[lug] Fedora *MEETS* KRUD comments wanted

[Nate Duehr]

Michael J. Hammel wrote:
> Lots of comments, but I'm afraid I agreed with almost none of them. 
> Since this is mostly a topic based on opinion I'll let it go.... <snip>

Agreed to disagree.  Although I don't think we're that far off from each 
other.  I'm only pointing out that things are on a particular strategic 
vector at RedHat and it's not a better one for the Community at large at 
first glance.  The direction is away from Community and more toward 
Corporate.  Isn't it?

One interesting thought came to mind during this though...

Just how *much* do developers from each of the distros contribute to the 
raw packages of software that make up Linux as we know it today.  And 
how would one figure that out...?   Hmmm....

Then of course, is the "how do you quantify it?" question... lines of 
code?  Uggh.

Obviously with Debian you could compare names and e-mail addresses from 
the GNUPG keyfile to changelogs of various distros or something with a 
script...

Is there a RedHat developer list that's "official" somewhere?  We all of 
course, know Alan Cox's name at RedHat -- even if we've never met him. 
That's an obvious one.

I have not a clue at all about how you'd get data on the other distros. 
  I think Gentoo has a published developer list... hmm.

It would be interesting to try to quantify where the "real work" in Free 
Software comes from.  Might lead to some interesting discoveries...

I'd make a large bet that there's specializations popping up due to 
where Linux is on the evolutionary timeline.  There are probably a lower 
  percentage of "distro maker" type developers who are regularly 
contributing upstream to the root packages these days -- as they spend 
their time working on the distribution of choice.  They work hard on 
packaging and distribution.  ??

(This is just a guess: The root packages of things I've looked at these 
days are all either large teams with their own developer crew, or small 
two and three person projects.  And few of those developers are 
"packaging" people that work on distributions.)

Obviously the work done by "both" types ("both" in quotes because 
there's no good way to label individual people) is important.  But it 
would be interesting to find out how a generic percentage of developers 
who work for any particular distribution relate back to the main 
packages that actually make up what we might call a "modern Linux distro".

Unfortunately it sounds hideously time-consuming and maybe kinda boring 
too.  :-)  Until the data was done.  Then it'd be neat.

Government grant to study open source development history, anyone?  :-)

Making up a brief list of stuff that'd need to be known ...

- Lists of developers from the distributions being studied.  (Hard for 
some, easy for others.)
- A "standard" package list of packages that all distros generically 
include.
- Lists of developers of those packages.
- Some way to know which developer changed what in the source and in 
which source.  (ouch... this would be hard without the help of a LOT of 
CVS admins... although if the project's at SourceForge that would help!)
- Some way to quantify when there's a distro-only change vs. a change 
that eventually went upstream to make the packages better overall.

There's probably a whole documentation project hiding here just on the 
ASSUMPTIONS that would have to be made to even attempt this... but it 
intriques me a little bit to see if it's possible.....

Has something like this ever been attempted?

Nate Duehr, nate at natetech.com