Linux-Advocacy Digest #19, Volume #34 Sat, 28 Apr 01 15:13:02 EDT
Contents:
Re: there's always a bigger fool (Johan Kullstam)
Re: there's always a bigger fool (Zippy)
Re: Does Linux support "Burn-Proof" CDRW's ([EMAIL PROTECTED])
Re: Article: Want Media Player 8? Buy Windows XP ([EMAIL PROTECTED])
Re: Blame it all on Microsoft (Bengt Larsson)
Endeavour shuttle and windows (Marcello Barboni)
Re: there's always a bigger fool (Johan Kullstam)
Re: Article: Want Media Player 8? Buy Windows XP (Donn Miller)
Re: Endeavour shuttle and windows (Donn Miller)
Re: Blame it all on Microsoft ([EMAIL PROTECTED])
Re: there's always a bigger fool (Zippy)
----------------------------------------------------------------------------
Crossposted-To: alt.destroy.microsoft,comp.os.ms-windows.advocacy,soc.singles
Subject: Re: there's always a bigger fool
From: Johan Kullstam <[EMAIL PROTECTED]>
Date: Sat, 28 Apr 2001 18:31:25 GMT
T. Max Devlin <[EMAIL PROTECTED]> writes:
> Said Johan Kullstam in alt.destroy.microsoft on Sat, 28 Apr 2001
> >[EMAIL PROTECTED] (Zippy) writes:
> >
> >> >on x86, linux does use the bios to boot. pci and io-apic are set up
> >> >by bios. once bootstrapping is complete, you are right, bios is not
> >> >used again.
> >>
> >> that's a limitation of "pc architecture", rather than a feature. of course,
> >> things like the crusoe architecture seriously blur the lines. in most
> >> cases, the way an architectural platform sets up the hardware is what gives
> >> that hardware an advantage when the OS loads. this is why Sun computers
> >> boot Solaris and Linux so sweetly compared to a PC, same thing for a Mac
> >> (from what i've heard - i've never actually SEEN linux running on a
> >> Mac).
> >
> >every machine has a boot ROM. the problem is that backward
> >compatibility trumps any design improvement. the IBM PC was designed
> >to compete with the Commodore-64. IBM didn't even think they'd sell
> >many with floppy drives. they were wrong, of course. still, we
> >suffer from the short-sightedness of that time.
>
> Why?
because
1) microsoft is an incompetent porter of software. look how long it
took them from the introduction of the 386 to get 32 bit
OS/applications. whether this is a planned incompetence to prevent
competition or just laziness, i don't know. i gather that microsoft's
greatest fear is a technology leap since that might let someone else
get their foot in the door.
2) backwards compatibilty is king. it trumps all else. microsoft
plays the proprietary data format game better than anybody else. part
of the lock-in to microsoft seems to make for lock-in to ia16 and then
ia32 PC architecture. "not having any applications" is a code phrase
for "doesn't run ms-office".
the bios lives on (with minor patches for, e.g., larger harddrives)
because it will boot the various flavors of windows. launching
windows being all anybody (who counts) cares about, that's enough.
> >> think about it - i found out recently that PCI devices don't even require
> >> an IRQ to function! it's Windows DOS/9x that freaks out if a device other
> >> than a video card doesn't have an IRQ, so we have to have two sets of PCI
> >> cards in the world - one for x86 architecture, and another for Alphas,
> >> SPARCs, Macs and everything else.
> >
> >i dunno.
>
> I do. Want me to explain it?
*requiring* an IRQ is stupid. however, *having* an interrupt line is
just plain smart for many devices.
as everyone know, windows is severely brain damaged in all kinds of
ways. however, microsoft has the luxury of having hardward cater to
and work around their flaws. this doesn't help the ia32 platform
much.
> >ia64 has been coming real soon now for like 2-3 years. from what i
> >gather from the net, ia64 has been totally blown away by recent clock
> >speed advances in ia32. you can always re-layout ia64 for the smaller
> >feature-size of todays process. however, my (totally wild-ass,
> >unsubstatiated) guess is that intel will drop ia64 in its current
> >form, take the lessons learned and make an ia64-2.
>
> That sounds more like an inevitable occurrence, then a wild-ass guess.
> :-D
thanks.
> --
> T. Max Devlin
> *** The best way to convince another is
> to state your case moderately and
> accurately. - Benjamin Franklin ***
--
J o h a n K u l l s t a m
[[EMAIL PROTECTED]]
Don't Fear the Penguin!
------------------------------
Crossposted-To: alt.destroy.microsoft,comp.os.ms-windows.advocacy,soc.singles
Subject: Re: there's always a bigger fool
From: [EMAIL PROTECTED] (Zippy)
Date: Sat, 28 Apr 2001 18:36:02 GMT
>the point of the IRQ is for the PCI card device to grab the attention
>of the CPU. an interrupt is activated when the card wants to tell the
>CPU something. this eliminates (wasteful) polling of the card by the
>CPU.
right, the card still sends the "interrupt request" regardless. but with
PCI it doesn't need to use the antiquated PC "IRQ" of 0-15. it uses the PCI
controller, which identifies it as "bus x: device x". actually maybe
windows DOES still use the old IRQs. that would be a hell of a waste, but i
wouldn't put it past them.
>
>in the SCSI card case, the SCSI card BIOS provides native machine
>instructions for the host. these machine instructions provide
>functions to allow the motherboard boot ROMs to access the SCSI
>drives.
scsi card is a really poor choice because it's a boot device. same thing
for network cards. hmmm... what else is left? sound card? modem? the
modem's a good example. it won't work in windows without an irq, because
COM ports require an IRQ. but a PCI modem SHOULD work anyway, even without
the IRQ.
actually the scsi card isn't such a bad example if you look at non-bootable
scsi cards. adaptec made one such card that worked in everything, mac and
pc alike, and i'd guess in a sun or alpha as well.
>i still don't think it's an IRQ question. a SCSI card of all things
>most certainly should have an IRQ. when the drive gets done seeking
>and reading what it needs to, then by all means busmaster it directly
>over to main memory and then wake the CPU by an interrupt to tell it
>that the data has arrived.
sure, all cards have an "irq" in the strictest sense of the word. but it
doesn't have to be 0-15, except for the fact that the PC motherboard BIOS
demands it to boot the machine, for backward compatibility reasons. many
cards can do both.
>
>alpha, sparc, hppa, powerpc are all much better arches than intel
>ia32. however, intel has a vastly greater volume of sales over which
>to ammortize non-recurring engineering costs. this means that despite
>a crappy arch, intel and amd can offer competitive CPUs
>price/performance.
actually, intel's sales of processors probably isn't as huge as you think.
a substantial portion of their income is from flash memory and other chips
and chipsets. sun still holds the market for internet servers, and ibm
still holds a commanding lead in the mainframe/minicomputer market (the
AS400 uses PPCs now, BTW). intel is going to be hard pressed to make any
gains in these areas any time soon, particularly if IA64 is the best they
can do.
people forget that they already tried this once before with the i960, a
dismal failure if there ever was one. they just don't know how to make RISC
processors.
>
>backward compatibility (ability to run windows and its applications)
>being nearly essential to most users, the other inherently superior
>arches have no chance. for many, it doesn't matter that alpha is 3x
>faster than intel, if alpha costs 10x more and won't run office 97.
sure. but it will run ANYTHING JAVA. and that's what's going to be really
important in the future. office 97 is irrelevant. you want my copy?
i guess that's my point. transmeta and linux have already proven that
architecture doesn't have to define the OS any longer. it's only a matter
of time before the rest of the industry catches on. you can create a VLIW
processor that can run any instruction set, or you can create a completely
new operating system for an existing architecture.
freedom!
>
>the ia64 may be a better architecture than ia32; i don't know,
>although i'd be hard pressed to design anything worse than ia32. i
>think it is unknown if intel can sell them because afaict they will
>not be directly compatible with ia32. any sufficiently powerful CPU
>could emulate the ia32 which could give competition a chance.
>
there already is such a thing. it's called the athlon. unfortunately, most
athlon boards use VIA chipsets. VIA makes crappy, extremely unreliable
chipsets.
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Does Linux support "Burn-Proof" CDRW's
Date: Sat, 28 Apr 2001 18:38:30 GMT
Thank you Lee, I wasn't sure if it was supported or not.
It makes a HUGE difference, even with the higher speed burners.
Flatfish
On Sun, 29 Apr 2001 00:05:51 +0800, Lee Wei Shun
<[EMAIL PROTECTED]> wrote:
>[EMAIL PROTECTED] wrote:
>>
><snip>
>>
>> How about Linux burner software?
>>
>> Are you guys behind the 8-ball yet again?
>>
>> Flatfish
>
>Very much up to date, thanks. Please see:
>
>http://www.fokus.gmd.de/research/cc/glone/employees/joerg.schilling/private/cdrecord.html
>
>Regards,
>Wei Shun
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: alt.destroy.microsoft
Subject: Re: Article: Want Media Player 8? Buy Windows XP
Date: Sat, 28 Apr 2001 18:43:19 GMT
Fact.
Vegas Audio/Video
Sonar
Cakewalk
Cubase
Logic Audio
SoundForge
Acid
And on and on....
Flatfish
On Sat, 28 Apr 2001 18:08:51 GMT, T. Max Devlin
<[EMAIL PROTECTED]> wrote:
>Said Todd in alt.destroy.microsoft on Fri, 27 Apr 2001 13:47:54 +0800;
> [...]
>>Well, see, with Windows XP, you have a choice. You can use the free Media
>>Player, or you can use one of a bazillion different other media players out
>>there... some of them quite good.
>>
>>You can choose whether or not to use XP or the media player. Just like you
>>chose to use Linux. So where is the problem?
>>
>>Of course, all of the good media applications are for Windows, but hey, if
>>you are fanatic about Linux, using substandard apps. is the norm.
>
>Guffaw.
------------------------------
From: [EMAIL PROTECTED] (Bengt Larsson)
Crossposted-To: comp.theory,comp.arch,comp.object
Subject: Re: Blame it all on Microsoft
Date: Sat, 28 Apr 2001 17:52:29 +0000
Reply-To: [EMAIL PROTECTED]
In comp.arch, [EMAIL PROTECTED] wrote:
> Actually, I do know some answers to making routing scale,
>but they have the disadvantage of being theoretical. See any comments
>made by Noel Chiappa on any of several mailing lists during the
>time IPv6 was being botched.
How does separating the address into end-point identifier and locator
solve the routing problem? I read the Big-internet list at the time;
it seemed to me nobody understood how that solved the routing problem.
------------------------------
From: Marcello Barboni <[EMAIL PROTECTED]>
Subject: Endeavour shuttle and windows
Date: Sat, 28 Apr 2001 18:55:07 GMT
It seems that the endavour space shuttle uses windows os on the onboard
computers (I picked this up on sci.space.shuttle), does anyone know if
thi is true? If it is, at least we know why they're having problems....
Marcello Barboni
------------------------------
Crossposted-To: alt.destroy.microsoft,comp.os.ms-windows.advocacy,soc.singles
Subject: Re: there's always a bigger fool
From: Johan Kullstam <[EMAIL PROTECTED]>
Date: Sat, 28 Apr 2001 18:55:27 GMT
[EMAIL PROTECTED] (Zippy) writes:
> >the point of the IRQ is for the PCI card device to grab the attention
> >of the CPU. an interrupt is activated when the card wants to tell the
> >CPU something. this eliminates (wasteful) polling of the card by the
> >CPU.
>
> right, the card still sends the "interrupt request" regardless. but with
> PCI it doesn't need to use the antiquated PC "IRQ" of 0-15. it uses the PCI
> controller, which identifies it as "bus x: device x". actually maybe
> windows DOES still use the old IRQs. that would be a hell of a waste, but i
> wouldn't put it past them.
> >
> >in the SCSI card case, the SCSI card BIOS provides native machine
> >instructions for the host. these machine instructions provide
> >functions to allow the motherboard boot ROMs to access the SCSI
> >drives.
>
> scsi card is a really poor choice because it's a boot device. same thing
> for network cards. hmmm... what else is left? sound card? modem? the
> modem's a good example. it won't work in windows without an irq, because
> COM ports require an IRQ. but a PCI modem SHOULD work anyway, even without
> the IRQ.
>
> actually the scsi card isn't such a bad example if you look at non-bootable
> scsi cards. adaptec made one such card that worked in everything, mac and
> pc alike, and i'd guess in a sun or alpha as well.
>
> >i still don't think it's an IRQ question. a SCSI card of all things
> >most certainly should have an IRQ. when the drive gets done seeking
> >and reading what it needs to, then by all means busmaster it directly
> >over to main memory and then wake the CPU by an interrupt to tell it
> >that the data has arrived.
>
>
> sure, all cards have an "irq" in the strictest sense of the word. but it
> doesn't have to be 0-15, except for the fact that the PC motherboard BIOS
> demands it to boot the machine, for backward compatibility reasons. many
> cards can do both.
>
> >
> >alpha, sparc, hppa, powerpc are all much better arches than intel
> >ia32. however, intel has a vastly greater volume of sales over which
> >to ammortize non-recurring engineering costs. this means that despite
> >a crappy arch, intel and amd can offer competitive CPUs
> >price/performance.
>
> actually, intel's sales of processors probably isn't as huge as you think.
> a substantial portion of their income is from flash memory and other chips
> and chipsets. sun still holds the market for internet servers, and ibm
> still holds a commanding lead in the mainframe/minicomputer market (the
> AS400 uses PPCs now, BTW). intel is going to be hard pressed to make any
> gains in these areas any time soon, particularly if IA64 is the best they
> can do.
>
> people forget that they already tried this once before with the i960, a
> dismal failure if there ever was one. they just don't know how to make RISC
> processors.
intel might suck at designing an architecture, but they do have the
resources to dump on it to make it go fast at a reasonable price.
> >backward compatibility (ability to run windows and its applications)
> >being nearly essential to most users, the other inherently superior
> >arches have no chance. for many, it doesn't matter that alpha is 3x
> >faster than intel, if alpha costs 10x more and won't run office 97.
>
> sure. but it will run ANYTHING JAVA. and that's what's going to be really
> important in the future.
depending on your circles, java isn't terribly relevant. scientific
computing has no use for it. we still use fortran at where i work.
> office 97 is irrelevant. you want my copy?
office 97 is extremely important. why do people at work run windows?
they want to run ms-word, ms-powerpoint, ms-excel &c. no, translation
of these formats to other more or less equivalent programs do not work
seemlessly. nobody actually cares about windows per se. they just
want to run word so they can attach some trivial message better done
in plain ascii in an email.
i don't have a copy of office 97, nor do i want one. i am a linux
user. but i still understand what people find important.
> i guess that's my point. transmeta and linux have already proven that
> architecture doesn't have to define the OS any longer. it's only a matter
> of time before the rest of the industry catches on. you can create a VLIW
> processor that can run any instruction set, or you can create a completely
> new operating system for an existing architecture.
>
> freedom!
let's hope so. i remain skeptical. this is simply because the
computer industry has historically lurched from one monopoly to the
next. it's really a matter of supply and demand.
1) on the production side, software has a large non-recurring initial
cost and virtually zero marginal costs.
2) on the demand side, the value of a piece of software increases the
more people use it since you can share files, knowledge of how to
operate it &c.
thus the supply and demand curve is inverted compared to the classic
textbook model. we have an unstable equalibrium in the middle and two
stable points at near zero and full market penetration.
microsoft, as bad as they are at actual software, understand this
point well.
> >the ia64 may be a better architecture than ia32; i don't know,
> >although i'd be hard pressed to design anything worse than ia32. i
> >think it is unknown if intel can sell them because afaict they will
> >not be directly compatible with ia32. any sufficiently powerful CPU
> >could emulate the ia32 which could give competition a chance.
> >
>
> there already is such a thing. it's called the athlon.
imho the athlon is not an emulation of ia32. it's an alternate
implementation of it. by emulation i mean, e.g., a powerpc
interpreting and running ia32 code.
> unfortunately, most athlon boards use VIA chipsets. VIA makes
> crappy, extremely unreliable chipsets.
nod. i would buy an athlon now if i felt the chipset was reliable in
linux. amd may be forced to make their own chipsets in the end.
intel seems to have lost the edge here too. the BX was good, but
i820 was a disaster, i810 crippled, i840 requires that crappy rambus.
i815 seems ok. still, it seems like a whole lot of motherboards --
both intel and amd cpu -- use VIA chipsets.
--
J o h a n K u l l s t a m
[[EMAIL PROTECTED]]
Don't Fear the Penguin!
------------------------------
Date: Sat, 28 Apr 2001 14:56:34 -0400
From: Donn Miller <[EMAIL PROTECTED]>
Crossposted-To: alt.destroy.microsoft
Subject: Re: Article: Want Media Player 8? Buy Windows XP
Chad Everett wrote:
> Except, of course, if there are radio stations or other broadcast
> entities that only push out Windows Media Player format. Microsoft
> is really trying to corner this market. They wanted to do this
> with Java and other Internet protocols too, but they seem to be
> having more success with Windows Media Player. I don't know why
> there isn't more outrage about this.
It's because of who those entities are being managed by. They are
mostly marketing people who have one or two "introduction to Windows"
courses, so they are just falling back to whatever limited computer
knowledge they have. They never look at the big picture at what impact
their decisions may ultimately have. Plus, there's the advertising
aspect of the whole thing. MS probably markets things like
content-controlled Media Player pretty aggressively. It's how MS got to
where they are today - aggressive marketing and strategically
controlling the software market.
====== Posted via Newsfeeds.Com, Uncensored Usenet News ======
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
======= Over 80,000 Newsgroups = 16 Different Servers! ======
------------------------------
Date: Sat, 28 Apr 2001 15:00:02 -0400
From: Donn Miller <[EMAIL PROTECTED]>
Subject: Re: Endeavour shuttle and windows
Marcello Barboni wrote:
>
> It seems that the endavour space shuttle uses windows os on the onboard
> computers (I picked this up on sci.space.shuttle), does anyone know if
> thi is true? If it is, at least we know why they're having problems....
Really? Hmmmm... Hey, wouldn't it be great if MS would come out with
Windows 2000 embedded? Damn, I'd bet they'd push VxWorks, QNX, and
RTLinux right out of the market. Of course, a huge memory upgrade would
be a small price to pay for this awesome multi-platform OS (NT) that
killed Unix back in 1992.
====== Posted via Newsfeeds.Com, Uncensored Usenet News ======
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
======= Over 80,000 Newsgroups = 16 Different Servers! ======
------------------------------
Crossposted-To: comp.theory,comp.arch,comp.object
Subject: Re: Blame it all on Microsoft
From: [EMAIL PROTECTED]
Date: Sat, 28 Apr 2001 19:08:06 GMT
In article <[EMAIL PROTECTED]>,
Bengt Larsson <[EMAIL PROTECTED]> wrote:
>In comp.arch, [EMAIL PROTECTED] wrote:
>
>> Actually, I do know some answers to making routing scale,
>>but they have the disadvantage of being theoretical. See any comments
>>made by Noel Chiappa on any of several mailing lists during the
>>time IPv6 was being botched.
>
>How does separating the address into end-point identifier and locator
>solve the routing problem? I read the Big-internet list at the time;
>it seemed to me nobody understood how that solved the routing problem.
Not to scratch up a long-dead debate here, but the basic
idea is to allow the 'locator,' the thing that appears in every packet
that tells the routers where to send it, can be structured optimally
for router hardware to use, and can be computed to be the right thing
based on where the two endpoints are and the current network state,
when you establish a channel of communication.
You're quite right, though, not many people Got It, which
is why IPv6 is the way it is.
The permanent identifier for the endpoint, for which we
currently use a 32 bit identifier (IP Address) is more or less fixed.
IPv6 fails to solve the problem.
I think of it as the difference between:
'I live at 411 Mississippi'
and
'the best way to get to my house from the airport now
is up 169 and over on 94 and then up university'
The former is fixed, the latter is computed on an as-needed
basis, works MUCH better for someone driving a car, and, being
dynamically computed, can account for issues like traffic loads
at the relevant time, or highway construction.
Now, to address the other issue -- I do not actually KNOW
this will scale better than IP. Theory is not practice.
To bring it back to computer architecture:
The idea is that another layer of indirection is sometimes
the right thing!
------------------------------
Crossposted-To: alt.destroy.microsoft,comp.os.ms-windows.advocacy,soc.singles
Subject: Re: there's always a bigger fool
From: [EMAIL PROTECTED] (Zippy)
Date: Sat, 28 Apr 2001 19:09:56 GMT
i'm not sure linux (at least in its current form) is going to be the
singular answer. different people have different needs, and linux in its
current state is a little bit too sloppy for the average small businessman
or office worker. i'm constantly getting advice and instructions from
people running mandrake or Suse that don't work in RedHat. it's very time
consuming.
nevertheless, linux is one of the first of the Unices that will actually
run well as a desktop. hybrids such as OS X, RedHat (is it stil considered
linux?) and yet unreleased operating systems may fill the void. or maybe
linux itself will start to congeal more. who knows? but a diversity at
least seems assured.
>That was a very nice analysis. Thank you.
>
>I think this goes back to the original issue, of the Linux "platform".
>I don't think it is any change in hardware which will fix these issues
>you talk about; thus, the comments you made on the IA64. Linux alone
>will be sufficient, once the monopoly is dealt with. Linux runs on any
>hardware platform, practically, so the definition of a PC based on the
>technical aspects of hardware will go away. A PC will be whatever
>hardware you use as a "personal computer" and there will not be any of
>the wholly artificial barrier between hardware platforms which used to
>be the rule. Years ago, computer scientists realized that the same
>programs can run on different hardware platforms. It is only the
>existence of the monopoly which has prevented the market from realizing
>the same thing.
yes this is the KEY to the whole question. the portable operating system
was invented a VERRRRRY long time ago - but desktop machines don't share
the same lineage as the rest of the computer world. the personal computer
was born out of ingenuity and surplus calculator chips, not defense
department grants, professional engineers and corporate board rooms.
as much as i resent microsoft, i will grant that they share (or at least
once did share) the "home enthusiast" mentality that is integral to the
consumer desktop.
i don't attribute bill gates's enthusiasm about linux to the evil and
sinister reasons others have assigned to him. bill gates is essentially a
hacker, like you and me. he has as much interest in the hardware as he does
the software (in fact, he's probably getting rather sick of the software
right about now). it is right and just that microsoft (division I) should
make hardware. and it is right and just that microsoft (division II) should
make operating systems. and it is right and just that microsoft (division
III) should make applications.
split microsoft up, and we will have 150 thriving computer companies in
this country (and the top three will still be the former divisions of
microsoft), rather than just one. it's a win/win situation.
>Once the free market is restored, the obvious choice for most people
>will be Linux. Once that happens, all distinctions between different
>hardware platforms become simply alternatives in one big market. The
>OEMs will be able to flush out old BIOS-based technologies like IRQs,
>just as easily and effectively as they would have been able to, years
>ago, in order to make their systems more competitive (cheaper or better)
>if compatibility with the monopoly was not the controlling issue.
>Without that controlling monopoly, any OEM can just build anything they
>want and call it a PC, and see if it sells, because Linux can run on
>just about anything. Sun and Apple and Dell will all just be different
>computer manufacturers, and they will probably all make PCs, but there's
>no need for their PCs to be similar in hardware (other than the market
>efficiencies allowed by interchangeable components, of course), as long
>as they run Linux.
or OS X, or 2K/NT, or Oracle or RPG or SunOS, or SAP or SQL or Novell or...
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list by posting to comp.os.linux.advocacy.
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Advocacy Digest
******************************
