[Eug-lug] Free O'Reilly book

[T. Joseph CARTER]

On Fri, Jan 28, 2005 at 04:39:06PM -0800, Alan wrote:
>  Sarge
> >hasn't been released yet, and Debian stable uses a 2.2 kernel still.
> 
> As much as I hate "distribution wars":
> 
> ajb at sprocket:~$ cat /etc/debian_version
> 3.0
> ajb at sprocket:~$ apt-cache search kernel-image
> kernel-image-2.2.20 - Linux kernel binary image for version 2.2.20.
> kernel-image-2.2.20-compact - Linux kernel binary image.
> kernel-image-2.2.20-idepci - Linux kernel binary image.
> ...
> kernel-image-2.4.18-1-386 - Linux kernel image for version 2.4.18 on386.
> kernel-image-2.4.18-1-586tsc - Linux kernel image for version 2.4.18
> kernel-image-2.4.18-1-686 - Linux kernel image 2.4.18 on
> kernel-image-2.4.18-1-686-smp - Linux kernel image 2.4.18 on
> ajb at sprocket:~$
> 
> While it's true that 2.2 is the default kernel(is this still the case? I 
> know you can use 'bf24' to boot directly into a 2.4 kernel),changing to 
> a 2.4 kernel is as easy as an apt-get.
> 
> This is _no_ different than updating your kernel on a RedHat box, or 
> even downloading the source and compiling your own.

Question for you: I just built a machine for $300.  Can YOU install Debian
stable on it?  (The answer is no.)

I built a machine before Woody was even released.  Could you install
Debian on it?  No, not even with 2.4.18.  You needed 2.4.19pre1's IDE
patch to make it work.  And Debian still doesn't have that IDE patch,
three years later.  You can't run Debian stable on any machine with an
ATA100 or ATA133 chipset.


What Debian needs to do is rename "stable" to "decrepit".  At the time
woody was released, ATA100 was standard and 133 was the latest thing.  I
protested more than a month before it was released that Debian needed to
have this patch, and the Debian kernel maintainer told me that he would
not incorporate an "experimental" patch because it wasn't a security fix
and because if it wasn't good enough for Linus, it wasn't good enough for
him.

Of course, in the grand fuckup that is Debian politics, nobody is in a
position to tell this developer to pull his head out of wherever he'd
managed to insert it and look at both the current state of the hardware
and of the maturity of the IDE patch.


On top of all of that, at the time, the Debian people said they were
rewriting their installer and it would be about a year before it was ready
for its first real test.  In that time, people with new machines could not
install Debian at all, and after it only if they went after the
experimental new sarge installer for another 6-8 months.  Them's the
breaks, say Debian.

This reminds me of a tale I heard about NetBSD.  (Details may be wrong
because it was a long time ago and I wasn't personally involved.)  Back
then, ISA SCSI cards were used and there was a problem with using some of
these on DEC Alpha systems with > 16 megs of memory.  For some ridiculous
amount of time (longer than a few weeks), NetBSD did not support > 16MB on
any system, regardless of its CPU architecture, because those damned
Alphas couldn't handle it.


In labour relations, there is this passive-aggressive way to go on strike
without actually doing so called work to rule.  I bring it up because I
ran into someone who had actually never heard of this kind of thing
before.  The basic idea is that most companies (or large community
projects) have a significant amount of policy and process in place.  So
much, in fact, that in order to get anything accomplished in a reasonable
timeframe, most of the policy gets ignored.

In a work to rule scenario, workers follow every single policy to the
letter, realizing that they are technically still doing their jobs, but
that by using the existing policies they can bring work to a near
standstill.  My first experience with this kind of thing was a Debian
developer applying the principle, but that's a long story I won't go into.
I will point out, however, that this demonstrates how policies can
sometimes get in the way of doing what needs to be done.

Okay, Debian stable has a policy that no new software goes in.  Period.
If there is a need for a backport of a security patch, it is backported.
There are times (required boot drivers!) when this doesn't make sense, but
the powers that be insist that the policy must be followed to the letter.
Consider that.