The Rise And Fall Of Sun Microsystems
Baby Peanut
about closed systems and Sun was about open systems. Sun prospered
and DEC withered.
Now it seems that Sun is past its peak, lost its vision of openness
and has turned into just another closed systems shop:
http://makeashorterlink.com/?D1F324043
How long before HP purchases what's left of Sun?
David Magda
[...]
> Now it seems that Sun is past its peak, lost its vision of openness
> and has turned into just another closed systems shop:
While I think it's unfortunate that Sun doesn't give US-III, you do
have to remember that they gave it to Linux for the UltraLinux
port. The Linux porter (David Miller?) had to sign an NDA. Theo de
Raadt is not agreeing to sign an NDA because of philosophical
reasons. Something I don't see anything wrong with.
The question you have to also ask is: how is Sun becoming more
"closed"? By:
* giving away the source to StarOffice
* opening up GridEngine (now on SourceForge?)
* helping to develop GNOME
* helping U. of Michigan with NFSv4 (to Solaris, Linux, OpenBSD)
Anything else?
> http://makeashorterlink.com/?D1F324043
The link is to a Slashdot story regarding OpenBSD trying to
get documentation for the US-III.
I have no idea what Sun is thinking on this one. They're a hardware
company: you want more systems running on that hardware.
> How long before HP purchases what's left of Sun?
Personally I think HP will sink before Sun, just because of
stupidity. There's an interesting thread in comp.arch on HP's
treatment of the Alpha processor.
Time will tell, he always does.
--
David Magda <dmagda at ee.ryerson.ca>
Because the innovator has for enemies all those who have done well under
the old conditions, and lukewarm defenders in those who may do well
under the new. -- Niccolo Machiavelli, _The Prince_, Chapter VI
Baby Peanut
> baby_...@yahoo.com (Baby Peanut) writes:
> [...]
> > Now it seems that Sun is past its peak, lost its vision of openness
> > and has turned into just another closed systems shop:
>
> While I think it's unfortunate that Sun doesn't give US-III, you do
> have to remember that they gave it to Linux for the UltraLinux
> port. The Linux porter (David Miller?) had to sign an NDA. Theo de
> Raadt is not agreeing to sign an NDA because of philosophical
> reasons. Something I don't see anything wrong with.
>
> The question you have to also ask is: how is Sun becoming more
> "closed"? By:
> * giving away the source to StarOffice
> * opening up GridEngine (now on SourceForge?)
> * helping to develop GNOME
> * helping U. of Michigan with NFSv4 (to Solaris, Linux, OpenBSD)
>
> Anything else?
Sun is only open when it's in Sun's interest to be open.
-----
Sun And Documentation... (Score:5, Funny)
by Bowie J. Poag (16898) Alter Relationship on Tuesday January 28,
@07:36AM (#5173447)
(http://www.ibiblio.org/propaganda)
Sun, with ANY kind of documentation, is going to be a royal pain in
the ass. Here, i'll give you a personal example.
One day, I picked up a SparcStation 1 at a surplus auction. Cool, I
thought, I'll learn SPARC architecture, a bit about disaster recovery
with Sun hardware, Solaris, you name it. So, I hacked the hell out of
it, and learned everything I could without documentation. When it came
time to look at a manual. I called Sun.
"Hi... I was wondering if you could send me the owners manual for a
SparcStation 1."
"Sorry. Thats handled by SunStore."
"Whats SunStore?"
"They handle all our documentation."
So, I call SunStore, and ask the same question.
"Hi.. I was wondering if I could order a user's manual for a Sun
SparcStation 1. I know the machine is like 10 years old, but do you
still have the manuals?"
"Yes, we do."
"Great, i'd like to order one, then. Is Visa ok?"
"Uhh.. Well, we can't sell it to you."
"What do you mean?"
"Well, we cant sell you just one."
"Huh?"
"You need to order in lots of 500."
"You mean in order to buy a SparcStation 1 manual, I need to buy
500?!"
"Yes."
"Uhhh.... Ooooh-kaaay.. How much is a lot of 500?"
"$39.95"
"Oh, okay..I guess thats fine.. I dunno what i'm gonna do with 499
Sparc manuals tho. I guess you can keep them, and just send me one.
Thats all I need."
"39.95 is the unit price, sir. You're looking at a total of....
$19,975."
"No way!"
"Yes sir. Will this be on a Visa or Mastercard?"
*click*
Bowie J. Poag
Project Founder, PROPAGANDA Desktop Enhancement Graphics [ibiblio.org]
> > http://makeashorterlink.com/?D1F324043
>
> The link is to a Slashdot story regarding OpenBSD trying to
> get documentation for the US-III.
>
> I have no idea what Sun is thinking on this one. They're a hardware
> company: you want more systems running on that hardware.
>
> > How long before HP purchases what's left of Sun?
>
> Personally I think HP will sink before Sun, just because of
> stupidity. There's an interesting thread in comp.arch on HP's
> treatment of the Alpha processor.
>
> Time will tell, he always does.
HP is in bed with Intel. They both want the Alpha sunk and the Itanic
to be marketed instead. If lousey CPUs sunk companies Intel would be
dead and buried. I don't see HP's aliance with Intel as sinking
either of them.
Tim Bradshaw
> The link is to a Slashdot story regarding OpenBSD trying to
> get documentation for the US-III.
Isn't OpenBSD done by the people who were *so* fractious and difficult
that not only could they not get along with the people who (now) do
FreeBSD, but they then fell out yet further with the people who (now)
do NetBSD and formed their own tiny cult OS. Perhaps Sun are finding
it hard to deal with them because they are, in fact, loonies[1]
--tim (hmm, two articles defending Sun in one day, I'll have to think
up something rude to say about them.)
Footnotes:
[1] Not that I think any such thing, of course.
Tim Bradshaw
> "Yes, we do."
> "Great, i'd like to order one, then. Is Visa ok?"
> "Uhh.. Well, we can't sell it to you."
> "What do you mean?"
> "Well, we cant sell you just one."
> "Huh?"
> "You need to order in lots of 500."
I don't suppose it occurred to you that the way they `still have' the
documentation for 10 year old machines is that they are willing to
print it on demand, and their costs for doing so make creating single
copies hopelessly uneconomic?
--tim
David Magda
> * David Magda wrote:
> > The link is to a Slashdot story regarding OpenBSD trying to
> > get documentation for the US-III.
>
> Isn't OpenBSD done by the people who were *so* fractious and
> difficult that not only could they not get along with the people
[...]
It depends on who you ask. :>
Tim Bradshaw
> It depends on who you ask. :>
I tend to go by the amount of foaming at the mouth...
--tim
Baby Peanut
I don't suppose you've heard of PDF or PostScript. Seems I can get
docs for old DEC boxes as PDF files online. Seems pretty economical
to make one machine copy and just share it electronically.
Meanwhile more about the setting Sun:
How Sun can pull out of its slump
By Paul Murphy Originally published Jan 29, 2003
Printed from LinuxWorld.com
http://www.linuxworld.com/site-stories/2003/0129.sun.html
Tim Bradshaw
> I don't suppose you've heard of PDF or PostScript. Seems I can get
> docs for old DEC boxes as PDF files online. Seems pretty economical
> to make one machine copy and just share it electronically.
And DEC are doing *so* much better than Sun, aren't they?
--tim
Baby Peanut
http://www.internalmemos.com/memos/memodetails.php?memo_id=1321
W C
> "Hi... I was wondering if you could send me the owners manual for a
> SparcStation 1."
>
> "Sorry. Thats handled by SunStore."
He could go to sunhelp.org and subscribe to rescue list,
those are bigger loonies [0] than Theo: never ending therapy session.
There is also sunhelp list which is mostly silent [1], and the infamous
geek list where complex compensation agenda is the agenda.
Yet eventually he'd get some pointers [2].
[0] collectively
[1] understandably and understandingly
[2] after a year of bonding by flame wars
Tim Bradshaw
> http://www.internalmemos.com/memos/memodetails.php?memo_id=1321
Um, yes? So Java has some problems on Solaris. Well, I knew that,
actually, because I've run SMC. Now, if you'd like to explain just
what this has to do with the availability of documentation for old
machines, or how well DEC are doing these days?
--tim
Baby Peanut
This thread is about "The Rise And Fall Of Sun Microsystems" just like
the title suggests.
Tim Bradshaw
> This thread is about "The Rise And Fall Of Sun Microsystems" just like
> the title suggests.
And a version of Sun's Java being slow on a version of Solaris will
kill Sun. The same way the terrible performance of any early Solaris
2 did, the same way all the hideous bugs in NFS did before that.
Right.
--tim
John Miller
"Baby Peanut" <baby_...@yahoo.com> wrote in message
news:c5cf6e8.03012...@posting.google.com...
Linux is the last great hope for SUN, the big question is Does SUN have the
reserves to last out until Linux envelops the entire computer industry ?
YTC#1
When linux has a stable stack, doesn't allow memory violations, has a
clean upgrade path that does not require re-compiles. Can utilse
multiple processors efficently and scalable. Then maybe it can be taken
more seriously on larger systems. Until then it is a disaster waiting to
happen.
--
Bruce Porter
XJR1300SP, XJ900F, GSX750W, GS550, GSX250, CB175
POTM#1(KoTL), WUSS#1 , YTC#1(bar), OSOS#2(KoTL) , DS#3 , IbW#18 ,Apostle#8
"The internet is a huge and diverse community
and not every one is friendly"
http://www.ytc1.co.uk
There *is* an alternative! http://www.openoffice.org/
John Miller
"YTC#1" <y...@ytc1.co.uk> wrote in message
news:3E5006E1...@ytc1.co.uk...
> John Miller wrote:
> > "Baby Peanut" <baby_...@yahoo.com> wrote in message
> > news:c5cf6e8.03012...@posting.google.com...
> >
> >>I think that the number one thing that killed DEC was Sun. DEC was
> >>about closed systems and Sun was about open systems. Sun prospered
> >>and DEC withered.
> >>
> >>Now it seems that Sun is past its peak, lost its vision of openness
> >>and has turned into just another closed systems shop:
> >>
> >>http://makeashorterlink.com/?D1F324043
> >>
> >>How long before HP purchases what's left of Sun?
> >
> >
> > Linux is the last great hope for SUN, the big question is Does SUN have
the
> > reserves to last out until Linux envelops the entire computer industry ?
> >
>
> When linux has a stable stack, doesn't allow memory violations, has a
> clean upgrade path that does not require re-compiles. Can utilse
> multiple processors efficently and scalable. Then maybe it can be taken
> more seriously on larger systems. Until then it is a disaster waiting to
> happen.
...and yet 7 out the 10 Most Powerful computers in the world run Linux.
Have you ever considered offering your services in a consultancy role to IBM
and NEC etc, just think how much better the Earth Simulator could have been
with your brains behind it !
Mike Jones
starl...@yahoo.co.uk says...
>
> "YTC#1" <y...@ytc1.co.uk> wrote in message
> news:3E5006E1...@ytc1.co.uk...
> > John Miller wrote:
> > > "Baby Peanut" <baby_...@yahoo.com> wrote in message
> > > news:c5cf6e8.03012...@posting.google.com...
> > >
> > >>I think that the number one thing that killed DEC was Sun. DEC was
> > >>about closed systems and Sun was about open systems. Sun prospered
> > >>and DEC withered.
> > >>
> > >>Now it seems that Sun is past its peak, lost its vision of openness
> > >>and has turned into just another closed systems shop:
> > >>
> > >>http://makeashorterlink.com/?D1F324043
> > >>
> > >>How long before HP purchases what's left of Sun?
> > >
> > >
> > > Linux is the last great hope for SUN, the big question is Does SUN have
> the
> > > reserves to last out until Linux envelops the entire computer industry ?
> > >
> >
> > When linux has a stable stack, doesn't allow memory violations, has a
> > clean upgrade path that does not require re-compiles. Can utilse
> > multiple processors efficently and scalable. Then maybe it can be taken
> > more seriously on larger systems. Until then it is a disaster waiting to
> > happen.
>
> ...and yet 7 out the 10 Most Powerful computers in the world run Linux.
> Have you ever considered offering your services in a consultancy role to IBM
> and NEC etc, just think how much better the Earth Simulator could have been
> with your brains behind it !
The problem is that those computers bear about as much resemblance to
commonly available machines as your average NASCAR ride does to the
Monte Carlo you'll find at the Chevy dealer. They're custom from the
ground up, hardware and software.
--
Mike Jones
Government is not establish'd merely by Power; there must be maintain'd
a general Opinion of its Wisdom and Justice, to make it firm and
durable.
-- Benjamin Franklin
YTC#1
> "YTC#1" <y...@ytc1.co.uk> wrote in message
> news:3E5006E1...@ytc1.co.uk...
>
>>John Miller wrote:
>>
>>>"Baby Peanut" <baby_...@yahoo.com> wrote in message
>>>news:c5cf6e8.03012...@posting.google.com...
>>>
>>>
>>>>I think that the number one thing that killed DEC was Sun. DEC was
>>>>about closed systems and Sun was about open systems. Sun prospered
>>>>and DEC withered.
>>>>
>>>>Now it seems that Sun is past its peak, lost its vision of openness
>>>>and has turned into just another closed systems shop:
>>>>
>>>>http://makeashorterlink.com/?D1F324043
>>>>
>>>>How long before HP purchases what's left of Sun?
>>>
>>>
>>>Linux is the last great hope for SUN, the big question is Does SUN have
>>
> the
>
>>>reserves to last out until Linux envelops the entire computer industry ?
>>>
>>
>>When linux has a stable stack, doesn't allow memory violations, has a
>>clean upgrade path that does not require re-compiles. Can utilse
>>multiple processors efficently and scalable. Then maybe it can be taken
>>more seriously on larger systems. Until then it is a disaster waiting to
>>happen.
>
>
> ...and yet 7 out the 10 Most Powerful computers in the world run Linux.
Based on what benchmark ? The one that makes them look best.
> Have you ever considered offering your services in a consultancy role to IBM
> and NEC etc, just think how much better the Earth Simulator could have been
> with your brains behind it !
The others are right, you can be tedious.
--
Bruce Porter
Anthony Mandic
>
> John Miller wrote:
......
> The others are right, you can be tedious.
Just ignore both peanut brains. They don't have a clue between
them.
-am © 2003
Stefaan A Eeckels
"John Miller" <starl...@yahoo.co.uk> wrote:
>
> Linux is the last great hope for SUN, the big question is Does SUN have
> the reserves to last out until Linux envelops the entire computer
> industry ?
--
Stefaan
--
"One man alone can be pretty dumb sometimes, but for real bona fide
stupidity there ain't nothing can beat teamwork." -- Mark Twain
Dave Uhring
> Well put, anyone knows that for secure scalable systems, people use
> FreeBSd..
SMP capability, and limited capability at that, of 2 processors is
"scalable"???
Chuck Swiger
> Rev. Don Kool, where are you when we need you most?
If I threw a Bible at the Rev, would he go up in a cloud of black smoke?
Yeah, yeah-- we should be so lucky...still, it'd be neat to see. :-)
-Chuck
Chuck Swiger | ch...@codefab.com | All your packets are belong to us.
-------------+-------------------+-----------------------------------
"The human race's favorite method for being in control of the facts
is to ignore them." -Celia Green
Rich Teer
> SMP capability, and limited capability at that, of 2 processors is
> "scalable"???
In PeeCee land, yes!
--
Rich Teer
President,
Rite Online Inc.
Voice: +1 (250) 979-1638
URL: http://www.rite-online.net
Dave Uhring
> On Mon, 17 Feb 2003, Dave Uhring wrote:
>
>> SMP capability, and limited capability at that, of 2 processors is
>> "scalable"???
>
> In PeeCee land, yes!
Casper pointed out quite some time ago that Solaris 8 x86 is capable of
scaling to 21 processors. And that OS is a part of PeeCee land :-)
Hell, even Linux can scale to 8 processors in a reasonably linear manner!
scz
>
> On Mon, 17 Feb 2003 18:10:43 +0000, Rich Teer wrote:
>
> > On Mon, 17 Feb 2003, Dave Uhring wrote:
> >
> >> SMP capability, and limited capability at that, of 2 processors is
> >> "scalable"???
> >
> > In PeeCee land, yes!
>
> Casper pointed out quite some time ago that Solaris 8 x86 is capable of
> scaling to 21 processors. And that OS is a part of PeeCee land :-)
The OS is willing but the hardware is weak, or something. :-)
scz
Baby Peanut
You can only make just so many mistakes before the end.
Meanwhile, here's Robert X. Cringely on the setting sun:
Baby Peanut
Sunset
How to Avoid the Almost Certain End of Sun Microsystems
By Robert X. Cringely
Every five to 10 years, Silicon Valley goes broke. This began in the
1950s and maybe long before, but the 1950s is as early as I care to
write about. The Valley then was filled with apricot and cherry
orchards only to see agriculture driven out first by the military and
aerospace, and then by semiconductor companies. It is fitting that
Shockley Semiconductor -- the first of many transistor companies --
was started in a shed previously used for drying apricots.
Transistors begat Integrated Circuits, which begat memory chips, which
begat microprocessors, which begat personal computers, which begat
consumer software, which begat networks, which begat the Internet,
which begat the day before yesterday and the day after tomorrow. And
each of those transitions was accompanied by a seismic shudder going
through the Valley as companies went under and home prices slowed, for
just a moment, their inexorable rise before continuing to climb again.
A few familiar names survived from each era, but most of the
companies went out of business because that's the way it is. We burn
our fields in Silicon Valley, then plow the ashes under and start
anew. It is perfectly natural, then, for companies to die here, but
that doesn't mean there is no room for regret and nostalgia. So today
I look with nostalgia on Sun Microsystems and hope -- probably in vain
-- that the company doesn't die.
Sun did not invent the engineering workstation, but they certainly
perfected it. But where are workstations today? Gone, for the most
part. Sun's workstation business is about the same size as SGI's,
which is to say small. Sun is now a server company, but that won't
last long either under the onslaught of Linux. Cheap Intel and AMD
hardware running Linux is going to kill Sun unless the company does
something so stop it, which they aren't.
Sun made a big show this week of rolling out its new product strategy,
called N1, which pits the company directly against both Microsoft and
IBM. Both Napoleon and Hitler learned the hard way that it is not a
good idea to fight a war on two fronts, and Sun, which can barely
afford to compete against one of those companies, much less both, is
about to get the same rueful lesson.
Sun's announcements were too little, too late, and they were made by
absolutely the wrong people -- a succession of marketing executives.
Sun is an engineering company, so where were the engineers? The
engineers were kept in the back rooms lest they reveal the despair
being felt right now in their company. The problem is that Sun has no
real technical leadership. CEO Scott McNealy doesn't know what to do
with the company. Ed Zander is gone, which is good, but that means it
has been years since the company had anything like charismatic or
visionary leadership. It doesn't look good.
Even Java is becoming superfluous. Java is the Dan Marino of
software. Just as the former Dolphins quarterback, Java affected the
world so much that history cannot be written without its mention. But
nonetheless, neither Java nor Dan ever won the big one.
So here is the prognosis. Sun lost $2 billion last year and will
probably lose another $2 billion this year. At that rate, the company
has at most five years to live. They have just renewed a commitment
to the Solaris operating system, which is no longer really viable from
an economic standpoint. I know, I know, Solaris users love Solaris,
but they don't love Solaris prices. And with a falling market share,
Sun can't afford to make Solaris any cheaper. Sun is having the same
problem in hardware where their SPARC architecture is falling behind,
and -- worse still -- has lost nearly all of its manufacturing support
in Japan. Both Solaris and SPARC will absorb vast sums in the coming
years and yield absolutely no increase in Sun's market share as a
result.
Here is something very important to understand: winning its current
anti-trust suit against Microsoft will not change the final outcome
for Sun. An award of $1 billion or even $3 billion (possible treble
damages) won't do anything except buy a little time.
It would be great if something happened to arrest Sun's fall. One
rumor going around is that Sun will merge with Apple, which is ironic
since Gil Amelio tried unsuccessfully to GIVE Apple to Sun back in
early 1997 before Gil was fired as Apple CEO. The logic behind this
rumor is that Apple is now effectively a Unix company, that Apple and
Sun could target the desktop and server markets, respectively, and
that Sun would drop SPARC in favor of PowerPC processors.
This is a nice rumor, but I don't believe it. Steve Jobs has done an
excellent job of turning Apple into a boutique computer company. He
can move Apple quickly to stay ahead of the market as he is doing
right now shifting the company more and more into notebooks, about the
only PC area that is still growing. But Jobs couldn't do the same
thing with a post-merger Apple/Sun. The company would be too big and
the cash reserves would be too low. The competition -- again
Microsoft and IBM -- would be too big and too rich. Steve is
ambitious, but he is not an idiot. There is nothing at Sun right now
that Apple needs.
So what is to be done? The answer is clearly two versions of the same
thing. Sun can either find a merger partner to take the company out
of its predicament or it can find its own strategy to achieve the same
result. Either way, this is a time for Scott McNealy to literally bet
the company.
To hear them talk, Sun's marketing folks think they are already
betting the company, but they aren't. They are throwing the company
away, which is very different. It is the difference between taking a
calculated risk that might turn the company around and the current
strategy of simply spending more money NOT trying to turn the company
around in hopes that some happy accident will take place before Sun is
completely broke.
I don't know exactly what Sun should do to save itself, but I know it
has to involve a bold and brash move that changes the entire company,
and with that, the entire game. Sun has to reinvent itself.
One way to do that is through a merger, but the logical merger partner
isn't Apple, it is Sony. The two companies have been talking about
some kind of strategic alliance. Maybe these are merger talks. Sony
is incredibly strong, having just posted its biggest-ever profit.
Sony leadership is changing, making possible a bold move as the new
management tries to put its own stamp on the company. Sony has both
the resources to support Sun and the need for technology Sun can
provide.
Sony is a leader in consumer electronics and home entertainment, but
not in computers. While the combined companies could field some very
good computer offerings, extending Sony's influence into the server
space, the real value of the combination lies in using Sun technology
and know-how to transform Sony's current bread-and-butter businesses,
which are TVs, video games, and movies.
With Sun's help, Sony could redefine television, bringing it into the
emerging broadband era. A Sony Internet TV could show Sony content
received over a Sony global network, all engineered by Sun. It is a
powerful attraction, and at around $3 per share, Sun is very
affordable for Sony.
I don't know if this will happen, but it might. If it doesn't happen,
then Sun will just have to go it alone, which means Scott McNealy will
have to stumble on a new business just as he stumbled on servers and
Java. This means getting new and energetic technical leadership for
the company, which desperately needs another Bill Joy. Then it means
finding a new product direction. And finally it involves betting the
whole darned company on that direction. Like Hernando Cortez
conquering Mexico, McNealy has to burn his ships to make retreat
impossible. While the risk in this strategy looks great, the
alternative is almost certain doom.
What can make the new strategy succeed is a particular CEO behavior:
McNealy has to not grow up. Be brash, be stupid even. Take enormous
risks and do it with élan. Only then will Sun return to greatness.
Rich Teer
> Sunset
> How to Avoid the Almost Certain End of Sun Microsystems
>
> By Robert X. Cringely
[...]
> which is to say small. Sun is now a server company, but that won't
> last long either under the onslaught of Linux. Cheap Intel and AMD
> hardware running Linux is going to kill Sun unless the company does
"Cheap Intel and AMD hardware running Linux" is not match for Sun's
bigger machines, although I agree that Sun needs to boost its low
end price/performance.
> Even Java is becoming superfluous. Java is the Dan Marino of
> software. Just as the former Dolphins quarterback, Java affected the
> world so much that history cannot be written without its mention. But
> nonetheless, neither Java nor Dan ever won the big one.
Based on what, exactly?
CJT
With all due respect, Cringely isn't much of an authority IMHO.
Emmanuel Florac
rich...@rite-group.com disait...
>
> "Cheap Intel and AMD hardware running Linux" is not match for Sun's
> bigger machines, although I agree that Sun needs to boost its low
> end price/performance.
>
No, but expensive IBM hardware do. Even expensive Intel-based hardware
(see SGI altix and more to come) do or will do within a few months. And
frankly, relying upon a market which is shrinking again and again...
> > [quoted text muted]
> > Even Java is becoming superfluous. Java is the Dan Marino of
> > software. Just as the former Dolphins quarterback, Java affected the
> > world so much that history cannot be written without its mention. But
> > nonetheless, neither Java nor Dan ever won the big one.
>
> Based on what, exactly?
>
Java sucks? Java is slow, a memory hog, and isn't half as well-designed
as C#, nor open-source as Python, nor as powerful as perl?
--
Quis, quid, ubi, quibus auxiliis, cur, quomodo, quando?
Phillip Fayers
>On 18 Feb 2003, Baby Peanut wrote:
>> Sunset
>> How to Avoid the Almost Certain End of Sun Microsystems
>> By Robert X. Cringely
>[...]
>> which is to say small. Sun is now a server company, but that won't
>> last long either under the onslaught of Linux. Cheap Intel and AMD
>> hardware running Linux is going to kill Sun unless the company does
>"Cheap Intel and AMD hardware running Linux" is not match for Sun's
>bigger machines, although I agree that Sun needs to boost its low
>end price/performance.
Sun needs either massively increase the performance of its CPUs or
dump UltraSPARC and use something else.
UltraSPARC simply isn't keeping up with the other chips out there.
It's totally outclassed at the low end and at the high end. A news
story today is reporting that IBM have a Power 5 system running in
their labs. Power 5 is due for release next year. Given that Power
4 already wipes the floor with Sun on performance they haven't got
a hope against the next generation. Heck, even Apple desktops
are faster than Sun desktops these days, actually even low end
portables are faster than low end Suns. You simply can't survive
in computing if you are that far behind the performance curve.
Sun is kidding itself if it thinks it can survive as a server company.
The reason Sun became successful was that they sold fast, comparitively
cheap systems in a market segment which was growing - UNIX workstations.
It was pretty easy to turn those workstations into cheap servers, then
make them bigger. If you read about the current SPARC chips you'll
find that the US-IIi was pretty much an accident, which shows you how
much Sun cares about the low end. Without cheap, competetive systems
to sell at the low end Sun will get squeezed into the job of selling
a few high end systems to a few huge companies.
We used to buy Suns. Not many I'l grant you but 4-5 years ago
there were close on 60 active Sun systems in this department, mostly
workstations. If we wanted a new UNIX box it was a Sun becuase it
was easy to integrate into the network and we had the skills base to
set it up and use it. In the last 2 years we've bought 1 Sun box.
We don't have anyone new learning how to use Suns, everyone wants
PCs running Linux becuase they are cheaper and faster. With Linux
skills growing it will mean that future servers run Linux because
they'll be easier to integrate and we'll have the skills. Within
a few years I'd expect most of the Suns to go. This is being repeated
in other departments in the university, and in other unis in the UK
and all over the world. Suns basic user base is moving away, without
it they'll die - I give them 3 years unless they make some radical
changes.
--
Phillip Fayers, SunAdmin/Support/Programming/Postmaster/Webmaster(TM)
Dept of Physics & Astronomy, University of Wales, College of Cardiff.
P.Fa...@astro.cf.ac.uk Attribute these comments to me, not UWCC.
Chuck Swiger
> Dans article <Pine.GSO.4.44.0302182102080.24976-100000@electron>,
> rich...@rite-group.com disait...
>> Based on what, exactly?
>
> Java sucks? Java is slow, a memory hog, and isn't half as well-designed
> as C#, nor open-source as Python, nor as powerful as perl?
Java doesn't suck. Java is a memory hog. Java is more open than GPL'ed
software, since I can use Java to write commercial software if I choose.
The expressive power of any Turing-complete language is equivalent; for sure,
I'd rather use Java than Perl for a program of any significant size.
All of the above being said, nobody answered by my question about the
dependencies Solaris has on older JVMs-- why java1.1, java1.2, and j2se all
being present under /usr under S8.
Emmanuel Florac
disait...
>
> Java doesn't suck.
We disagree on this part...
> Java is a memory hog. Java is more open than GPL'ed
> software, since I can use Java to write commercial software if I choose.
You can write commercial software with GPL software. Obviously you
misunderstood the GPL.
> The expressive power of any Turing-complete language is equivalent; for sure,
> I'd rather use Java than Perl for a program of any significant size.
>
Actually I'm getting sure that Perl and Python are probably better
languages overall, whatever size is the project...
Baby Peanut
Poisoning the well/damning the source is the poorest of rebuttals.
Emmanuel Florac
disait...
>
> What *does* Bill Gates' cock taste like?
>
Should I care? Do you think that Microsoft and its thousand engineers can
only do crap, just because it's Microsoft? Then surely SUN can only
produce marvels, like U5. Go buy a brain, be quick.
Emmanuel Florac
disait...
>
> Jeez, how did you get from C# to that? Oh, hang on, you're a moron.
>
Hm? write something meaningful, just prove you have a brain and you're
not some sort of ELIZA robot.
Steve Kappel
> In article <Pine.GSO.4.44.0302182102080.24976-100000@electron>, Rich Teer wrote:
>>On 18 Feb 2003, Baby Peanut wrote:
>
>>> Sunset
>>> How to Avoid the Almost Certain End of Sun Microsystems
>
>>> By Robert X. Cringely
>
>>[...]
>
>>> which is to say small. Sun is now a server company, but that won't
>>> last long either under the onslaught of Linux. Cheap Intel and AMD
>>> hardware running Linux is going to kill Sun unless the company does
>
>>"Cheap Intel and AMD hardware running Linux" is not match for Sun's
>>bigger machines, although I agree that Sun needs to boost its low
>>end price/performance.
>
> Sun needs either massively increase the performance of its CPUs or
> dump UltraSPARC and use something else.
>
> UltraSPARC simply isn't keeping up with the other chips out there.
> I give them 3 years unless they make some radical changes.
How many times have we heard a similar prediction? About every
5 years we go through this. Every time Sun has turned around
and responded favorably - continuing to maintain and grow its
market position.
Anybody who has watched this industry for more than a few years
knows full well that the fastest chip don't mean diddle. Plenty
of competitors have frequently had chips outperforming Sun -
some by a wide margin. Most of these are dead or dying.
The industry is about more than the fastest chip or the cheapest
O/S. Winners need integrated systems - the triad of scalable
hardware with common processor architecture, robust industrial-
strength operating system, and wealth of applications.
> We used to buy Suns. Not many I'l grant you but 4-5 years ago
> there were close on 60 active Sun systems in this department, mostly
> workstations. If we wanted a new UNIX box it was a Sun becuase it
> was easy to integrate into the network and we had the skills base to
> set it up and use it. In the last 2 years we've bought 1 Sun box.
> We don't have anyone new learning how to use Suns, everyone wants
> PCs running Linux becuase they are cheaper and faster. With Linux
Sun isn't out to make the cheapest and fastest desktop. Their
value comes in the triad above. Binary compatibility with
high-end servers, O/S, applications. Before choking on that
last one, it hasn't been that long since Sun was the reference
platform for most "open source" software. There is nothing
"reference" about Linux (like, which one??). Most desktop
apps on Linux are medicore at best.
It isn't a popularity contest. For the last decade we've
heard how Wintel was going to annihilate Unix and non-Intel
platforms. Didn't happen. Although Visual Basic was awful
popular! Linux isn't going to annihilate Unix either.
Desktops are becoming less and less important in a networked
world. That trend is going to hurt Microsoft a whole lot
more than Sun.
The Linux hype will level off once companies figure out
there is little to no money to be made.
Emmanuel Florac
disait...
>
> How many times have we heard a similar prediction? About every
> 5 years we go through this. Every time Sun has turned around
> and responded favorably - continuing to maintain and grow its
> market position.
That's a perfect evidence that this sort of FUD is useful, if not
necessary! :)
rachel polanskis
Rich Teer <rich...@rite-group.com> wrote:
> On 18 Feb 2003, Baby Peanut wrote:
>
> > Sunset
> > How to Avoid the Almost Certain End of Sun Microsystems
> >
> > By Robert X. Cringely
>
> [...]
>
> > which is to say small. Sun is now a server company, but that won't
> > last long either under the onslaught of Linux. Cheap Intel and AMD
> > hardware running Linux is going to kill Sun unless the company does
>
> "Cheap Intel and AMD hardware running Linux" is not match for Sun's
> bigger machines, although I agree that Sun needs to boost its low
> end price/performance.
If Sun goes under, the technologies that replace it will be
inferior for sure. The price point does need to come down though,
agreed. However I think Sun has been doing that to some degree
in recent times overall.
I shudder to think of what will happen if Sun disappears. I will
have to retreat to Apple/OSX I guess, since I am not happy with Linux
and certainly have a dislike for x86, for my own reasons! I haven't
used an Intel box for over 3 years now, except under duress.
It's a sad state of affairs that Linux, which has been deemed to
be a "Windoze killer" has instead completed the Balkanisation of UNIX
and because of Linux's tendency to be available on commodity hardware
has cheapened the whole game considerably, both in hardware, software
and skill levels. I feel that while Linux has a lot of merit,
there still needs to a space for a commercial, world class UNIX.
And I don't think HP-UX or True64 is it, from my own experience.
(please don't bother flaming me - I know what the usual arguments are)
I have watched Solaris grow from an awkward kludge into a really
slick and smooth running OS, that is a real joy to use. Likewise
Sun hardware seems to be improving all the time. These things do come
at a cost though, but while ever there's people prepared to cut corners
to weasel in the bottom line, the inferior solutions will always make
headway, sadly.
rachel
CJT
<snip>
>
>
> Should I care? Do you think that Microsoft and its thousand engineers can
> only do crap, just because it's Microsoft? Then surely SUN can only
> produce marvels, like U5. Go buy a brain, be quick.
>
Perhaps they could do better if it weren't for the need to maintain
compatibility with an intrinsically flawed design, compounded by a
corporate culture that values the length of a list of "features" over
basic reliability.
Stefaan A Eeckels
Emmanuel Florac <efl...@imaginet.fr> wrote:
> Actually I'm getting sure that Perl and Python are probably better
> languages overall, whatever size is the project...
Actually, that depends on whether your target audience
can install Perl or Python. The moment you have to
distribute Perl/Python to distribute your app, you're
up the proverbial creek without a paddle. Perl and Python
are fine programmers languages, but not good for stuff
that's distributed to clueless sites.
Phillip Fayers
>In article <slrnb56mg...@sadalsud.astro.cf.ac.uk>, Phillip Fayers wrote:
>> In article <Pine.GSO.4.44.0302182102080.24976-100000@electron>, Rich Teer wrote:
>>>On 18 Feb 2003, Baby Peanut wrote:
>>>> which is to say small. Sun is now a server company, but that won't
>>>> last long either under the onslaught of Linux. Cheap Intel and AMD
>>>> hardware running Linux is going to kill Sun unless the company does
>>>"Cheap Intel and AMD hardware running Linux" is not match for Sun's
>>>bigger machines, although I agree that Sun needs to boost its low
>>>end price/performance.
>> Sun needs either massively increase the performance of its CPUs or
>> dump UltraSPARC and use something else.
>> UltraSPARC simply isn't keeping up with the other chips out there.
>> I give them 3 years unless they make some radical changes.
>How many times have we heard a similar prediction? About every
>5 years we go through this. Every time Sun has turned around
>and responded favorably - continuing to maintain and grow its
>market position.
>Anybody who has watched this industry for more than a few years
>knows full well that the fastest chip don't mean diddle. Plenty
>of competitors have frequently had chips outperforming Sun -
>some by a wide margin. Most of these are dead or dying.
The problem now is that we have a few "living" architectures
(ie. discount Alpha as it's basically dead) and ALL of them
outperform Sun. They are also all committed to increasing
performance at a faster rate than Sun. I think Sun don't
have the right attitude to UltraSPARC (or many of their
products). There's a nice quote from Raju Vegesna in
an article (http://www.theworkcircuit.com/story/OEG20020731S0037)
I read recently. Raju was an architect on the Ross HyperSPARC
CPUs, he now run Serverworks. At the end of the article
he is quoted as saying:
"You have to win the business with every new silicon generation.
Execution is vital." And "if you blink you fail,"
Sun have been blinking for a while now.
>The industry is about more than the fastest chip or the cheapest
>O/S. Winners need integrated systems - the triad of scalable
>hardware with common processor architecture, robust industrial-
>strength operating system, and wealth of applications.
>> We used to buy Suns. Not many I'l grant you but 4-5 years ago
>> there were close on 60 active Sun systems in this department, mostly
>> workstations. If we wanted a new UNIX box it was a Sun becuase it
>> was easy to integrate into the network and we had the skills base to
>> set it up and use it. In the last 2 years we've bought 1 Sun box.
>> We don't have anyone new learning how to use Suns, everyone wants
>> PCs running Linux becuase they are cheaper and faster. With Linux
>Sun isn't out to make the cheapest and fastest desktop. Their
>value comes in the triad above. Binary compatibility with
>high-end servers, O/S, applications. Before choking on that
>last one, it hasn't been that long since Sun was the reference
>platform for most "open source" software.
"It hasn't been that long since" - exactly. Sun haven't produced
a competitive desktop (at a reasonable price) for a long time.
That hole in their product line is now manifesting itself - the
majority of developers have moved to using Linux boxes.
On the application front Sun have never really been serious about
that one either. StarOffice was far too little and far too late
for anyone who might want to keep a Sun on their desk for general
computing. Take a look at their new N1 initiative. One of the
main advantages of the new Sun File Blade servers is supposed to
be the N1 software - but look at the price! (The Sun Fire B1600
Intelligent Shelf lists at $4,795, a licence for N1 Provisioning
Server 3.0 Blades Edition is $3,290 for EACH shelf.)
>There is nothing
>"reference" about Linux (like, which one??). Most desktop
>apps on Linux are medicore at best.
But again they are getting better faster than the desktop
apps on Solaris.
>It isn't a popularity contest. For the last decade we've
>heard how Wintel was going to annihilate Unix and non-Intel
>platforms. Didn't happen. Although Visual Basic was awful
>popular! Linux isn't going to annihilate Unix either.
>Desktops are becoming less and less important in a networked
>world. That trend is going to hurt Microsoft a whole lot
>more than Sun.
>The Linux hype will level off once companies figure out
>there is little to no money to be made.
The Linux hype will continue as long as IBM/HP etc see that
it hurts Sun (which it does).
Emmanuel Florac
disait...
> Perl and Python
> are fine programmers languages, but not good for stuff
> that's distributed to clueless sites.
>
Though py2exe (for Python) is getting reaaly good.
Emmanuel Florac
disait...
>
> Perhaps they could do better if it weren't for the need to maintain
> compatibility with an intrinsically flawed design, compounded by a
> corporate culture that values the length of a list of "features" over
> basic reliability.
>
Microsoft seldom build good products; however the overall design of C# is
fine, and probably there'll be a good implementation (probably not from
Microsoft...)
Stefaan A Eeckels
Emmanuel Florac <efl...@imaginet.fr> wrote:
> Dans article <20030219225857.3...@ecc.lu>, hoen...@ecc.lu
> disait...
> > Perl and Python
> > are fine programmers languages, but not good for stuff
> > that's distributed to clueless sites.
> >
>
> Though py2exe (for Python) is getting reaaly good.
Last time I checked (several months ago) it had its problems.
I'll have to revisit it.
Thanks!
Steve Kappel
> In article <slrnb57tqs....@isis.visi.com>, Steve Kappel wrote:
>>In article <slrnb56mg...@sadalsud.astro.cf.ac.uk>, Phillip Fayers wrote:
>>> We used to buy Suns. Not many I'l grant you but 4-5 years ago
>>> there were close on 60 active Sun systems in this department, mostly
>>> workstations. If we wanted a new UNIX box it was a Sun becuase it
>>> was easy to integrate into the network and we had the skills base to
>>> set it up and use it. In the last 2 years we've bought 1 Sun box.
>>> We don't have anyone new learning how to use Suns, everyone wants
>>> PCs running Linux becuase they are cheaper and faster. With Linux
>
>>Sun isn't out to make the cheapest and fastest desktop. Their
>>value comes in the triad above. Binary compatibility with
>>high-end servers, O/S, applications. Before choking on that
>>last one, it hasn't been that long since Sun was the reference
>>platform for most "open source" software.
>
> "It hasn't been that long since" - exactly. Sun haven't produced
> a competitive desktop (at a reasonable price) for a long time.
> That hole in their product line is now manifesting itself - the
> majority of developers have moved to using Linux boxes.
And that is why "open source" software is sucking more and
more every day. For those of us on non-Linux Unix it is
becoming harder and harder to get said software to compile
and run. Back when Sun was the reference things were alot
better (for more than just Sun users). Linux bigots just
don't give a damn about anybody else. In that sense they
are == M$.
The Linux crowd has taken the phrases "open source" and
"open" and twisted the meaning. Sun used to be considered
the premier "open" platform. Partly due to the standards-
based approach. Note that the "used to be" isn't because
anything has changed with or on Sun - what has changed is
a bunch of newbies have redefined the term, for the worse.
The last time Sun had a "competitive hole" in their desktop
line they came out with the Ultra 10. They sold more of
those than any other desktop in their history. It will
happen again when the UIIIi/jalapeno comes out (soon).
> On the application front Sun have never really been serious about
> that one either. StarOffice was far too little and far too late
Wrong. You have a limited view of *free* software. In many
categories the premier applications are still on Sun. I
think Sun claims something like 4,500 applications -
far beyond ANY other Unix. And lots of people will pay
extra for value - even if it is a commercial product.
>>There is nothing
>>"reference" about Linux (like, which one??). Most desktop
>>apps on Linux are medicore at best.
>
> But again they are getting better faster than the desktop
> apps on Solaris.
Possibly. Now if those app developers cared about anybody
else they would make those apps portable to Solaris and
other Unixes. There is nothing about these desktop apps
that is specific to Linux - other than the authors
philosophical orientation.
As I've suggested in the past, Sun should either merge
with Apple or license the Apple desktop & apps. Wouldn't
that be interesting.
>>The Linux hype will level off once companies figure out
>>there is little to no money to be made.
>
> The Linux hype will continue as long as IBM/HP etc see that
> it hurts Sun (which it does).
Agreed. That is what failing companies do - grab whatever
they can to keep from sinking further.
Fortunately, buyers will eventually see the truth. Linux
is here to stay and it is good for some purposes. It
is not the grail of operating systems.
Raymond Toy
Steve> And that is why "open source" software is sucking more and
Steve> more every day. For those of us on non-Linux Unix it is
Steve> becoming harder and harder to get said software to compile
Steve> and run. Back when Sun was the reference things were alot
Steve> better (for more than just Sun users). Linux bigots just
Steve> don't give a damn about anybody else. In that sense they
Steve> are == M$.
I, unfortunately, agree with this. And it's only going to get worse.
Steve> The last time Sun had a "competitive hole" in their desktop
Steve> line they came out with the Ultra 10. They sold more of
Steve> those than any other desktop in their history. It will
Steve> happen again when the UIIIi/jalapeno comes out (soon).
I, unfortunately, disagree with this. That UIIIi/Jalapeno would have
to be 1+ GHZ and sell for about $1500 or so to even have a chance to
compete. The 650 MHz Blade goes for $2000. I don't need
fibre-channel disks, I don't need all the other fancy stuff. I want
MIPS and FLOPS and apps. And even apps are getting harder to get.
We're a key customer for one of our vendors. We pleaded and begged
for years to get a Solaris version of their tools. All we got was a
cold shoulder, saying they have no real interest, even when their
other key customers would also like a Solaris version. Finally, they
produce a version. It's functional, but lacks some of the key
features that the Windows version has. And then they wonder why no
wants it, after saying for years we'll never get it. Idiots.
Ray
Dragan Cvetkovic
> We're a key customer for one of our vendors. We pleaded and begged
> for years to get a Solaris version of their tools. All we got was a
> cold shoulder, saying they have no real interest, even when their
> other key customers would also like a Solaris version. Finally, they
> produce a version. It's functional, but lacks some of the key
> features that the Windows version has. And then they wonder why no
> wants it, after saying for years we'll never get it. Idiots.
If you are one of thier key customers, why didn't you say you will abandon
them unless they produce a Solaris version? Or are they also one of the key
vendors for you?
Bye, Dragan
--
Dragan Cvetkovic,
To be or not to be is true. G. Boole No it isn't. L. E. J. Brouwer
Phillip Fayers
>In article <slrnb59c8...@sadalsud.astro.cf.ac.uk>, Phillip Fayers wrote:
>> "It hasn't been that long since" - exactly. Sun haven't produced
>> a competitive desktop (at a reasonable price) for a long time.
>> That hole in their product line is now manifesting itself - the
>> majority of developers have moved to using Linux boxes.
>And that is why "open source" software is sucking more and
>more every day. For those of us on non-Linux Unix it is
I certainly aggre with you there. New developers are coming
along who think that as long as it works on Linux on x86 then
it must be OK. Gone aer any attempts to adhere to set standards
or ensure portability. The Linux crowd don't seem to have learned
the lesson that nothing lasts for ever. Linux is flavour of
the month (maybe flavour of the next few years) and when it dies
it'll take a load of the applications with it.
...
>The last time Sun had a "competitive hole" in their desktop
>line they came out with the Ultra 10. They sold more of
>those than any other desktop in their history.
And still they didn't learn. When they saw that they should
have stepped up development efforts at the low end to make
sure that they never had the same product line hole again.
>It will
>happen again when the UIIIi/jalapeno comes out (soon).
Soon. Now we really have been saying that for a long time.
The Jalapeno would have been a huge hit if it was available
this time last year. Sadly it wasn't.
>> On the application front Sun have never really been serious about
>> that one either. StarOffice was far too little and far too late
>Wrong. You have a limited view of *free* software. In many
>categories the premier applications are still on Sun. I
>think Sun claims something like 4,500 applications -
>far beyond ANY other Unix. And lots of people will pay
>extra for value - even if it is a commercial product.
You don't usually have a choice but to pay on Sun. Example:
We have a backup system, Veritas, a licence to backup a single
UNIX client costs 10 times a licence for a single Windows or
Linux machine. Running the server side on a Solaris box means
that the licence cost double what it would have if I'd put it
on NT. Example 2: A licence for Autocad on a Sun costs more
than for the same product on a faster PC. Not entirely Suns
fault I know but something they should have been working for
a long time and reflecting in their own pricing.
>> But again they are getting better faster than the desktop
>> apps on Solaris.
>Possibly. Now if those app developers cared about anybody
>else they would make those apps portable to Solaris and
>other Unixes. There is nothing about these desktop apps
>that is specific to Linux - other than the authors
>philosophical orientation.
As long as it works on Linux/x86 ...
>As I've suggested in the past, Sun should either merge
>with Apple or license the Apple desktop & apps. Wouldn't
>that be interesting.
Probably too late for that now. Sun had an escape route
lined up. Solaris on x86 and PowerPC but they didn't know
how to handle it.
>> The Linux hype will continue as long as IBM/HP etc see that
>> it hurts Sun (which it does).
>Agreed. That is what failing companies do - grab whatever
>they can to keep from sinking further.
>Fortunately, buyers will eventually see the truth. Linux
>is here to stay and it is good for some purposes. It
>is not the grail of operating systems.
Nothing is.
Rich Teer
> And that is why "open source" software is sucking more and
> more every day. For those of us on non-Linux Unix it is
> becoming harder and harder to get said software to compile
> and run. Back when Sun was the reference things were alot
> better (for more than just Sun users). Linux bigots just
> don't give a damn about anybody else. In that sense they
> are == M$.
That is very sad, and very true. What's even more ironic
is how insulting Linux users would find that statement!
> The Linux crowd has taken the phrases "open source" and
> "open" and twisted the meaning. Sun used to be considered
> the premier "open" platform. Partly due to the standards-
> based approach. Note that the "used to be" isn't because
> anything has changed with or on Sun - what has changed is
> a bunch of newbies have redefined the term, for the worse.
Indeed. :-(
> The last time Sun had a "competitive hole" in their desktop
> line they came out with the Ultra 10. They sold more of
> those than any other desktop in their history. It will
> happen again when the UIIIi/jalapeno comes out (soon).
I agree with the other (Phil Fayers?) poster who said that
when Sun noticed this hole in their market, they should have
done more to make sure it didn't happen again. I'm pretty
sure that US-IIIi based machines will be part of Sun's NC03Q2
announcement, but I hope they won't be too little, too late,
for too much.
> Wrong. You have a limited view of *free* software. In many
> categories the premier applications are still on Sun. I
> think Sun claims something like 4,500 applications -
Actually, they claim over 12,000 apps!
> Possibly. Now if those app developers cared about anybody
> else they would make those apps portable to Solaris and
> other Unixes. There is nothing about these desktop apps
> that is specific to Linux - other than the authors
> philosophical orientation.
Agreed again.
> As I've suggested in the past, Sun should either merge
> with Apple or license the Apple desktop & apps. Wouldn't
> that be interesting.
iDVD et al on SOlaris would be a neat idea!
> Fortunately, buyers will eventually see the truth. Linux
> is here to stay and it is good for some purposes. It
> is not the grail of operating systems.
Amen to that.
Raymond Toy
Dragan> Raymond Toy <t...@rtp.ericsson.se> writes:
>> We're a key customer for one of our vendors. We pleaded and begged
>> for years to get a Solaris version of their tools. All we got was a
>> cold shoulder, saying they have no real interest, even when their
>> other key customers would also like a Solaris version. Finally, they
>> produce a version. It's functional, but lacks some of the key
>> features that the Windows version has. And then they wonder why no
>> wants it, after saying for years we'll never get it. Idiots.
Dragan> If you are one of thier key customers, why didn't you say you will abandon
Dragan> them unless they produce a Solaris version? Or are they also one of the key
Dragan> vendors for you?
Both. The tools are for their chips and switching chips is extremely
hard too.
Ray
Michael T Pins
>On 20 Feb 2003, Steve Kappel wrote:
>> And that is why "open source" software is sucking more and
>> more every day. For those of us on non-Linux Unix it is
>> becoming harder and harder to get said software to compile
>> and run. Back when Sun was the reference things were alot
>> better (for more than just Sun users). Linux bigots just
>> don't give a damn about anybody else. In that sense they
>> are == M$.
>That is very sad, and very true. What's even more ironic
>is how insulting Linux users would find that statement!
While it is sad, it's not true. The only thing that has changed is that
free software is no longer based on "all the world is a Sun", but is now
"all the world is Linux/x86". Sun fanatics never understood this, as most
software used to compile out of the box for them. They're now discovering
what the rest of us have been putting up with for many years.
>> The Linux crowd has taken the phrases "open source" and
>> "open" and twisted the meaning. Sun used to be considered
>> the premier "open" platform. Partly due to the standards-
>> based approach. Note that the "used to be" isn't because
>> anything has changed with or on Sun - what has changed is
>> a bunch of newbies have redefined the term, for the worse.
>Indeed. :-(
Well, no. SunOS was never any more standard than Linux is today.
Solaris, today, is more standards compliant than SunOS4 ever was, but far
more software needs to be ported to it. Why? Because it's no longer the
dominant platform of people who have time to write free software.
>> Possibly. Now if those app developers cared about anybody
>> else they would make those apps portable to Solaris and
>> other Unixes. There is nothing about these desktop apps
>> that is specific to Linux - other than the authors
>> philosophical orientation.
>Agreed again.
And you can substitute "SunOS" for Linux in the above, and any other unix
for "Solaris", and it was true ten years ago. Again, nothing has changed
other than the platform of choice for people who have time to write free
software.
--
**************************************************************************
* Michael T Pins | mtp...@nndev.org *
* keeper of the nn sources | mtp...@visi.com *
* ftp://ftp.nndev.org/pub | #include <std.disclaimer> *
Anthony Mandic
> Microsoft seldom build good products; however the overall design of C# is
> fine, and probably there'll be a good implementation (probably not from
> Microsoft...)
Sure, sure. Its not pronounced "C hash" for nothing.
-am © 2003
Steve Kappel
Michael T Pins wrote:
> Well, no. SunOS was never any more standard than Linux is today.
> Solaris, today, is more standards compliant than SunOS4 ever was, but far
> more software needs to be ported to it. Why? Because it's no longer the
> dominant platform of people who have time to write free software.
SunOS *4* (and earlier) were of a different time - before
standards gained importance. But that was ~12+ years ago.
When you say SunOS I assume you mean SunOS 4. Note that
Solaris means SunOS 5 + environment.
Solaris (SunOS 5) is far more standard than Linux is today
or will be anytime in the forseeable future. We've already
established why - many Linux developers don't care about
standards because they don't care about anything other than
Linux.
Barry Margolin
Steve Kappel <ska...@visi.com> wrote:
>Solaris (SunOS 5) is far more standard than Linux is today
>or will be anytime in the forseeable future. We've already
>established why - many Linux developers don't care about
>standards because they don't care about anything other than
>Linux.
What you're seeing is the distinction between de jure standards and de
facto standards. What may be happening is that Linux's popularity has made
it a de facto standard among open-source software developers.
De jure standards are useful when writing contracts for software
acquisition and development, but de facto standards are often more
significant in the real world.
--
Barry Margolin, barry.m...@level3.com
Genuity Managed Services, Woburn, MA
*** DON'T SEND TECHNICAL QUESTIONS DIRECTLY TO ME, post them to newsgroups.
Please DON'T copy followups to me -- I'll assume it wasn't posted to the group.
Rich Teer
> Solaris (SunOS 5) is far more standard than Linux is today
> or will be anytime in the forseeable future. We've already
> established why - many Linux developers don't care about
> standards because they don't care about anything other than
> Linux.
That reminds me of a thread from a couple of years ago. SOme
Linux programmer wanted to use VGAlib for some application he
was writting. His reaction to people suggesting that he avoid
that, and use something like curses for portablity instead (and
the implictaion that hardware speedups will more than compensate
for the loss of efficiency)? "I don't care about portability".
He didn't change his position when someone pointed out that one
of the (supposedly) great things about Linux is that it's
available for many platforms, and that maybe people on those
platforms would like to use his software.
Yes, Linux is the Windoze of the GNU generation[tm]...
Rich Teer
> While it is sad, it's not true. The only thing that has changed is that
> free software is no longer based on "all the world is a Sun", but is now
> "all the world is Linux/x86". Sun fanatics never understood this, as most
> software used to compile out of the box for them. They're now discovering
> what the rest of us have been putting up with for many years.
Hmm. I've heard that it's much easier to port software from
Solaris to some other UNIX (even Linux), than vice versa. Why?
Apparently because there are so many easy to use Linuxisms that
newbie programmers aren't aware of.
> Well, no. SunOS was never any more standard than Linux is today.
I take it by SunOS, you're referring to SunOS 4.x and earlier,
which was EOLed how many years ago?
> Solaris, today, is more standards compliant than SunOS4 ever was, but far
> more software needs to be ported to it. Why? Because it's no longer the
> dominant platform of people who have time to write free software.
Right; but if the people writing the software wrote it in a
portable manner, it wouldn't be (so much of) an issue.
> And you can substitute "SunOS" for Linux in the above, and any other unix
> for "Solaris", and it was true ten years ago. Again, nothing has changed
> other than the platform of choice for people who have time to write free
> software.
The difference is that these days, there are more standards available,
and they're more important. I mean, a POSIX compliant app should
compile fairly easily on any other POSIX compliant version of UNIX.
But Linux isn't (yet) POSIX compliant, nor, from what I can see, is
POSIX compliance a high priority. N.b. I'm not talking about getting
Linux officially UNIX 98 or whatever branded - that costs money. But
the SUSv2 spec is freely available to anyone, so there's no reason
why the Linux developers can't make Linux SUSv2 compliant in principle.
Heck, they can't even use the "we don't want to break backwards
compatibility" chestnut, as it hasn't been a burden for them in
the past, from I gather (glibc nonsense, for example).
Steve Kappel
> In article <slrnb59sbl....@isis.visi.com>, Steve Kappel wrote:
>
>>The last time Sun had a "competitive hole" in their desktop
>>line they came out with the Ultra 10. They sold more of
>>those than any other desktop in their history.
>
> And still they didn't learn. When they saw that they should
> have stepped up development efforts at the low end to make
> sure that they never had the same product line hole again.
I agree. Not sure why they bumbled it again. I would
think some of these reasons:
1. Volume desktops are not profitable. The PC market has
commoditized desktops to the point of almost no margin.
2. Sun sales droids don't make commission of any significance
on desktops. They want to spend their time making maximum
$ - on servers.
3. Given 1 & 2, most business people would put their R&D
money into servers and not desktops.
4. It wouldn't be unreasonable to think that Mhz were
going to level off at some point for desktops. Increasing
it now is having diminishing returns. Yet Intel must keep
pushing it up because what sells is the #, not the real world
performance. How long can it keep going and still be
profitable for Intel? If desktop chips outperform their
server chips then how do they justify higher prices for
the server chips? If they drive server chips to
commodity prices and the sales volume continues to
level off... Intel is going to be in a world of hurt.
The same reason that Sun can't release USIIIi until they
are closer to releasing USIV. If the entire world of
computing devolves to commodity status then innovation is
going to come to a terrible halt.
It wouldn't have been unreasonable for Sun to think that
desktops weren't going to keep growing at this rate.
>>It will
>>happen again when the UIIIi/jalapeno comes out (soon).
>
> Soon. Now we really have been saying that for a long time.
> The Jalapeno would have been a huge hit if it was available
> this time last year. Sadly it wasn't.
Why has it been delayed? One line of reasoning is that
it had to be held back because of price/performance verses
USIII. That USIV had to be closer. If that is true then
they have had more time to refine it which could mean
substantial increases in performance. Don't count them
out yet. USIIIi & USIV could be home runs. They don't
need to become fastest processors - just get back into
the pack; the other legs of the triad will put Sun ahead
again.
> You don't usually have a choice but to pay on Sun. Example:
> We have a backup system, Veritas, a licence to backup a single
> UNIX client costs 10 times a licence for a single Windows or
> Linux machine. Running the server side on a Solaris box means
> that the licence cost double what it would have if I'd put it
> on NT. Example 2: A licence for Autocad on a Sun costs more
> than for the same product on a faster PC. Not entirely Suns
> fault I know but something they should have been working for
> a long time and reflecting in their own pricing.
That really is a shame. It is a historical problem
that the marketers that set prices still don't get it that
UNIX clients should be priced the same as equivalent
Windows or Linux clients.
Steve Kappel
> On 20 Feb 2003, Michael T Pins wrote:
>
>> While it is sad, it's not true. The only thing that has changed is that
>> free software is no longer based on "all the world is a Sun", but is now
>> "all the world is Linux/x86". Sun fanatics never understood this, as most
>> software used to compile out of the box for them. They're now discovering
>> what the rest of us have been putting up with for many years.
>
> Hmm. I've heard that it's much easier to port software from
> Solaris to some other UNIX (even Linux), than vice versa. Why?
> Apparently because there are so many easy to use Linuxisms that
> newbie programmers aren't aware of.
This is fact.
I've been a software engineer for over 20 years and have worked
on 10s of millions of lines of code that was portable to more
than 10 Unixes. Follow the standards and write code on Solaris
and all the rest will be easier.
There are 2 families - Windows and Unix. On the Unix side the
range of standards compliance is pretty wide - from Solaris
on one end to Linux on the other with lots of others in
the middle. I've seen companies try to go from HP/UX or
AIX to others; it wasn't a pretty sight.
Another poster suggested that Linux is a de facto standard.
That is certainly the way things are going. Which means
there are now 3 families - Windows, Unix, and Linux.
That isn't an improvement. It's really too bad; Linux
could have taken a much different path and been a much
better system today.
Emmanuel Florac
disait...
>
> The OmoronP stated that C# is "better designed than Java".
>
Hey, you should go for the "Turing test contest", I almost believed
you're real. Please congratulate your programmer for me, will you? Nice
job.
Barry Margolin
Steve Kappel <ska...@visi.com> wrote:
>Another poster suggested that Linux is a de facto standard.
>That is certainly the way things are going. Which means
>there are now 3 families - Windows, Unix, and Linux.
It's not too different from the situation of the previous decade, which was
Windows, AT&T Unix, and BSD Unix. While BSD and AT&T have been unified by
the various published Unix standards, Linux has come along.
>That isn't an improvement. It's really too bad; Linux
>could have taken a much different path and been a much
>better system today.
Yes. I thought an original goal of Linux was to be a free OS that could
be dropped in as a Unix replacement, implying that it should conform to the
important Unix-related standards.
One of the realities of open-source development is that there's no Director
of Engineering establishing priorities that direct the efforts of all the
developers. In the bazaar model of the development, people contribute
whatever they want; if it happens to improve standards conformance, that's
a happy occasion, but people are just as likely to implement things that
seem more useful to them than standards conformance.
Robin KAY
> One of the realities of open-source development is that there's no Director
> of Engineering establishing priorities that direct the efforts of all the
> developers. In the bazaar model of the development, people contribute
> whatever they want; if it happens to improve standards conformance, that's
> a happy occasion, but people are just as likely to implement things that
> seem more useful to them than standards conformance.
Well said. The principle of open source development is "if you want something
doing, hack it yourself". I think we have to accept that the lack of portability
among open source software is our problem rather than theirs.
--
Wishing you good fortune,
--Robin Kay-- (komadori)
Paul Eggert
> New developers are coming along who think that as long as it works
> on Linux on x86 then it must be OK. Gone aer any attempts to adhere
> to set standards or ensure portability.
I dunno, I often see cases where Solaris breaks something that works
on all other Unix and/or GNU/Linux boxes that aren't running Sun code.
For example, the implementation of NFS on Solaris breaks GNU "make",
because Sun's 'stat' system call fails with errno==EINTR when it is
interrupted on an filesystem mounted with the intr option, even if the
program has set the SA_RESTART signal-handling option. That is just
plain lame, since the whole point of SA_RESTART is to avoid having to
code around the errno==EINTR possibility. And this lame behavior
breaks programs like GNU make that work perfectly well on GNU/Linux
and other Unix hosts.
I complained about this to Sun recently, and they politely told me to
go jump in the lake. It's not a standards-conformance issue, since
Sun doesn't claim that NFS conforms to POSIX. (If you want POSIX
conformance, you have to stick with POSIX-conforming filesystems, and
NFS isn't one of them.) It's merely a case where Solaris is brain
damaged, and Sun doesn't want to fix it.
I agree that Sun conforms to POSIX better than GNU/Linux does, but I'm
not sure I agree that it's a more "standard" platform. They're about
the same: from an application writer's viewpoint, both OSes have their
own special quirks and disasters.
Phillip Fayers
>In article <slrnb5a28...@sadalsud.astro.cf.ac.uk>, Phillip Fayers wrote:
>> In article <slrnb59sbl....@isis.visi.com>, Steve Kappel wrote:
>> And still they didn't learn. When they saw that they should
>> have stepped up development efforts at the low end to make
>> sure that they never had the same product line hole again.
>I agree. Not sure why they bumbled it again. I would
>think some of these reasons:
>1. Volume desktops are not profitable. The PC market has
>commoditized desktops to the point of almost no margin.
Volume desktops can be profitable. They also get you better value
for your research dollars. All that cash spent on developing the
chip gets spread over more chip sales which results in a lower
research cost per chip.
>2. Sun sales droids don't make commission of any significance
>on desktops. They want to spend their time making maximum
>$ - on servers.
You don't need sales droids to sell desktops, or low end servers,
or anything costing less than, say $20,000. Just stick them on
the web and let people order them. Sales droids are only really
required when you have silly pricing policies - who has ever paid
the list price for an E3800 or above, for example.
>3. Given 1 & 2, most business people would put their R&D
>money into servers and not desktops.
R&D put into servers can also be used for desktops, if you do
it right. A couple of "not done right" examples from Sun. The
new V1280, based on different boards to the V480 and V880 which
means that you can't chuck the XVR-4000 into the V1280.
>4. It wouldn't be unreasonable to think that Mhz were
>going to level off at some point for desktops. Increasing
>it now is having diminishing returns.
...
But Intel aren't simply pushing MHz. They improve everything else
as well. Look at the memory bandwidth on a modern PC, 5+ years
ago if you wanted decent memory bandwidth you had to have a RISC
workstation. Currently PC design moves so fast that new technologies
are incorporated way faster than slow coach designers like Sun.
>If desktop chips outperform their
>server chips then how do they justify higher prices for
>the server chips?
Becuase the desktop chips are not capable of multiple CPU operation
and have smaller caches.
...
>> Soon. Now we really have been saying that for a long time.
>> The Jalapeno would have been a huge hit if it was available
>> this time last year. Sadly it wasn't.
>Why has it been delayed? One line of reasoning is that
>it had to be held back because of price/performance verses
>USIII. That USIV had to be closer. If that is true then
If that's true then it would confirm how monumentally stupid
Suns management have become. In the early days Sun used to obselete
their own systems faster than anyone else could do it for them.
It's true that a quad processor US-IIIi machine should be
able to compete directly with a V480. So what's the problem?
The quad US-IIIi - let's for the sake of this arguement call
it a V440 - would be cheaper to produce and could be sold for
less than a V480 whilst giving the same performance. Where's
the problem? The V440 increases your market share whilst still
making you money and driving up your volumes.
>they have had more time to refine it which could mean
>substantial increases in performance. Don't count them
Which, by the earlier logic, they won't release becuase it'll
be too fast compared to US-IV.
>out yet. USIIIi & USIV could be home runs. They don't
>need to become fastest processors - just get back into
>the pack; the other legs of the triad will put Sun ahead
>again.
Not only do they need to get back into the pack but they need
to show a commitment to stay there.
>That really is a shame. It is a historical problem
>that the marketers that set prices still don't get it that
>UNIX clients should be priced the same as equivalent
>Windows or Linux clients.
More than a shame. It shows more bad business sense. In the
Veritas case the software has been available on Solaris for
ages, the Linux port is recent so should actually cost more becuase
of the extra work involved.
Tom Tobin
> Emmanuel Florac <efl...@imaginet.fr> writes:
>
> Oh, Ghod, and I just noticed you're a cheese eating surrender
> monkey.
<stands up>
I, too, am a cheese-eating surrender monkey.
I'm not French, but I've just finished some cheese...
I have surrendered my future to a politician who has
made a monkey of me.
A politician who proves that democracy is just the
modern equivalent of the Divine Right of Kings, or
- in this case, the Divine Right of Bushes.
<sits down>
--
Tom.
Alex Buell
> I'm not French, but I've just finished some cheese... I have surrendered
> my future to a politician who has made a monkey of me.
>
> A politician who proves that democracy is just the modern equivalent of
> the Divine Right of Kings, or - in this case, the Divine Right of
> Bushes. <sits down>
Yeah but we all know what happned to people who believed in the Divine
Right to Rule. They were executed.
Steve Kappel
> In article <slrnb5af62....@isis.visi.com>, Steve Kappel wrote:
>>In article <slrnb5a28...@sadalsud.astro.cf.ac.uk>, Phillip Fayers wrote:
>>> In article <slrnb59sbl....@isis.visi.com>, Steve Kappel wrote:
>
>>> And still they didn't learn. When they saw that they should
>>> have stepped up development efforts at the low end to make
>>> sure that they never had the same product line hole again.
>
>>I agree. Not sure why they bumbled it again. I would
>>think some of these reasons:
>
>>1. Volume desktops are not profitable. The PC market has
>>commoditized desktops to the point of almost no margin.
>
> Volume desktops can be profitable. They also get you better value
> for your research dollars. All that cash spent on developing the
> chip gets spread over more chip sales which results in a lower
> research cost per chip.
More chip sales mean lower cost to the customer. Would
you want to pay the same price for a chip that sells 5,000,000
a year as one that sells 50,000 a year? Doesn't work for
Intel - most the savings are passed on to the customer.
Intel makes money because the volume is millions upon millions.
Ask Intel how much money they make on Itanium. The reason
they want it to succeed so badly is because they know they can
charge a lot more for it and make a lot more money. Of
course it will take a decade to recover the R&D. And
that assumes it is moderatly successful which it may not be.
>>2. Sun sales droids don't make commission of any significance
>>on desktops. They want to spend their time making maximum
>>$ - on servers.
>
> You don't need sales droids to sell desktops, or low end servers,
> or anything costing less than, say $20,000. Just stick them on
> the web and let people order them. Sales droids are only really
> required when you have silly pricing policies - who has ever paid
> the list price for an E3800 or above, for example.
What business people care about is margin. Low-end
products are a necessity - investment in them is done
grudgingly. Unless you are a company with no IP
(Intellectual Property) - but then you don't need
investment because somebody else has all the IP.
Sun is an IP-oriented business; there is no reason for
them to become anything else.
>>3. Given 1 & 2, most business people would put their R&D
>>money into servers and not desktops.
>
> R&D put into servers can also be used for desktops, if you do
> it right. A couple of "not done right" examples from Sun. The
> new V1280, based on different boards to the V480 and V880 which
> means that you can't chuck the XVR-4000 into the V1280.
Is there a market for a V1280 scale visualization product?
The V1280 has a 9.6 GB/sec fireplane; the V880 is 4.8 GB/sec.
It's called progress. The next incarnation of an 8 processor
box will probably have the faster interconnect.
I'd say Sun is one of the best companies for reusing
technology in multiple models.
>>4. It wouldn't be unreasonable to think that Mhz were
>>going to level off at some point for desktops. Increasing
>>it now is having diminishing returns.
> ...
>
> But Intel aren't simply pushing MHz. They improve everything else
> as well. Look at the memory bandwidth on a modern PC, 5+ years
> ago if you wanted decent memory bandwidth you had to have a RISC
> workstation. Currently PC design moves so fast that new technologies
> are incorporated way faster than slow coach designers like Sun.
And PC power is far beyond the needs of the vast majority
of users. That is why sales have slowed and the PC industry
has plateaued. The rate of PC technology advancement is
a desperate attempt to fight off a collapse. At some
point the R&D is going to become too expensive and it will
slow. Or the prices will go up.
>>out yet. USIIIi & USIV could be home runs. They don't
>>need to become fastest processors - just get back into
>>the pack; the other legs of the triad will put Sun ahead
>>again.
>
> Not only do they need to get back into the pack but they need
> to show a commitment to stay there.
For the vast majority of Sun's customers they aren't out
of the pack now. Remember the triad. The roadmap is
sufficient to prevent most defections.
If being at the head of the pack is so important, why
is Alpha dying?
Only a minority of customers need the bleeding edge
of performance. I'm sitting here typing this on my
Ultra 60 with 2 300Mhz UltraSPARCs - it is more than
enough for what I need. It happens to be in my home
and I paid less than $800US for it a year ago.
I wouldn't even consider a PC.
>>That really is a shame. It is a historical problem
>>that the marketers that set prices still don't get it that
>>UNIX clients should be priced the same as equivalent
>>Windows or Linux clients.
>
> More than a shame. It shows more bad business sense. In the
> Veritas case the software has been available on Solaris for
> ages, the Linux port is recent so should actually cost more becuase
> of the extra work involved.
That certainly isn't Sun's issue. Software pricing is
based on what the market will bear or for secondary
motives (like maintaining a monopoly). Companies are
going to have a hard time selling much Linux software
unless it is cheaper than Windows versions - since $ is
the driving force behind Linux.
Don't get me wrong, I'd love to have Sun selling desktops
that are equivalent in power to PCs, have SPARC binary
compatibility, cost similar to PCs, and applications
similar to Apple. But that isn't likely. So I have
to make a value choice. Other people have different
ways to measure value.
Baby Peanut
> On Mon, 17 Feb 2003 18:10:43 +0000, Rich Teer wrote:
>
> > On Mon, 17 Feb 2003, Dave Uhring wrote:
> >
> >> SMP capability, and limited capability at that, of 2 processors is
> >> "scalable"???
> >
> > In PeeCee land, yes!
>
> Casper pointed out quite some time ago that Solaris 8 x86 is capable of
> scaling to 21 processors. And that OS is a part of PeeCee land :-)
>
> Hell, even Linux can scale to 8 processors in a reasonably linear manner!
Tut-tut, must keep up with the actual state of the software (FreeBSD)
being discussed. FreeBSD 5.0 has SMPNG ("NG" as in "Next
Generation".)
YTC#1
> On Mon, 17 Feb 2003 15:22:22 +0000, Sniper wrote:
>
>
>>Well put, anyone knows that for secure scalable systems, people use
>>FreeBSd..
>
>
> SMP capability, and limited capability at that, of 2 processors is
> "scalable"???
>
Dave, sniper knows me IRL, he was tring to wind me up.
--
Bruce Porter
"The internet is a huge and diverse community
and not every one is friendly"
http://www.ytc1.co.uk
There *is* an alternative! http://www.openoffice.org/
YTC#1
> Captains Log Star Date Sun, 16 Feb 2003 21:47:13 +0000,alien life
> YTC#1 <y...@ytc1.co.uk> contacted us and said:
>
>
>>John Miller wrote:
>>
>>>"Baby Peanut" <baby_...@yahoo.com> wrote in message
>>>news:c5cf6e8.03012...@posting.google.com...
>>>
>>>
>>>>I think that the number one thing that killed DEC was Sun. DEC was
>>>>about closed systems and Sun was about open systems. Sun prospered
>>>>and DEC withered.
>>>>
>>>>Now it seems that Sun is past its peak, lost its vision of openness
>>>>and has turned into just another closed systems shop:
>>>>
>>>>http://makeashorterlink.com/?D1F324043
>>>>
>>>>How long before HP purchases what's left of Sun?
>>>
>>>
>>>Linux is the last great hope for SUN, the big question is Does SUN have the
>>>reserves to last out until Linux envelops the entire computer industry ?
>>>
>>
>>When linux has a stable stack, doesn't allow memory violations, has a
>>clean upgrade path that does not require re-compiles. Can utilse
>>multiple processors efficently and scalable. Then maybe it can be taken
>>more seriously on larger systems. Until then it is a disaster waiting to
>>happen.
>
>
> Well put, anyone knows that for secure scalable systems, people use
> FreeBSd..
>
>
>
FOYRNB
Baby Peanut
[...big snip...]
> One way to do that is through a merger, but the logical merger partner
> isn't Apple, it is Sony. The two companies have been talking about
> some kind of strategic alliance. Maybe these are merger talks. Sony
> is incredibly strong, having just posted its biggest-ever profit.
> Sony leadership is changing, making possible a bold move as the new
> management tries to put its own stamp on the company. Sony has both
> the resources to support Sun and the need for technology Sun can
> provide.
>
> Sony is a leader in consumer electronics and home entertainment, but
> not in computers. While the combined companies could field some very
> good computer offerings, extending Sony's influence into the server
> space, the real value of the combination lies in using Sun technology
> and know-how to transform Sony's current bread-and-butter businesses,
> which are TVs, video games, and movies.
>
> With Sun's help, Sony could redefine television, bringing it into the
> emerging broadband era. A Sony Internet TV could show Sony content
> received over a Sony global network, all engineered by Sun. It is a
> powerful attraction, and at around $3 per share, Sun is very
> affordable for Sony.
[...small snip...]
Seems that Sun and SONY are not strangers either.
http://www.sun.com/smi/Press/sunflash/2002-04/sunflash.20020408.4.html
http://www.sun.com/smi/Press/sunflash/2002-04/sunflash.20020408.6.html
http://www.sony.com/SCA/press/980327.shtml
Remeber also that SONY once tried to release a 68K workstation and it
basically was a spiff box that bombed.
Dave Uhring
> "Dave Uhring" <daveu...@yahoo.com> wrote in message news:<pan.2003.02.
> 17.18.31....@yahoo.com>...
>> Hell, even Linux can scale to 8 processors in a reasonably linear manner!
>
> Tut-tut, must keep up with the actual state of the software (FreeBSD)
> being discussed. FreeBSD 5.0 has SMPNG ("NG" as in "Next
> Generation".)
Tried to load it on a 4-way AlphaServer. Machine locked up as soon as the
2d processor was started.
Michael T Pins
>On 20 Feb 2003, Michael T Pins wrote:
>> While it is sad, it's not true. The only thing that has changed is that
>> free software is no longer based on "all the world is a Sun", but is now
>> "all the world is Linux/x86". Sun fanatics never understood this, as most
>> software used to compile out of the box for them. They're now discovering
>> what the rest of us have been putting up with for many years.
>Hmm. I've heard that it's much easier to port software from
>Solaris to some other UNIX (even Linux), than vice versa. Why?
>Apparently because there are so many easy to use Linuxisms that
>newbie programmers aren't aware of.
I've "heard" many things over the years, many of them untrue. It's as easy
to write Solaris specific code as it is to write Linux specific code. The
difference is, most of the people who have time to write free software are
using Linux instead of Solaris these days. Once again, for the people who
missed it the first three times, the only difference is what platform we
need to port from. Those of us who port code have been dealing with these
issues for nearly twenty years, only the "default" platform has changed.
>> Well, no. SunOS was never any more standard than Linux is today.
>I take it by SunOS, you're referring to SunOS 4.x and earlier,
>which was EOLed how many years ago?
SunOS 4.1.4 is still under support until the end of September. Your point?
SunOS 5.x has never been the dominant desktop unix platform. SunOS4 used
to be, and Linux is today.
>> Solaris, today, is more standards compliant than SunOS4 ever was, but far
>> more software needs to be ported to it. Why? Because it's no longer the
>> dominant platform of people who have time to write free software.
>Right; but if the people writing the software wrote it in a
>portable manner, it wouldn't be (so much of) an issue.
Which has always been true, and has always been ignored by 90% of the
people writing free software.
>> And you can substitute "SunOS" for Linux in the above, and any other unix
>> for "Solaris", and it was true ten years ago. Again, nothing has changed
>> other than the platform of choice for people who have time to write free
>> software.
>The difference is that these days, there are more standards available,
>and they're more important. I mean, a POSIX compliant app should
>compile fairly easily on any other POSIX compliant version of UNIX.
>But Linux isn't (yet) POSIX compliant, nor, from what I can see, is
>POSIX compliance a high priority. N.b. I'm not talking about getting
>Linux officially UNIX 98 or whatever branded - that costs money. But
>the SUSv2 spec is freely available to anyone, so there's no reason
>why the Linux developers can't make Linux SUSv2 compliant in principle.
Are there more standards available? Possibly.
Are they more important? Not in the real world.
Yes, it would be nice if people wrote portable code to begin with, but it
didn't happen when "all the world is a Vax", it didn't happen when "all the
world is a Sun", and it's unlikely to happen now that "all the world is
Linux/x86".
The really ironic thing is that as Sun was moving to a more standards
compliant OS, the default desktop unix OS was moving away from Sun.
Paul Eggert
> SunOS 4.1.4 is still under support until the end of September.
Sun may call it "support", but it's not really support. If you find a
new bug in SunOS 4.1.4, and you have a standard software support
contract, then Sun is not obligated to (and probably will not) issue a
patch. That may well be smart business, but it's not "support" in my
book.
My impression is that Sun doesn't fix even old, well-known bugs in
SunOS 4.1.4 any more. For example, this bug in the SunOS 4.1.4
/bin/sh:
Sun bug description
1123136 /bin/sh uses wrong PID for name of the temporary file created by cat << EOF
1134744 in-line input redirection into awk after exec of stdin and stdout
1170383 "here documents" within "command substitutions" fail much of the time in 4.1.x
has been around for ages, and Sun hasn't fixed it. The bug breaks
valid "configure" scripts generated by recent versions of GNU
Autoconf. It's not easy for Autoconf to work around the bug, so the
GNU Autoconf developers have decided to stop supporting SunOS 4.1.4
/bin/sh. Nobody uses SunOS 4.1.4 these days for new applications, so
there's little point to having Autoconf support it. And if you're a
computer museum operator and really want to port to that ancient OS,
you can work around the problem by installing an old Bash version
first, and using it to bootstrap.
Many longtime developers have rose-colored opinions about how
"standard" and "stable" SunOS 4.x was. 4.x was pretty good by the
standards of the day, but that was a long time ago. The Debian
GNU/Linux x86 box I'm typing this on (a new $249 x86 server from Dell
+ $30 graphics card + $800 monitor) is far more stable and reliable
than the SunOS SPARCstation 1 that was on my desk in 1989 ($13,000,
including monitor).
Emmanuel Florac
disait...
>
> Oh, Ghod, and I just noticed you're a cheese eating surrender
> monkey.
>
Yep. Like 75% of people on this planet or so (well about the so-called
surrender side, slightly less for cheese I suppose). Actually if you were
real you couldn't have enough brain NOT to believe any bullshit written
on the Sun toilet paper (looks in-chart, isn't it;), but who cares?
You're not real (thanks god).
> But thanks for reminding me I'd forgotten to killfile you.
>
nice proggie, really. Is the source available somewhere?
Anthony Mandic
> Tut-tut, must keep up with the actual state of the software (FreeBSD)
> being discussed.
Why's that peanut brain?
> FreeBSD 5.0 has SMPNG ("NG" as in "Next Generation".)
Like changing the name is going to make that crap run any better.
-am © 2003
Dave Uhring
FreeBSD does indeed have improved SMP performance in the newest version
thanks to code provided by BSDi. But that improvement seems to work only
on i86 so far. And even on that architecture FreeBSD-5.0 is not yet ready
for prime time.
John Miller
> So here is the prognosis. Sun lost $2 billion last year and will
> probably lose another $2 billion this year. At that rate, the company
> has at most five years to live.
The cost of all the layoffs and changes required to succeed will wipe out
Suns reserves. In 12 months time from now with another $2 Billion loss, the
company won't have the required reserves it needs to change and adapt for
the future, it would be more cost effective to close the company, pay the
redundancy and debt charges and return any remaining funds to the
shareholders. The situation is that bad.
>They have just renewed a commitment
> to the Solaris operating system, which is no longer really viable from
> an economic standpoint. I know, I know, Solaris users love Solaris,
> but they don't love Solaris prices. And with a falling market share,
> Sun can't afford to make Solaris any cheaper.
Solaris is economically bankrupt, the only people who support it are those
who think it makes them super intelligent to name drop it in every
conversation. The film is shit the operating system even worse.
>Sun is having the same
> problem in hardware where their SPARC architecture is falling behind,
> and -- worse still -- has lost nearly all of its manufacturing support
> in Japan. Both Solaris and SPARC will absorb vast sums in the coming
> years and yield absolutely no increase in Sun's market share as a
> result.
Sun needs to share costs if it's going to go forward with SPARC, which means
it needs a strong asian partner like Samsung to lead the R&D and
manufacturing effort and to lay off all it's american people who are
currently working on it.
Steve Kappel
Can you send me some of whatever it is that you're smoking?
Or are you related to Bill G?
Georges A. Tomazi
John -
On Sat, 22 Feb 2003 15:20:32 GMT, "John Miller"
<starl...@yahoo.co.uk> wrote:
[...]
>Solaris is economically bankrupt, the only people who support it are those
>who think it makes them super intelligent to name drop it in every
>conversation. The film is shit the operating system even worse.
I don't want to get into some religion war, and I don't pretend having
such an extensive knowledge in the computer market to judge if Solaris
is economically bankrupt. However, I have about 18 years of Unix
experience, from Multics to Linux, including some
OpenVMS/Windows/MacOS/Novell exposure, and I guess that helps me
having a pretty good and objective view of the pros and cons of many
different environments.
All that to say that until now, I haven't found a reliable, scalable
and fully featured multi purpose operating system such Solaris 2.
I don't know if that makes me "super inteliigent" but talking from a
technical point of view, the facts are there. Period.
[...]
Georges
--
Georges A. Tomazi - g...@sunwizard.net
Rich Teer
[A load of BS as usual]
___________________
/| /| | |
||__|| | Please do |
/ O O\__ NOT |
/ \ feed the |
/ \ \ trolls |
/ _ \ \ ______________|
/ |\____\ \ ||
/ | | | |\____/ ||
/ \|_|_|/ \ __||
/ / \ |____| ||
/ | | /| | --|
| | |// |____ --|
* _ | |_|_|_| | \-/
*-- _--\ _ \ // |
/ _ \\ _ // | /
* / \_ /- | - | |
* ___ c_c_c_C/ \C_c_c_c____________
rrbrown
everyone here is adults debating about the fate of a company we all have
grown to love ? hate etc... if you dont really have anything important to
say why dont you just ignore the thread sheesh..
- RB
"Rich Teer" <rich...@rite-group.com> wrote in message
news:Pine.GSO.4.44.0302221042460.17439-100000@electron...
Steven Hill
>
> Or are you related to Bill G?
He was smoking Bill Gates' cock.
--
Steven Hill
``Libenter homines id quod volunt credunt''
Steven Hill
> who think it makes them super intelligent to name drop it in every
> conversation. The film is shit the operating system even worse.
Or perhaps those who see it as a useful, robust operating environment for
the datacenter? Since linux on a 16 processor box just doesn't work
reliably.
How about a nice bug cup of shut the fuck up?
Can't comment on the film though.
--
Steven Hill
``It is by will alone I set my mind in motion.
It is by the juice of Sapho that thoughts acquire speed,
the lips acquire stains, stains become a warning.
It is by will alone I set my mind in motion.''
- Piter De Vries, Dune.
Emmanuel Florac
s...@waroffice.net disait...
>
> Or perhaps those who see it as a useful, robust operating environment for
> the datacenter? Since linux on a 16 processor box just doesn't work
> reliably.
>
Did you by any chance hear about the SGI Altix? Up to 64 CPUs with a
single Linux image.
--
In girum imus nocte ecce et consumimur igni
John Miller
"Emmanuel Florac" <efl...@imaginet.fr> wrote in message
news:MPG.18c35dd26...@news.free.fr...
> Dans article <Pine.LNX.4.33.03022...@homer.mathie.cx>,
> s...@waroffice.net disait...
> >
> > Or perhaps those who see it as a useful, robust operating environment
for
> > the datacenter? Since linux on a 16 processor box just doesn't work
> > reliably.
> >
>
> Did you by any chance hear about the SGI Altix? Up to 64 CPUs with a
> single Linux image.
Thankfully the ignorance shown by solaris users isn't shared with Sun
management who are moving towards Linux as their standard OS at top speed.
Anthony Mandic
> Solaris is economically bankrupt, the only people who support it are those
> who think it makes them super intelligent to name drop it in every
> conversation. The film is shit the operating system even worse.
Sounds like you have a severe case of sour grapes because Linux is so
crappy it can't make it into the enterprise and no one will ever get
a decent job with it - let alone make a movie about it.
-am © 2003
Dave Uhring
[nothing of any consequence]
begin virus.vbs
Steven Hill
> single Linux image.
That they have just brought out this year, and no doubt have a heavily
customised distro, and a kernel that has been rewritten.
16 processor sun servers have been around for quite some time. SC2000 for
example. I am surprised you didn't wheel out the IBM z/OS with Linux
running on partitions.
I am a believer in right tool for the job. Solaris is the right tool for
some of the stuff I was doing. Furthermore, you are also a tool.
Eyes front. Nothing more to see here.
--
Steven Hill
``Neca eos omnes. Deus suos agnoscet.''
Emmanuel Florac
s...@waroffice.net disait...
>
> That they have just brought out this year, and no doubt have a heavily
> customised distro, and a kernel that has been rewritten.
>
No, standard distro (RH advanced server).
> 16 processor sun servers have been around for quite some time. SC2000 for
> example.
Sun machines with lots of processors sucked quite recently (SMP vs NUMA).
> I am surprised you didn't wheel out the IBM z/OS with Linux
> running on partitions.
>
This is a different story.
> I am a believer in right tool for the job. Solaris is the right tool for
> some of the stuff I was doing.
Solaris is the right tool for a lot of things. Linux may be the right
tool for some identical things, and some other different things.
However, you just can't ignore it, it' gaining huge momentum, and not
only for bad reasons.
Unix Guru
Dreamer. Sun has provided Linux for the bottom feeders. Did you know
that there was a big fight internally at Sun about the Intel box?
Some (Cobalt lusers) wanted it to run Linux only. After many Sun
customers slapped some sense back into Sun management we now have
Solaris x86 fully supported. So no, Sun management isn't moving
towards Linux at top speed.
Unix Guru
> s...@waroffice.net disait...
> >
> > Or perhaps those who see it as a useful, robust operating environment for
> > the datacenter? Since linux on a 16 processor box just doesn't work
> > reliably.
> >
>
> Did you by any chance hear about the SGI Altix? Up to 64 CPUs with a
> single Linux image.
Does it work? Does it scale? Will it prevent SGI from going out
of business? Unlikely.
Emmanuel Florac
unix...@engineer.com disait...
> After many Sun
> customers slapped some sense back into Sun management we now have
> Solaris x86 fully supported.
That is indeed a GOOD thing, that Solaris x86 has been brought back to
life (at least for Sun).
Emmanuel Florac
unix...@engineer.com disait...
> >
> > Did you by any chance hear about the SGI Altix? Up to 64 CPUs with a
> > single Linux image.
>
> Does it work?
Yes. Actually they sell it like hot cakes (relatively to its price), and
they worked on it for 4 years now, it's really solid stuff.
> Does it scale?
No doubt about it. The 256*Itanium1 version has run for 2 years now...
> Will it prevent SGI from going out
> of business? Unlikely.
Following market rules, probably not. Right now it seems that SGI depends
more and more on DOD and other governement/military contracts to
keep afloat.
--
Sutor ne ultra crepidam.
Mike Jones
efl...@imaginet.fr says...
> Dans article <84d35129.0302...@posting.google.com>,
> unix...@engineer.com disait...
> > > Did you by any chance hear about the SGI Altix? Up to 64 CPUs with a
> > > single Linux image.
> > Does it work?
> Yes. Actually they sell it like hot cakes (relatively to its price), and
> they worked on it for 4 years now, it's really solid stuff.
> > Does it scale?
> No doubt about it. The 256*Itanium1 version has run for 2 years now...
That's not the real answer, though. The interesting question is not
"will it run with 256 CPUs", but "how many times the 1-CPU performance o
you get with 256 CPUs"? I haven't seen the numbers on the F15K, but the
E10K performance with 64 CPUs was ~58 times the single-CPU performance
over a fairly wide range of applications (including scenarios with
multiple applications). NUMA-type systems tend to run in the ~40x
single-CPU range if they're well tuned and the problem is partitionable
into chunks that will fit into local node memory (i.e., best case).
--
Mike Jones
Government is not establish'd merely by Power; there must be maintain'd
a general Opinion of its Wisdom and Justice, to make it firm and
durable.
-- Benjamin Franklin
Emmanuel Florac
m...@nycap.rr.com disait...
> NUMA-type systems tend to run in the ~40x
> single-CPU range if they're well tuned and the problem is partitionable
> into chunks that will fit into local node memory (i.e., best case).
>
Funnily, I always met benchmarks stating exactly the opposite: NUMA
scales almost perfectly as far as you want, SMP can't keep up over 8 CPUs
or so. Especially the E10K won't scale that well AFAIK...
Actually if I remember well, Sun used to bash NUMA as long as they didn't
have NUMA machines to announce.
Baby Peanut
Mon Feb 24, 1:45 PM ET
Add Technology - NewsFactor to My Yahoo!
Brad Hill, www.EcommerceTimes.com
In the latest round of a corporate boxing match that has persisted for
more than five years, Microsoft (Nasdaq: MSFT - news) has filed a
court brief accusing Sun Microsystems (Nasdaq: SUNW - news) of unfair
competition and violation of a previous settlement agreement. The
brief is part of an ongoing private antitrust case brought against
Microsoft by Sun.
At issue is Microsoft's freedom to distribute its own version of Java
bundled with Windows. Java is a cross-platform programming language
that delivers applications to the desktop, sometimes embedded within a
Web browser. Sun has claimed that Microsoft's version of the language
is incompatible with Sun's version, creating problems for users and
violating the Java licensing agreement.
Case Background
The ongoing antitrust case is the second suit filed over this issue.
The original complaint was brought by Sun in October 1997. Just over
three years later, in January 2001, Microsoft settled that case by
paying Sun US$20 million.
The second lawsuit, heard by U.S. District Judge J. Frederick Motz, is
broader in scope. Sun claims Microsoft is attempting to persuade
developers to write programs for the .NET platform instead of Java,
but the case also includes complaints that Microsoft has refused to
export its Office suite to non-Windows operating systems, and that
Microsoft is trying to pressure enterprise customers to use its
Exchange and SQL Server products.
On January 21st, Motz issued a preliminary injunction ordering
Microsoft to remove its proprietary version of Java and begin
distributing Sun's version with the Windows operating system. That
injunction was quickly put on hold by a circuit court in Richmond,
Virginia, to which Microsoft appealed. The appeal will be argued by
both sides April 3rd.
But Giga Information Group research fellow Rob Enderle told the
E-Commerce Times that there is now an inherent contradiction in Sun's
position. By thwarting Microsoft's ability to distribute Java in the
first case, then complaining that Microsoft is not using Java enough
in the second case, Sun is on untenable ground, according to Enderle.
"Microsoft is in a stronger position," he said.
Alleged Violations
Microsoft's new counterclaim accuses Sun of two violations. First, the
software giant alleges that Sun's prosecution goes against
California's Business and Professional Code 17200, which stipulates
fair business practices. The code supports litigious action against
"unfair, unlawful, or fraudulent practice."
Microsoft asserts that Sun has brought the antitrust case "willfully
and deliberately with an intent to cause competitive injury to
Microsoft and to aid Microsoft's competitors." In so doing, according
to the brief, Sun is "seeking to relitigate the United States
government's action against Microsoft."
At the same time, Microsoft also is arguing that Sun's case has
breached the 2001 Java settlement terms, depriving Microsoft of "quiet
enjoyment" of the technology it licensed under that settlement.
According to the software giant, Sun agreed that Microsoft could
distribute its own version of the Java Virtual Machine until 2008.
Sun has complained about the method by which Microsoft chose to
distribute that technology -- the software giant offered Sun's Java to
Windows XP (news - web sites) users as an optional download, rather
than bundling it with Windows. Microsoft has countered that Sun had no
comment on this arrangement when Microsoft first announced it in July
2001, but instead waited eight months before launching its complaint.
No Mincing of Words
True to the acrimonious spirit between these two arch-rivals, the
language of Microsoft's latest court filing is caustic and
contemptuous. The brief refers to "Sun's technologically inferior Java
software" and claims that "Sun [seeks] to obtain a free ride on
Microsoft's success." Pressing the point with deepening scorn,
Microsoft declares, "Sun claims that this success -- earned through
years of costly research and development -- obligates Microsoft to
shore up Sun's business and compensate Sun for its own failures."
Microsoft also refers to Sun's complaints as "antitrust epithets."
Asserting a we-can-do-it-better attitude, Microsoft's litany of
insults about Java range from incompatibilities within Sun's own
product to a slam against all cross-platform technologies aimed at the
lowest common denominator so that they will work with several
operating systems. Specifically, Microsoft claims that Java
applications run poorly and consume too much memory.
"Microsoft smells blood, because Sun is stumbling," Yankee Group
software analyst Laura DiDio told the E-Commerce Times. However, she
added, the two companies are evenly matched on the rhetorical
battleground. "Sun has had its pens dipped in acid for longer than
Microsoft has."
Microsoft has not specified an amount it seeks in monetary damages.
That amount, the company says, should be determined by the court.
Chris Newport
There are lies, damn lies, and benchmarks.
Having done tests on this I can categorically state that the E10k
scales very well up to 64 processors and 256G memory in a single
domain under Solaris. The Linux kernel, however, scales rather
badly on the same configurations with 24 processors being the usefull
limit on the E10k.
The E10k reached the end of growth with the 480MHz processors
because the backplane could become a limiting factor. Hence the
new Fireplane system in the SunFire range, which has the potential
for growth with up to 108 processors for the next several years.
The SMP/NUMA argument is not really valid for either of these systems
simply because they are not classic SMP or NUMA architectures.
In any multiprocessor system the problems arise when a context
moves from one processor to another because the cache is no longer
valid. Good SMP kernels avoid the coherency problem by maintaining
a sticky relationship between processor and context. The CC in CC/NUMA
means that that NUMA maintains cache coherance by continual copying
and/or the use of shared cache. There are obvious limits to the
scaleability of classic CC/NUMA which dictate an efficiency which falls
rapidly as the number of processors increases.
This is the point at which the performance becomes highly dependant
on applications. A typical commercial system typically uses multiple
processes, each of which is single threaded, so the ideal situation
has one CPU per simultaneous process, minimising context switches.
This is the case where SMP wins. Many scientific applications are
highly parallel, which at first sight means that CC/NUMA is the
obvious choice. The poor scaleability of CC/NUMA, however, means that
hybrid solutions as used in the Fireplane will almost always win in
the vast majority of applications using more than about 16 processors.
Naturally is is possible to design a benchmark which will prove almost
anything in such complex situations, and compiler design becomes a
major issue.
Emmanuel Florac
c...@NOSPAM.netunix.com disait...
> The poor scaleability of CC/NUMA, however, means that
> hybrid solutions as used in the Fireplane will almost always win in
> the vast majority of applications using more than about 16 processors.
Yes, you said it. However that's the very first time I hear that; I was
always told the opposite way for years, go figure.
> Naturally is is possible to design a benchmark which will prove almost
> anything in such complex situations, and compiler design becomes a
> major issue.
>
Sure. So we never know, except that rea lity is made of 1 and 2
processors machines...