On Tue, 5 Jun 2001, Will Newton wrote:
>> In theory, yes. In practice? Not really. Sure, the odd patch is
>> submitted, but that's about it. Not much else has come from the "open
>> source community"...
>
>Probably because:
>
>a) XFree86 is a big horrible beast (esp. if you're using a dial-up)
Yes, but that is minimized by using CVSup IMHO. If it would be
beneficial to other developers out there, I would be willing to
consider making XFree86 source tarballs, with incremental patches
available somewhere similar to the Linux kernel's patch sets.
Then one only need download the source once, and then get patches
from then on.
>b) NDAs are often involved, which deters volunteers
Perhaps, but it is a very very very minimal NDA. Definitely not
an evil NDA, or many people myself included wouldn't agree to it.
>c) There are limited projects - cards are usually not supported because of
>an nVidia-like situation or have a well known "guru" doing the coding. No
>nice easy intros to coding drivers.
The DESIGN document is pretty much a good start.
>d) There's limited real world experience in the field. Most
>driver level programmers do it as a day job, it's not something
>you pick up everywhere.
While that may be true, there are many people working on drivers
who are college/university students as well.
>e) The mere fact that a company like VA Linux is working on XFree makes
>people think they are doing everything when they aren't
I don't think that is accurate myself. Looking at changelog
entries shows who is doing the work. Statistically, the most
checkins come from VA employee's or ex-VA employees. Any code
checkins that originated as patch submissions usually credit
those who submitted and/or wrote the patch.
Either way, I don't see how any of that would impede someone new
from writing code for XFree86. If someone gets started, and has
questions, they are usually answered quite well on either
[EMAIL PROTECTED], [EMAIL PROTECTED], dri lists, etc.. and quite
often from core developers.
>There are other reasons too of course.
The main reason IMHO is that XFree86 is very large and takes many
months to be able to just navigate the source tree, and begin to
understand the Imake build system, etc. As in kernel
development, one must have a deep interest in overcoming the
initial learning curves to get seriously involved in something
like this. It certainly isn't a one-nighter thing one can just
pick up.
What might however help newcomers to XFree86 development (of
which I still very much consider myself BTW) is more
documentation on things, examples, etc. That will come in time
though I believe. Perhaps new developers would be interested in
coordinating some FAQ docs or a "Learning to hack XFree86"
website? Sourceforge project?
Just some ideas to throw around. Comments/suggestions/flames?
----------------------------------------------------------------------
Mike A. Harris Shipping/mailing address:
OS Systems Engineer 190 Pittsburgh Ave., Sault Ste. Marie,
Red Hat Inc. Ontario, Canada, P6C 5B3
http://www.redhat.com Phone: (705)949-2136
----------------------------------------------------------------------
Latest XFree86 test RPMS: ftp://people.redhat.com/mharris/testing
_______________________________________________
Dri-devel mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/dri-devel
