[OT] really?

[Daniel Gimpelevich]

On Tue, 14 Jun 2005 10:54:47 -0400, Eric Dunbar wrote:

> I'll only touch on a few points (I disagree with you on a number of
> fronts but it's not worth re-hashing tired old arguments that have
> already been articulated (I'm no politician)).

I'm no politician either; I tend to be rather blunt at times. I'd rather
not rehash that much myself, so if there are any new points not yet
mentioned...

> Umm. That's why there are APIs -- to avoid direct h/w access. OpenGL

Yes, but there are OS-specific APIs and there are multi-OS APIs.

> and other such graphics programming languages provide a platform
> agnostic means of accessing graphics cards. Some programmers *do* go
> straight to the source, and, having i86 for both their Windows and Mac
> OS code will make that job that much easier, if they do. Plus, drivers

That's not what I mean. OS-specific APIs are used far more than multi-OS
APIs and direct hardware access put together. Code using multi-OS APIs may
be built for any supported OS, and code using direct hardware access may
be ported to any OS using that hardware (but not as easily), but code
using OS-specific APIs is locked to a specific OS and the hardware
supported by that OS.

> for Windows x86 tend to be of much higher quality than for Mac PPC.

I challenge you to name even one Apple-supplied driver that is of worse
quality than its Windows counterpart.

> The reason they don't exist for Linux (with a few oddball exceptions)
> is that Linux just doesn't constitute a large enough a paying base and
> there are too many variants to support (I'd hate to take the support
> calls of Linux computer "users" (i.e. don't know the first thing about
> sudo)).

If you're talking about hardware drivers, that can only apply to
vendor-supplied Linux drivers, which as a rule with some exceptions, are
complete crap. Again, not using Linux does not make someone any more a
paying customer than a Linux user, because use of the drivers tends to be
free as in beer on any OS. Demand for such vendor-supplied drivers is very
low not just because of Linux's market share, but for two more important
reasons: First, few such drivers are open-source. Second,
community-supplied open-source drivers are of much higher quality than
vendor-supplied open-source drivers for Linux.

> They don't have to but efficiency demands that they do!

Efficiency does not demand that they do under Mac OS 10.3 and later, and
OSX-x86 demands that they don't. It's efficiency under Linux that demands
that they do. If you want to see further discussion of such efficiency
under Linux, look at any Gentoo-related discussion.

> Can you expand on "not affected the availability of non-free (as in
> speech) codecs for Linux much."? Do you mean that these codecs *are*
> available for Linux (there are MANY more codecs available on x86 Linux
> than on PPC Linux)? Officially, unofficially? What?

The codecs available on x86 Linux but not PPC Linux are so precisely
because they are non-free as in speech. That proves they exist, and
therefore their availability has not been prevented. The vast majority of
these codecs, by the way, are Windows codecs with Linux wrappers.

> I really do think you underestimate the psychological importance of
> "Intel Inside" for computer users and coders. Mac may be widely known
> *outside* of the Mac-computing community but PowerPC/PPC IS NOT!!!
> Even for those who know PPC is a niche market, and not necessarily
> associated with Mac.

PowerPC is not widely known among users outside of Mac, but the same is
true of Intel Inside outside of Windows. Believe it or not, there are
still computer users that have never heard of Linux, FreeBSD, or anything
else that runs on Intel.

Psychological importance of Intel == Psychological importance of Windows

> Even if the benefits from switching to i86 are no more than
> psychological and speed (and I think there are more that come with
> low-level x86 code compatibility) they're worth it.

Whether or not it's worth it is not something that can be measured, so
everyone has to make their own determination of that.

> Yep, that's a problem that's already been pointed out. The trade-offs
> are minor though since Apple is quite good about keeping system
> "closed" by default and patched (it's a hell of a lot easier to patch
> an OS X system than YDL or even Ubuntu (though, Ubuntu is now
> apparently testing an auto-update daemon)).

My comment was only intended to pertain to the use of Linux on Intel Macs,
but the prospect of CPU-based attacks on OSX is worth exploring.

Keeping the system "closed" by default isn't exclusively due to Apple, and
Apple never makes OSX as "closed" as it could be. Patches address
vulnerabilities in the OS, not in the CPU.

Ubuntu has dpkg and apt-get. What's easier than that? Apple's Software
Update? (And don't say "Yes" until you consult some MUG FAQ about "Should
I ever use Software Update?")

> You are correct. I did say "Debian as the Debian distro of choice".
> And, I meant what I said, no more, no less!
> 
> Perhaps you're not familiar with the ecology of the Debian community?
> There are a tonne of Debian-based distros based on the Debian
> repositories. Ubuntu is itself nearly 100% Debian (unstable) except
> that is has a different philosophy guiding its development and

It is because I know this that I said what I said.

> operation, and, Ubuntu appears poised to supplant Debian as the most

Most widely used among whom? Certainly not desktop users. Use of Debian
itself is quite rare among Windows converts.

> widely used Debian-based distro. This does not say that Debian will
> die. It will remain the gold standard upon which Ubuntu is based and
> Ubuntu's changes will feed back into Debian's development. I also

That is true of any Debian-based distro, but Ubuntu's changes have an
extra layer of management before they can get into Debian.

> expect to see Ubuntu chew into the relative "market share" of other
> distros, including MEPIS. DistroWatch shows that there's a phenomenal
> level of interest in Ubuntu, but, that's not a 100% accurate measure
> since those numbers can be inflated through a concerted effort on the
> part of fan(atics).
> 
> Time will tell which ones climb to the top. Ubuntu will definitely
> become one of the "big boys".

It already has. Within the sub-category that is Debian-based distros for
desktop users, Ubuntu and MEPIS are the big boys that should be measured
against. There may be bigger boys that are either not Debian-based or not
suited to desktop users, but that's a whole other discussion.

> As for YDL vs. FC... don't know, don't care. As long as my server runs
> Apache, Samba, AppleTalk, Webmin, Webalizer, Gallery and GNOME via
> XDCMP (for experimentation) I'm happy (plus, I may start experimenting
> with LTSP or FreeNX).

It is possible to do practically all of that under OSX, too. But ease and
straightforwardness is what shapes preferences, and recommendations of
distros to try are based on expected preferences.

> They are not always the first adopters but are definitely early
> adopters. Apple takes what is normally a painful process and makes it
> easy. That's Apple's magic formula. Others in turn copy Apple once

Preaching to the choir.

> Apple has shown the model's value. In 1983/1984 they (including Lisa)
> were "first adopters" with the Mac GUI. In 1987 they were early
> adopters with high quality colour graphics (unseen except in the tiny

PCs had 24-bit color long before the Mac did, relatively speaking.

> niche of Commodore's world). The elimination of floppy drives in 1998.
> Etc.

Is that a typo? Steve Jobs permanently eliminated floppy drives is 1988,
not 1998. Or is that era now forgotten?

> The iPod is a disposable consumer product that does one task, and one
> task only! It's not like a computer.

"Disposable consumer product" is what describes a computer to people that
only know Windows, and Apple sees a bit of wisdom in that.

> Tiger is "proven" technology. 10.4 is little different from 10.3.
> Moving to 10.0 WAS a major shift, however!

Only the technology that is inherited from 10.3 is proven. The new stuff
is new.

> Didn't know which ones Apple had dropped support for. I'm guessing
> that G3s may be next on the chopping block (perhaps Apple could use
> graphics card size as its next discriminator?).

All of those were choices Apple had for the Tiger release, but their
decision was not to do those things, proving that speculation as to what's
next on the chopping block is useless.

> That still doesn't change the fact that Apple is now making more off
> selling OSes than it ever did (in relative income).

That is because of their strategy, not in spite of it. Regardless of how
much more or less they make from selling OSX in the future, it is
currently still not their main business.