He's speaking to businessmen who buy
particular vendor's products. They
didn't buy DOS, they bought MS-DOS,
and ignored DR-DOS.
Similarly they buy Red Hat in the U.S.,
so he's obviously adressing U.S. businessmen.
If he were adressing German businessmen
he'd have said "SuSE's Linux".
In neither case would I expect him to say
"version of". The listener is expected to
get that from context.
surprise_audit wrote: I don't know that it would be fraud exactly, because I wouldn't getting money from you under false pretences.
Fraud is defined a bit wider than that, at least
in Canada. The competitor is attempting to make Google
take money from the advertiser under false pretences. Google, not unexpectedly, disapproves!
The Solaris x86 IBSC equivalent has
been around for about five years: I remember
a colleague working on it when I
was a new-hire. Janus seems
to be IBSC for SPARC...
I think you're misunderstanding what I said:
the Samba team is concerned about anyone
contributing who was legally bound to an
agreement with Microsoft. Anyone bound by
that agreement who contributes, taints
the Samba source code.
Samba implements a protocol which
was analyzed "off the wire", and so is
not legally encumbered by a license. They
wish to stay that way.
Not just FUD, but also lock-in. Please see
the warning at the
Samba Development page: In order to avoid any potential licensing issues we also ask that anyone who has signed the Microsoft CIFS Royalty Free Agreement not submit patches to Samba, nor base patches on the referenced specification.
Anyone who voluntarily licenses, for example, eating fish, must then abide by the
fish-eating license (:-))
And it even deals with the biggest problem
of Indian outsourcing, as you mentioned: timezones. The person you
need to deal with is usually on the
opposite side of the world,
so you get up at 5 AM to talk to them
at 5 PM and hope they haven't left
for the day.
With a Canadian base, the outsourcer
opens an office in the same time zone,
and staffs a call center and maintenance
team there.
The law is surprising logical, it just has
a very practical bent, so you find some
very fine distinctions... stating one
of the decisions in case law in logical
terms ends up looking something like
for all accused-of-fraud X, there exists a person to go to trial T, such that prequisites p, q, r, s and t and corequisites (u but not w), v and not (w, x y z) all apply
We used to actully do such translations
at U of WIndsor, which had both a school of law
and of philosophy. I don't claim the lawyers
thought this was logical, though (;-))
The buses have lots of bandwidth, but poor
response times. Running lots of threads
on lots of decoders (in this case, cores)
means that while the latency is still
horrid, there's programs running on the
data that arrived earlier...
Erk! You don't want to use a Windows terminal server, that's a money-pit. Use the VMware or
Win4Lin terminal-server offerings, and recycle your
old Window 9x licences instead of buying new CALs.
I use Win4Lin to run those last, annoying
Windows-only programs under, and it runs
faster than Windows on the same hardware.
An MMU and a real filesystem have some
definite benefits (;-)).
Threading is best done in an environment which
understands threads and monitors, where by
environment one usually means language.
I used threading in PL/1 (processes) and C (threads), and the amount of reinvent-the-wheel
I had to do was really excessive. I eventually
learned threading via Brinch
Hansen's Concurrent Pascal, where classes
and monitors (the model for Java's "protected classes") were first-class
objects in the language and runtime.
Net result? I don't want to do threads
in C ever again. Much like I never want to
do typsetters in assembler again. So I'll
use fork.
Getting a large quantity of Sun stock
at the current very low price might be a
useful move for Kodak. After all, the
only time "buy low, sell high" works is
when something is thought to be not very
valuable.
Getting about a quarter of Sun's cash
on hand might be resisted a lot harder.
Excuse me, is that not imposing a changed set of
terms after the offer and acceptance? If so,
it's a breach of the oldest commercial law ever,
the (methinks 16th century) statute of frauds.
I think that in Canada, they would be
selling a program with terms and conditions
concealed from the purchaser, and if so
the terms would have to be strictly in keeping
with normal business practice to prevent it from
being considered a changed term.
I wonder what the U.S. variant of
the statute of frauds implies...
--dave
Who isn't a laywer, but who taught
civil (civilian) and military law in the reserve force, way back in his ill-spent youth.
Time to turn processor affinity up higher, or
pbind Oracle to a particular set of processors.
And you're right, you don't know what an
app is going to do on a new platform or a much
larger set of processors. My favorite
hear-tearing situation is the application that
only scales to 4 or so processors, and then
degrades as you add more. Alas, that's
not uncommon!
--dave (I used to be a performance tuner (;-)) c-b
AKAImBatman wrote: Solaris is designed around high availability, easy problem diagnosis, and fault recovery. In exchange it sacrifices speed and kernel size.
Although it isn't mentiond in this discussion,
a big selling point of Solaris is that it's
setadily optimized and tuned to scale to
large numbers of processors. 64 or so, if
you really need that many (;-)).
Every few years the Solarii notice that
they've undertuned uniprocessors, and put
a push on to get uniprocessor performance
back. This happened around 2.5.1, and
for x86, around the boundary between 9 and 10. A former colleague
who runs Solaris 10 on x86 said "it sure ain't
Slolaris any more".
--dave (I run Solaris on SPARC and Linux on Intel, myself) c-b
I think Groklaw is a little too trusting of Z-D, but
it's an honestly held opinion, and I support them holding it.
I happen to disagree with it, and see Sun as
more interested in good engineering than FUD for
PHB's. Look at the articles on their multithreaded
cores. That's an elegant engineering approach to
closing the brutal speed gap between main memory and
the CPU. That's the stuff I'm interested in.
Well, folks like Z-D who get paid by the attention
they draw do tend to overstate and oversimplify a teentsy bit (;-))
For the large-system, huge-customer market,
this might be a real Sun marketing push. The
PHBs at that level are unlikely to recognize
the word Linux, but have heard about Red Hat, as they've read about it in Forbes.
For these folks, Sun might well need to say
"Sun is bettter than Red Hat". For a consultant or a publisher specializing in talking to PHBs,
this sort of ovrsimplification actually sounds
intelligent. That might not be my opinion,
but then I'm Dilbert, not Wally.
Ok, assume you have a full 19" rack of dual
1U Intels. Into the same rack you can
put 3U blade chassis, with 16 blades per,
and 32 logical CPUs per blade.
The benefit is (1/3) * (16*32/2)
or something like 85.3 to one.
And the chip that they stareted with is
pretty good at graphics, bitblit and
floaty-point.
No,actually it's for any cloud of applications
that run on a machine where the memory speed is much less than the cpu speed. This is true of everything
save 390s (;-))
Similarly they buy Red Hat in the U.S., so he's obviously adressing U.S. businessmen. If he were adressing German businessmen he'd have said "SuSE's Linux".
In neither case would I expect him to say "version of". The listener is expected to get that from context.
--dave
Congrats to Tridge and to Benjamin Pierce for Unison.
--dave
Fraud is defined a bit wider than that, at least in Canada. The competitor is attempting to make Google take money from the advertiser under false pretences. Google, not unexpectedly, disapproves!
--dave
Sure enough, one of his colleagues in name withheld by request got to pocket a substantial bonus.
--dave (who used to work for Sun) c-b
Samba implements a protocol which was analyzed "off the wire", and so is not legally encumbered by a license. They wish to stay that way.
Which protocol? Every protocol.
--dave
-- dave
Not just FUD, but also lock-in. Please see the warning at the Samba Development page: In order to avoid any potential licensing issues we also ask that anyone who has signed the Microsoft CIFS Royalty Free Agreement not submit patches to Samba, nor base patches on the referenced specification.
Anyone who voluntarily licenses, for example, eating fish, must then abide by the fish-eating license (:-))
--dave
With a Canadian base, the outsourcer opens an office in the same time zone, and staffs a call center and maintenance team there.
Evil, but elegant.
--dave
We used to actully do such translations at U of WIndsor, which had both a school of law and of philosophy. I don't claim the lawyers thought this was logical, though (;-))
The buses have lots of bandwidth, but poor response times. Running lots of threads on lots of decoders (in this case, cores) means that while the latency is still horrid, there's programs running on the data that arrived earlier...
--dave
For an example with Project, see my old article, Windows Compatability on the Linux Desktop.
--dave
I used threading in PL/1 (processes) and C (threads), and the amount of reinvent-the-wheel I had to do was really excessive. I eventually learned threading via Brinch Hansen's Concurrent Pascal, where classes and monitors (the model for Java's "protected classes") were first-class objects in the language and runtime.
Net result? I don't want to do threads in C ever again. Much like I never want to do typsetters in assembler again. So I'll use fork.
--dave
Getting a large quantity of Sun stock at the current very low price might be a useful move for Kodak. After all, the only time "buy low, sell high" works is when something is thought to be not very valuable.
Getting about a quarter of Sun's cash on hand might be resisted a lot harder.
--dave
--dave (ask a U.S. lawyer) c-b
I think that in Canada, they would be selling a program with terms and conditions concealed from the purchaser, and if so the terms would have to be strictly in keeping with normal business practice to prevent it from being considered a changed term.
I wonder what the U.S. variant of the statute of frauds implies...
--dave
Who isn't a laywer, but who taught civil (civilian) and military law in the reserve force, way back in his ill-spent youth.
And you're right, you don't know what an app is going to do on a new platform or a much larger set of processors. My favorite hear-tearing situation is the application that only scales to 4 or so processors, and then degrades as you add more. Alas, that's not uncommon!
--dave (I used to be a performance tuner (;-)) c-b
Although it isn't mentiond in this discussion, a big selling point of Solaris is that it's setadily optimized and tuned to scale to large numbers of processors. 64 or so, if you really need that many (;-)).
Every few years the Solarii notice that they've undertuned uniprocessors, and put a push on to get uniprocessor performance back. This happened around 2.5.1, and for x86, around the boundary between 9 and 10. A former colleague who runs Solaris 10 on x86 said "it sure ain't Slolaris any more".
--dave (I run Solaris on SPARC and Linux on Intel, myself) c-b
I happen to disagree with it, and see Sun as more interested in good engineering than FUD for PHB's. Look at the articles on their multithreaded cores. That's an elegant engineering approach to closing the brutal speed gap between main memory and the CPU. That's the stuff I'm interested in.
--dave
For the large-system, huge-customer market, this might be a real Sun marketing push. The PHBs at that level are unlikely to recognize the word Linux, but have heard about Red Hat, as they've read about it in Forbes.
For these folks, Sun might well need to say "Sun is bettter than Red Hat". For a consultant or a publisher specializing in talking to PHBs, this sort of ovrsimplification actually sounds intelligent. That might not be my opinion, but then I'm Dilbert, not Wally.
--dave
--dave
The benefit is (1/3) * (16*32/2) or something like 85.3 to one. And the chip that they stareted with is pretty good at graphics, bitblit and floaty-point.
--dave
--dave