Re: [Sheflug] RAID and kernel upgrading.

["Stephen J. Turnbull" <nospam [at] sk.tsukuba.ac.jp>]

>>>>> "Will" == Will Newton <will [at] misconception.org.uk> writes:

    Will> On Thu, 8 Jun 2000, Stephen J. Turnbull wrote:
    >> >From what I've heard, 2.4.0-pre is stable only with Red
    >> Hat-supported hardware.  Older Sparcs, many Alphas, Amigas,
    >> Macs, are typically shit-out-of-luck.

    Will> The kernel has exactly bugger all to do with RedHat. Tell it
    Will> to Mulder & Scully. I'm sorry to sound like that, but RH
    Will> supports alpha, sparc and i386 in it's main distro tree.

Not the Sparc ELC I happen to have.  Unless you call oopsing before
starting /sbin/init "support."  If Red Hat pays people to work on
Sparcs, and gives them deadlines and puts priority on the ARCHs that
big commercial users have lots of, what do you think happens to older,
or less common, ARCHs?

Exactly what happens to those ARCHs, apps, etc that MSFT doesn't
support.

    Will> I'm just sick of people demonising RedHat because 90% of
    Will> people started out with RedHat and now they see themselves
    Will> as "above" it.

I started with 386BSD, but no apps, then SLS, OK?  Then went to BOGUS,
didn't have the hardware (disk space, mostly) to support building all
my stuff, so then Slackware.  I have only had one partially successful
Red Hat install of 3 tries, the first two hosed Red Hat installs
resulted in a move to Debian as Slackware didn't cut it.  No way have
I "grown out of" Red Hat.  Red Hat has never grown into my needs,
certainly not enough to justify switching from Debian.

But (this time) my complaints don't have anything to do with whether
Red Hat is a good distro or not; they have to do with Red Hat being
the biggest of the commercial distros, and therefore the ones
redirecting developer resources toward competing with Windows 9x and
NT, 'cause that's where the money is.

Or, to take a more reasonable position.  If I write an email to Alan
Cox about my Sparc[1], and Bob Young walks into his office with a request
to support a VA Linux Sparc/SMP initiative, which do you think is
going to get priority?

    Will> Hmmm, "people you respect" sounds like "bloke in the pub" to
    Will> me.

No, they're core XFree86 and XEmacs developers, both with kernel
contribs.[2]  Little bugfixes; that's not the main use of their time.

Their complaints about the code are mostly that the modularization is
getting broken more frequently in the name of "features" and
"efficiency," and that much of the kernel code they need to understand
to debug their own projects is poorly designed, unreadable and
unmaintainable.  And getting worse.

Allegedly the BSD kernels are going in exactly the opposite
direction.  I'm definitely planning to change my Sparc over to NetBSD
for these reasons.

    Will> The dev. kernels have been very "alpha" recently, but if it
    Will> stabilizes OK I will have no complaints.

Exactly my point.  Except that I noticed both the "very alpha"
behavior and the "pre-*" version tags.  Doesn't that bother you?  It
certainly would if it were a Microsoft product!

    Will> There have been many new additions to the kernel since the
    Will> early days (PCMCIA, ISDN, USB, video, I2C, I2O, etc.).

None of those options are on in my Sparc .config, and therefore have
should have squat to do with kernels that oops before starting
/sbin/init.  One suspects that what is happening is that people are
putting in "exciting features" or "bug fixes" and breaking working
code.  Alan and Linus cannot check all the platforms now "supported"
by Linux.

The problem is that in the old days when free software programmers
were driven by reputation, a code change that breaks an obsolete
platform was considered a threat to that reputation, and the
developers had strong incentive to fix them.  Now that they are partly
driven by money, there is a strong incentive to ignore such bugs
because "it only affects a half-dozen machines" or "we've only had one
bug report," and "we need to release 2.4 soon!"

Sure, since (as any economist will tell you) this money-driven demand
reflects real market needs, paying attention to it is a good thing.
But remember that the free software ecology has produced unique
benefits through a process that we economists understand very poorly
as yet.  We want to be careful that building the "Three Gorges Dam"
distribution doesn't end up with the flow of open source innovation
buried under 50 meters of silt!

    Will> The difference between those paid to work on the kernel and
    Will> others is some do actually do something and not sit on their
    Will> arses and whinge.

Uh-huh.  What really is going on is that the "whingers" are working on
_other_ subsystems (like XFree86 and XEmacs), and would just like to
have a stable platform that boots for their hardware, subsystems, and
apps.

I don't see having money go into Linux as a net bad thing.  But I'm
not going to keep my mouth shut when I see possibly preventable bad
things coming in along with the good.



Footnotes: 
[1]  No, I haven't filed a bug report; others have---Sparc ELC support
has been non-existent for a couple of years.

[2]  _I'm_ just a bloke in the pub on kernel code; that's exactly why
I cited "people I respect."

-- 
University of Tsukuba                Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences       Tel/fax: +81 (298) 53-5091
_________________  _________________  _________________  _________________
What are those straight lines for?  "XEmacs rules."
---------------------------------------------------------------------
Sheffield Linux User's Group - http://www.sheflug.co.uk
To unsubscribe from this list send mail to
- <sheflug-request [at] vuw.ac.nz> - with the word 
 "unsubscribe" in the body of the message. 

  GNU the choice of a complete generation.