And SUN's has just a little more muscle than Next ... )
--
jam...@techbook.COM "2091 newsgroups & nothing on ..."
PDaXs gives free access to news & mail. (503) 220-0636 - 1200/2400, N81
Full internet (ftp, telnet, irc) access available. Voice: (503) 223-4245
PC GrabDisc is a monthly CD-ROM of interesting PC files. Email for info.