Openness: Apache as a guiding model (was Re: GFS removed??? )

[Ville Skyttä]

On Fri, 2005-03-18 at 09:45 -0500, Jeff Spaleta wrote:
> On Fri, 18 Mar 2005 10:15:33 +0100 (CET), Nicolas Mailhot
> <nicolas.mailhot at laposte.net> wrote:
> > Choosing to ignore Axel's input on this subject is so utterly ridiculous
> > from a technical point of view one can only conclude other considerations
> > are taking over now.
> 
> This is completely and utterly wrong. No one is being ignored. 
> Let me be clear... opinions expressed in -devel-list are not ignored. 
> Axel is not being ignored. He's expressed an opinion, its been heard.   
> Whether contributors inside the Fedora project incorporate Axel's
> ideas into the Extras will be self-evident eventually.  From what has
> been said here, clearly Ville wants to know Axel's opinion.. and if
> thats the case Ville will most likely incorporate anything he finds
> useful in any of his contributions to fedora contributor-wide
> discussions.

Right.  And sorry about the late reply.  I think I could have prevented
some of the crap flying around this time by following this list more
closer and responding earlier, my bad.

> I would find it rather odd and malicious if Ville now
> ignored Axel, considering Ville has explicitly desired Axel's input.

My goal is the exact opposite of ignoring people who have valuable
contributions to the issue at hand.  To prevent further guesswork, an
essay on political short term history follows.  (Regarding the input
seen in this thread, I haven't really noticed any that I would consider
being in the scope of my RFD, more on that at end of this mail.)

There's a lot of IMO's and "I'd like"s here on purpose.  Please note
that I have absolutely no authority on the issue at hand, just happened
to start the discussion this time, and presented my POV. 

Dams, Thorsten, and I discussed in private mail about importing the
fedora.us kernel-module-devel to livna.org since various module packages
therein already depend on it, and Extras/Core alone doesn't currently
provide "good enough" infrastructure for building those kernel module
packages.  (Various bits have been already improved in the latest
Rawhide kernel(-devel) packages, but there's still room for
improvement.)

At some point, Warren was brought into the discussion, and he suggested
discussing this once more on a mailing list to see if we could find a
way to benefit a wider audience instead of just applying a livna.org
specific solution.  He suggested -maintainers due to the noise involved
with these discussions in earlier tries eg. on this list.

I wasn't entirely comfortable with using -maintainers, but had no better
suggestions; the earlier related noise levels on this list had been
unacceptable and had resulted in things actually being _worse_ than
before the "discussion".  I thought -maintainers would be better than
private mail, which would also have been an option.

Anyway, I thought Axel's and Dag's opinions, input, and experience would
be valuable, they do package a lot of kernel modules for a big set of
vanilla and custom kernels for a wide user base.  Rik (who was also
brought into the PM discussion shortly after Warren) agreed, so I pinged
Axel and Dag to see if they were interested, and they were.  I notified
the previously Cc'd folks in the PM thread above about that, and
suggested Axel and Dag would be subscribed to -maintainers in order to
be able to participate in this discussion.

At this point, I wondered whether I wanted to engage in this discussion
one more time in the first place due to earlier somewhat bad experiences
where the discussion had turned away from technical issues to nonsense,
so I didn't act immediately.  Then, I got busy with other stuff for a
week or so.

After that week or so, I noticed Rik had started to incorporate some of
my Bugzilla'd RFE's to the Rawhide packages, so I rushed to post the
request for discussion to -maintainers before it would be too late to
discuss things and sanity check those RFE's.  Many eyes etc.  My mistake
was to not ensure that the suggestion/request for subscribing Axel and
Dag actually reached people who can do something about making that
happen, sorry about that.

  Now: who should this subscription request be sent to?

It's not too late yet, the discussion on -maintainers hasn't really
started.  Thorsten responded to the list, and I got a few "seconded,
well thought out" -like private mails, but that's about it.

In the meantime, feel free to comment on this list, or the Bugzilla'd
RFE's mentioned in my RFD, or send me private mail and I'll forward the
opinions.  But I'd like everyone to stay within the "infrastructure"
scope for now.  How to name the module packages and how to get
depsolvers to Do The Right Thing with pulling them in in various
scenarios is more or less an orthogonal issue, and anyway one that _can_
be tackled separately.  And because it can, IMO it also should, due to
the whole big picture of kernel module packaging being a relatively wide
and complex one.

So, let's avoid the political/personal fuss, keep things technical, get
the basic building blocks done first, then proceed to the
naming/depsolver part, ok?  Keeping the noise level down only improves
one's chances of being heard, no matter where the discussion goes on.
Thanks in advance!