Debian flamewar (was: OpenOffice doc...)

[Neil Joseph Schelly]

On Tuesday 15 February 2005 09:51 pm, Derek Martin wrote:
> > I don't buy that.  It takes a LOT longer for it to hit stable, but by
> > that time it's ludicrously rock solid.
>
> Um, huh?  It strikes me that you said, "I don't buy that," and then
> proceeded to agree with everything Ben said...
I said it takes a lot longer to hit stable, but the rest of my post discussed 
how waiting for things to hit stable isn't necessary.  It's convenient how 
snipping out the point of my paragraph lets you poke holes in my stance.

> And so what if it's ludicrously rock-solid, if it doesn't recognize my
> hardware?  Not so useful, regardless of how stable it may be...
Debian uses the same kernels as everyone else.  It's trivial to make your own 
if you're finding that the stock kernel doesn't support whatever setup you 
have.  

> I can't agree with that, and just the fact that you said it suggests
> to me that you're not a system administrator.  Ignoring for the moment
> the lack of vendor support options from Debian (being not a company),
> most businesses have little tolerance for unstable software.  The
> non-stable branches of Debian update far too often to be useful as a
> standard desktop platform for support reasons at most companies who
> have their heads on straight.  Notable exceptions for companies whose
> business is directly Linux-related...
I am a network administrator, for what that's worth.  I find statements like 
that just silly in a discussion such as that.  It's like calling names or 
something equally childish.

As for lack of vendor support, if that's important, I think HP recently 
started supporting Debian, didn't they?  Where I work, the servers run RHEL.  
They pay for up2date and nothing else.  Vendor support is essentially 
worthless as in this case Debian also updates packages for security problems 
free of charge.   Some businesses just have issues trusting free - they have 
all the tolerance in the world for unstable software, as evidenced by the 
number of times I have to restart systems or services in our Windows server 
systems and the annoyance of having Outlook 2003 start scrambling my folder 
structure in need of a reboot.

And business desktops by the way, since you brought it up, rarely have need 
for things past stable.  If Debian Testing is unsuitable as business desktop 
OS, then I'd say nothing in the Linux world is particularly ready yet. just 
close.

> > 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).  If I really want, I can
> > install anything from there, but then again, I don't need that on my
> > TV.  If I needed the full repositories, then a spin-off wasn't the
> > right choice I'd say.
>
> You appear to be contradicting yourself...
You appear to be grasping for straws.  Spin-offs are great for single-purpose 
or specialized machines.  If you want a general purpose machine, then you 
stick to the big repositories.  Pick what suits you - it's a simple concept.

> > Stable/Testing/Unstable are just names.  If you don't like them called
> > that, then call them Woody/Sarge/Sid.
>
> You're missing the point, which is something like, "If it ain't
> stable, it ain't usable."  This doesn't mean that YOU can't use it, it
> means that the management of an organization can't risk using it,
> because if there's a problem, it could mean a serious loss of
> work/time/money/etc.
>
> In practice, so-called "stable" releases of certain software may be no
> better, but you're never going to convince a non-technical manager
> type that it's a good idea to use something which is not considered
> production- quality by the people who are developing it...
And you're missing the point.  Don't ask your manager to approve the use of 
testing/unstable because it's just a name.  Call it Debian Sarge and call it 
a solid release that is under modern development and always up to date, 
within a reasonable few weeks timeframe to work out any bugs in new 
development.  Pick the appropriate release for the job - it's silly to 
consider anything else when you're making a decision like this.  


These are tired arguments... Testing is quite stable and reliable and 
up-to-date.  Take that assumption and you realize that everything you said 
above is meaningless.  If we disagree on that point, fine... then we can 
disagree on anything you said above and discussing it back and forth won't 
change anything.  If you haven't tried running Sarge though, then you're 
really not qualified for further telling me I don't know what I'm talking 
about.  And if you're not willing to try, then just leave it be at that - 
that you're not interested in either learning what I mean or confirming what 
you say.
-N