Using the same logic, if the keyboard driver were to become fuxored, then you'd not be able to type anything anyway, and will eventually ahve to restart. Sure, you could ssh, telnet etc. etc. and get your work done, and/or restart the keyboard driver from within an ssh session, and fix the problem. Oh, wait a second, that's not how your argument goes... Hmmm, maybe you've only used really poor quality OSes?
"The winner in this case is Open Source software."
I disagree. Conditionally.
I'd agree if WinNT (a lot of infrastructural stuff is still running NT) and Whatever MS's latest OS is (are we at 2003XPProPlus?) were included in the benchmark. NT is vaguely POSIX, and I think all of the benchmarks should have been portable onto things Windowsish. And if OSS had then shown to be no worse than MS's output.
It's not fair to make any comparison with any OS utterly unrepresented by the tests.
If MS's graphs had all flatlined below the lowest *BSD/Linux graphs then it would have shown that OSS was significantly poorer than commercial software. You've just _assumed_ where the MS lines would be.
Assumption and benchmarking are incompatible concepts.
Lots of blah blah blah, but I don't notice you supply any figures. He's shown me the 5.1-RELEASE DNF and the 5.1-CURRENT figures, and you've shown me hot air.
And "this study did everything it could to slant against freebsd." is nothing but _paranoia_.
Lock your doors and your windows, remember. Just in case. Don't answer the phone either - it's Linux hackers trying to brainwash you with telephonic telepathy.
Alpha was the first architecture to be able to achieve >1.0 IPC across the board at the SPEC benchmarks.
Alpha was the first architecture to run at 1GHz.
Alphas were repeatedly at the top of the SPEC/FP rantings, occasionally letting HPPAs climb ahead. However, Alphas almost always clawed back the #1 spot fairly soon afterwards. (Grab a dump of the Spec tables from ~1997, say, from www.spec.org to see how dominant it was (it was Spec CPU95 in those days.)
They were a little weaker with integer ops, but they were designed as for numerical applications, and FP was all that mattered. By default IEEE-conformance was disabled, as they could eke out more performance through not handling things like exact exceptions, denormalised numbers, and operations on NaN.
However, bits of Alpha (tech and engineers) were being sold off _years_ ago. The reasons that Intel and AMD are at the top of the SPEC tables now is because they took so much from Alpha/Compaq/HP. (And OK there may have been the occasional patent infringement before even then, *cough* Intel *cough*).
Egon Ronay has released a new gastronomic periodic table, with the elements arranged by taste. Along the bottom are the tasteless ones such as the inert metals, and higher up are more astringent ones such as iodine and sulphur. Egon has yet to place fluorine and potassium, but is said to be recovering well, according to his wife who visits him daily.
I've had a 64-bit desktop machine for 5 years. I bought it when the technology was a bit old, and thus cheap. Cheaper than a wizzo-graphics/sound gaming PC, anyway. Is "Samsung" a well-enough-known company for you? You don't get much more "consumer electronics" than Samsung.
I believe the benchmarks on www.spec.org more than these, but Apple have got all snotty and don't post any anymore. YAW.
You're maybe thinking of Ron Rivest, of Rivest Shamir Aldleman (RSA) fame. The magic words to google for are "40 quadrillion years", if I remember correctly. (Sometimes numbers just stick in your head when they're so embarassingly wrong!)
Possibly fantastic for the sieving stages for factoring the RSA challenges. No number of nodes is too many for a task like that. (where each has comparatively low processing power)
Thy mention 64 processing elements, so 25GFlops is only 390MHz. 390MHz is a pretty slow clock speed. Maybe there's _also_ a DMA controller moving data around, and every number that gets moved into a PE or moved out of a PE counts as a FlOp too.
It doesn't sound unreasonable. I remember in the 90s working with TI c80MVP chips which had an integer DSP, 4 multiple-issue (OK, LIW) FP DSPs, and one DMA controller. For only a 66MHz piece it was cranking out nearly 1GOps.
""" T-Minus 10 second and counting til someone starts bashing Gentoo and recommending Debian. """
??? I'm a died-in-the-wool Debian user, and yet I've recommended Gentoo to many more people recently than I've recommended Debian.
Do you really see that much Gentoo bashing? Are you sure they're real debian users and not just MS plants? Either way, they're idiots, Gentoo's perfect for this situation.
Of course every copy is as old as the date when the copy was taken. Your argument is a straw man. However, there was a freely available up-to-date copy available. If you're not prepared to look at the source of the information for the latest information then you have a problem.
I am referring to the side bar, and I consider Opera to be a real web browser. The box is not labeled "local links" in my Opera, and the source of the page does not contain the string "local links", as far as I can tell. By the "standards" they promoted 5 years ago, they must fail themselves.
Hmmm, I can see HS saying teh exact opposite. Maybe it was pre-HS then, and that could be why HS was dragged on board in order to repair their reputation.
Probably c.l.c.m. Maybe it wasn't HS, but it was definitely an official MS spokesperson. Probably over a year ago. I'm googling now, but all the terms yield a million hits...
""" I swore off C++ almost a year ago (wish I had done it sooner), and in retrospect, getting "the maximum" from C++ felt like getting blood from a turnip. """
Gawd bleshya for your honesty.
Fess-up time: I was a C++ _lecturer_ about a decade ago, and would with a clear conscience recommend it for almost everything(*), but I more often than not (i.e. 90% of the time) recommend _other_ languages than C++ nowadays.
I've even heard occasional derogatory remarks from those on the national standards committees who ten years ago thought the road to programming heaven was paved with C++.
YAW.
(* I even wrote a real-time microkernel in C++ in order to prove to my students that C++ didn't have to be slow. I could task-switch 6000-times a second on a 486/33, for example (but not do anything in the tasks obviously!))
(that is the regulars on comp.lang.c++(.moderated))
Personally I prefer Sutter, he actually is a 'language lawyer', Myers is more of an amateur (who writes books that needed to be written, so all respect for him for doing that).
I'd have said that as soon as things like property lists hit LISP, then it was object-based. There's a fine line between object-based and object oriented. (Some even say that early (<=cfront2) C++ is more properly described as object-based.) I have no idea when property lists hit LISP though? They've been in every dialect I've ever encountered, but I've never read any early McCarthy of even MACLISP docs.
Are you now, or have you ever been, a LISP programmer? -- McCarthy
The W3 CSS page, http://www.w3.org/Style/CSS/ , contains links that don't look like links. For reasons such as that, the W3 ceased being a credible source of information years ago. They're just as bad as the other schlock-merchants. So what if you make something a standard, that _does't_ make it fundamentally a better thing to do.
Slashdot Mirror
Using the same logic, if the keyboard driver were to become fuxored, then you'd not be able to type anything anyway, and will eventually ahve to restart. Sure, you could ssh, telnet etc. etc. and get your work done, and/or restart the keyboard driver from within an ssh session, and fix the problem. Oh, wait a second, that's not how your argument goes... Hmmm, maybe you've only used really poor quality OSes?
YAW.
"The winner in this case is Open Source software."
I disagree. Conditionally.
I'd agree if WinNT (a lot of infrastructural stuff is still running NT) and Whatever MS's latest OS is (are we at 2003XPProPlus?) were included in the benchmark. NT is vaguely POSIX, and I think all of the benchmarks should have been portable onto things Windowsish.
And if OSS had then shown to be no worse than MS's output.
It's not fair to make any comparison with any OS utterly unrepresented by the tests.
If MS's graphs had all flatlined below the lowest *BSD/Linux graphs then it would have shown that OSS was significantly poorer than commercial software. You've just _assumed_ where the MS lines would be.
Assumption and benchmarking are incompatible concepts.
YAW.
Lots of blah blah blah, but I don't notice you supply any figures.
He's shown me the 5.1-RELEASE DNF and the 5.1-CURRENT figures, and you've shown me hot air.
And
"this study did everything it could to slant against freebsd."
is nothing but _paranoia_.
Lock your doors and your windows, remember. Just in case. Don't answer the phone either - it's Linux hackers trying to brainwash you with telephonic telepathy.
YAW.
"""
Microsoft learned a lot about making a 64 bit OS from it's Alpha experience
"""
Alpha NT was 32-bit. At least the version I have is. (Not that I've booted into it for years, strictly an alphalinux user.)
YAW.
Yes, yes, yes, yes and no,
Alpha was the first architecture to be able to achieve >1.0 IPC across the board at the SPEC benchmarks.
Alpha was the first architecture to run at 1GHz.
Alphas were repeatedly at the top of the SPEC/FP rantings, occasionally letting HPPAs climb ahead. However, Alphas almost
always clawed back the #1 spot fairly soon afterwards.
(Grab a dump of the Spec tables from ~1997, say, from www.spec.org to see how dominant it was (it was Spec CPU95 in those days.)
They were a little weaker with integer ops, but they were designed as for numerical applications, and FP was all that mattered. By default IEEE-conformance was disabled, as they could eke out more performance through not handling things like exact exceptions, denormalised numbers, and operations on NaN.
However, bits of Alpha (tech and engineers) were being sold off _years_ ago. The reasons that Intel and AMD are at the top of the SPEC tables now is because they took so much from Alpha/Compaq/HP.
(And OK there may have been the occasional patent infringement before even then, *cough* Intel *cough*).
YAW
Egon Ronay has released a new gastronomic periodic table, with the elements arranged by taste. Along the bottom are the tasteless ones such as the inert metals, and higher up are more astringent ones such as iodine and sulphur. Egon has yet to place fluorine and potassium, but is said to be recovering well, according to his wife who visits him daily.
Hahahah, you mentioned 'stable' in the context of BIND.
... ...
Man you have a sense of humour!
Here's an exerpt from Dan Bernstein's (of djbdns fame) view on BIND's stability:
bug bug bug bug bug bug bug bug bug bug
bug bug bug bug bug bug bug bug bug bug
bug bug bug bug bug bug bug bug bug bug
bug bug bug bug bug bug bug bug bug bug
http://cr.yp.to/djbdns/blurb/unbind.html
YAW
6.12
7.any crash 20 times a day; and the tabbed browsing isn't 100% disableable -- it's mixed-paradigm nonsense that's a pain to use.
I've had a 64-bit desktop machine for 5 years.
I bought it when the technology was a bit old, and thus cheap.
Cheaper than a wizzo-graphics/sound gaming PC, anyway.
Is "Samsung" a well-enough-known company for you? You don't get
much more "consumer electronics" than Samsung.
I believe the benchmarks on www.spec.org more than these, but Apple have got all snotty and don't post any anymore.
YAW.
Altivec's SIMD, which is not vector.
YAW.
You're maybe thinking of Ron Rivest, of Rivest Shamir Aldleman (RSA) fame. The magic words to google for are "40 quadrillion years", if I remember correctly. (Sometimes numbers just stick in your head when they're so embarassingly wrong!)
YAW.
Possibly fantastic for the sieving stages for factoring the RSA challenges. No number of nodes is too many for a task like that.
(where each has comparatively low processing power)
YAW.
Thy mention 64 processing elements, so 25GFlops is only 390MHz.
390MHz is a pretty slow clock speed. Maybe there's _also_ a DMA
controller moving data around, and every number that gets moved
into a PE or moved out of a PE counts as a FlOp too.
It doesn't sound unreasonable. I remember in the 90s working with TI c80MVP chips which had an integer DSP, 4 multiple-issue (OK, LIW) FP DSPs, and one DMA controller. For only a 66MHz piece it was cranking out nearly 1GOps.
YAW.
"""
T-Minus 10 second and counting til someone starts bashing Gentoo and recommending Debian.
"""
??? I'm a died-in-the-wool Debian user, and yet I've recommended Gentoo to many more people recently than I've recommended Debian.
Do you really see that much Gentoo bashing? Are you sure they're real debian users and not just MS plants? Either way, they're idiots, Gentoo's perfect for this situation.
Foot to the metal!
YAW.
So it appears CISCO have never heard of Due Diligence?
I have a lovely bridge I can sell them.
YAW.
Kiitos.
Of course every copy is as old as the date when the copy was taken. Your argument is a straw man.
However, there was a freely available up-to-date copy available.
If you're not prepared to look at the source of the information for the latest information then you have a problem.
YAW.
I am referring to the side bar, and I consider Opera to be a real web browser. The box is not labeled "local links" in my Opera, and the source of the page does not contain the string "local links", as far as I can tell. By the "standards" they promoted 5 years ago, they must fail themselves.
YAW.
Hmmm, I can see HS saying teh exact opposite.
Maybe it was pre-HS then, and that could be why HS was dragged on board in order to repair their reputation.
Still googling...
Probably c.l.c.m.
Maybe it wasn't HS, but it was definitely an official MS spokesperson. Probably over a year ago.
I'm googling now, but all the terms yield a million hits...
Didn't HS himself say that making MSVC++ standards compliant extremely low on the MS priority list?
YAW.
"""
I swore off C++ almost a year ago (wish I had done it sooner), and in retrospect, getting "the maximum" from C++ felt like getting blood from a turnip.
"""
Gawd bleshya for your honesty.
Fess-up time:
I was a C++ _lecturer_ about a decade ago, and would with a clear conscience recommend it for almost everything(*), but I more often than not (i.e. 90% of the time) recommend _other_ languages than C++ nowadays.
I've even heard occasional derogatory remarks from those on the national standards committees who ten years ago thought the road to programming heaven was paved with C++.
YAW.
(* I even wrote a real-time microkernel in C++ in order to prove to my students that C++ didn't have to be slow. I could task-switch 6000-times a second on a 486/33, for example (but not do anything in the tasks obviously!))
Oi!
_We_ wrote a fair proportion of Sutter's EC++!
(that is the regulars on comp.lang.c++(.moderated))
Personally I prefer Sutter, he actually is a 'language lawyer', Myers is more of an amateur (who writes books that needed to be written, so all respect for him for doing that).
YAW.
I'd have said that as soon as things like property lists hit LISP, then it was object-based. There's a fine line between object-based and object oriented. (Some even say that early (<=cfront2) C++ is more properly described as object-based.)
I have no idea when property lists hit LISP though? They've been in every dialect I've ever encountered, but I've never read any early McCarthy of even MACLISP docs.
Are you now, or have you ever been, a LISP programmer?
-- McCarthy
Or not.
YAW
The W3 CSS page, http://www.w3.org/Style/CSS/ , contains links that don't look like links. For reasons such as that, the W3 ceased being a credible source of information years ago. They're just as bad as the other schlock-merchants. So what if you make something a standard, that _does't_ make it fundamentally a better thing to do.
YAW.