[Members] Re: disconnect from board to active developers

Jim Gettys jg at laptop.org
Fri Oct 20 07:53:13 PDT 2006


First, since I'm not running for the board this year, you can draw what
you will from these opinions, which I'm voicing in public much of what
I've expressed in the BOD meetings the last several years.

I will continue to oppose spending money to fund full time long term
employees or contractors to write code for X.Org.

It is all too easy to "vote" for things when spending money,
particularly when you look at it as just a few dollars out of your own
(or other's) pockets, rather than having the commitment to "make it
happen" yourself.

I do believe that money should be spent on some topics that money can
solve well (machines, system management/infrastructure, professional
documentation help for our developers, many of whom are not native
speakers of English), and most particularly for community building and
outreach.  

Our community is two fold: those who develop X itself, out of interest
and passion and need, and those who use it, which, since the
fragmentation of the community into desktop projects and the base window
system, is primarily toolkit developers.  

We no longer have a significant role to play in the end user customer
facing shows, for the same reason you don't see the "Linux Kernel" booth
at Linux World, we should not.  Others (e.g. Gnome/KDE and the Linux
distro's) now play that role.  We are represented every time X is shown
on a desktop, laptop or embedded device anywhere. If people are funded
to go to such a show, it should only be because that is a convenient
location to meet with others in the community to work on common
projects.  Technical events that directly encourage developers
participation are a good use of our funds, and we must realize that
there is a great untapped resource in other parts of the world than the
US and Europe.

Most importantly, I believe:
  o Open source code without a *community* around it is useless useless.
It isn't maintained, it doesn't evolve, it doesn't get adopted.  X's
history is littered with such dead whales, from Xie (mostly DEC, IIRC),
to XCMS, to PEX (mostly to avoid doing OpenGL, as far as I could tell,
due to concerns at the time about dependence on a competitor), to Shiman
and Associates' MAS, which has seen no uptake in the community despite
large expenditures of funds from the predecessor organization to the
X.Org Foundation.  

Google's SOC is a different animal: it is funding (usually) students, to
work on small projects that are *integral* parts of an existing
community for a limited period: in fact, on of the purposes of SOC is
explicitly to help teach students and others how to join the community.
Example: Manu Cornet's work on GTK+ and on Xephyr this summer will live
on, and I hope Manu continues to be part of these communities in the
future. I could see X.Org doing a similar program on its own given
funding; again, it is a community building program, and the failure of
any given project is not a problem to the overall community.

If people don't care enough to put in their (or make the case to their
management that this is in their companies management), it probably
isn't worth doing.  If, as a company, you see a need that you should be
done and can't find other people and organizations to help, you should
question whether that need is really urgent enough to be worth doing.

I do think the X.Org Foundation can and should play a role as a
facilitator between people and organizations on topics that of shared
interest to the community and help organize cooperative efforts between
companies and people to work on areas of interest.

But it has to be as part of the community that technical decisions and
evolution of the system, for without this community, code is a dead
whale on the beach (as I once had X described some years ago).  And the
history of X is littered with dead whales, which, one by one, we are
dragging out and burying at sea.
                     Regards,
                        - Jim Gettys


-- 
Jim Gettys
One Laptop Per Child





More information about the xorg mailing list