Debian flamewar (was: OpenOffice doc...)

[Benjamin Scott]

On Tue, 15 Feb 2005, at 6:26am, neil at jenandneil.com wrote:
>>> The single-large repository part is what's important.  It takes a little
>>> longer to unify configuration ...
>>
>>   No, it takes a **LOT** longer.  If the number of components in a
>> Configuration Management scenario is N, then the number of potential
>> interactions is (N^2)-N.  Think about that for a minute.
>
> I don't buy that.

  Do I need to draw a picture?  :-)

  (Actually, (N^2)-N is wrong.  That assumes all components have precisely
one interaction to worry about, and that all interactions are one-way.  But
I don't remember the right one, and I flunked combinatorics. :) )

  If I have 3 packages, A, B, and C, I need to test A with B, A with C, and
B with C -- 3 interactions.  If I have 4 packages -- A, B, C, and D -- I
need to test A with each of B, C, and D, and B with each of C and D, and C
with D: 6 interactions.  Here's a table:

# Pkgs	# Ints	Ints
2	1	AB
3	3	AB AC BC
4	6	AB AC AD BC BD CD
5	10	AB AC AD AE BC BD BE CD CE DE
6	15	AB AC AD AE AF BC BD BE BF CD CE CF DE DF EF

  Notice how the progression for the number of interactions is distinctly
non-linear.  I may not be good at abstract mathematics, but I can tell the
one value is getting bigger a lot faster the the other one is.

> It takes a LOT longer for it to hit stable, but by that time it's
> ludicrously rock solid.

  It's not a matter of quality, it's a matter of time-to-test.

> It really doesn't take too long for decent packages to hit Sarge now (or
> for the last year really).

  Exactly my point.  testing and unstable are moving targets.  It's in flux.  
To test something, it needs to be *unchanging*.  Once it's tested, you can't
change it again, or you have to test again.  Since packages build on one
another, you have a real hard time getting a release out.  That's why woody
is two-and-a-half years old at this point.  While Red Hat's offerings are
definitely of a lower quality then Debian, is woody *three years worth of
testing* better?  Hell no.  You're way past the point of diminishing
returns.

> This assumes they are too slow, but I don't feel too limited by that
> release cycle anyway.

  I'm happy you don't feel that way.  Seriously.  However, many *do* feel
that way, for reasons I've explained elsewhere (see my comments WRT CM IRT
Derek Martin).

>>   As soon as you switch to a "spin-off", you lose the benefit of the huge
>> Debian repository.
>
> Not true.  KnoppMyth does a great job of running my TV.  And they manage
> their own repository (in addition to the Debian testing/unstable ones and
> a few others).

  Right, but now I just can't type "apt-get install foo" and magically have
everything work.  And one will quite quickly get into the "dependency hell"  
that people are all too quick to blame on RPM.

> That's sorta what I said above, but a different kernel, even for the
> install, is rather painless and can fix your storage problems.

  And I get this kernel how...?  :)

> The install kernel is 2.2 and typically doesn't support this, but there's
> an optional 2.4 kernel on the install for this kind of stuff.

  Cool.  Wanna tell me how I use it?  I've got Debian 3.0r2 images on my
hard disk.  (I see 3.0r4 is out now, but they keep telling me not much has
changed...)  I've attempted installs of this Debian before, but my HD is 160
GB and most of the stuff below the 24-bit barrier is already allocated.  
The woody kernel is so old it can't see above the 24-bit barrier.  I've got
a tiny space (a couple hundred MB) for the base system to install in.  That
I can do.  But I can't fit a C compiler to build my own kernel.  I tried
downloading an updated 2.4 kernel, but I couldn't get the initrd to work.  
I even went to far as to build my own initrd image with a custom shell, and
played around with it as much as I could.  It all looks to be there, it
just pukes once I leave the initrd and the kernel tries to pivot the roots.  
It eventually gets to:

	pivot_root: No such file or directory
	/sbin/init: cannot open dev/console: No such file or directory
	Kernel panic: Attempted to kill init!

and then halts.  Then I reboot back into FC3.  :)

> On that note, I've been through the new installer a few times and while I
> never minded the old one much, the new one is really slick.

  The Debian zealots I know have been telling me the installer is going to
get much better Real Soon Now for over five years.  You'll pardon me if I
don't hold my breath.  :)

> Within a few months, I think it'll be released.  But if you want that
> functionality, it's effectively available now.  It's not like you have to
> wait for it's "release" to use it.

  Sure.  How do I install it?  In the past, I've been told the way to
install it is to install stable and then "apt-get dist-upgrade" (I think
that was the command).  See above.  :)

>>> And as for non-servers, just use Testing.  It's stable enough for any
>>> desktop.
>>
>> Further evidence that Debian zealots lack any concept of Configuration
>> Management.  :-)
>
> I don't really see anyone doing anything better than APT, even on a large 
> scale here.

  Read my keystokes: It's not the frelling package manager.  :-)

  Configuration management is completely hopeless if one's configuration
varies depending on when you happened to pull your package set from
testing/unstable/sarge/sid/pixar/whatever.

> As for deploying hundreds of machines, I have no idea how that's connected
> to choice of distro ...                ^^^^^^^^^^^^^^

  Exactly my point.  :-)

> tab-completion has ruined my memory.

  Me too.  Hell, it's ruining my mind.  I've lately found myself trying to 
tab-complete file names I haven't created yet!  Damn shell doesn't have 
that DWIM module installed yet...

> apt-src is the package.

  I'll check it out.  But not tonight.  Don't have time.  (Sleep is for the
weak.  Caffeine is my sacrament.  Are those spiders real?)

-- 
Ben Scott <bscott at ntisys.com>
| The opinions expressed in this message are those of the author and do  |
| not represent the views or policy of any other person or organization. |
| All information is provided without warranty of any kind.              |