[Members] Re: disconnect from board to active developers

[Donnie Berkholz]

Egbert Eich wrote:
> Do we actually know any precedence where a free project employes
> hired people to do development work - together with contributors
> from the community?

Mozilla.

In fact, there's a discussion happening related to this on another list
right now. They cite projects that are worthwhile but don't see
developers coming forward, such as accessibility. I'll paste a large
chunk of the email:

"... There are worthwhile projects that for whatever reason
don't see developers spontaneously coming forward to work on them as
volunteers, at least initially. For example in the Mozilla context we
found this to be the case with the task of improving the accessibility
of Firefox and other Mozilla products for people with disabilities.
 ...

By funding work in this area (through "grants" in the form of short-term
consulting contracts) and providing other support (e.g., sponsoring
people to attend conferences and developer meetings) we've been able to
build up a larger pool of people who are excited about doing
accessibility-related work and are accomplishing useful things (e.g.,
improving Firefox accessibility on Mac OS X). This has also led to a
significantly increased positive perception of Firefox among the
community of people concerned with accessibility issues.

I think this type of approach can be successfully replicated in other
areas, and I think there are likely other areas in the Mozilla project
and beyond where it might make sense. Here are the main elements that I
think are key to making it work:

* Have a strong lead developer who can identify promising areas in which
to fund work (basically areas which have a high potential payoff but no
one working on them) and also find promising people to support. ... In
some cases we start with the project and attempt to find a
person; in other cases we have a promising candidate and try to find a
project that matches their interests and is of use to us.

Note that this approach is the exact opposite of ideas like code
bounties where you throw a wish list out to the world and let people bid
on items. Based on my experience I think that's an unworkable approach.

* Keep the amount of funding within reasonable limits. Almost all of the
"money for code" arrangements we do are for relatively small amounts, in
the $5-15K range per project: big enough to make a difference to someone
who's a student or doing Mozilla development as a spare time activity,
but not big enough for anyone to be able to live off of. Our goal is not
to support a group of full-time developers. Instead our hope is that
people who'd like to do this sort of work full-time will use their work
with us as a springboard to paid employment with companies like IBM,
Sun, Novell, etc (or, for that matter, with the Mozilla Corporation).

* Balance "money for code" arrangements with other forms of support. A
lot of the funding we've done is for things like helping to pay people's
travel expenses to attend conferences and meetings. We've found this is
a good way to build a sense of community among the various developers
and help get them connected into the informal social networks in
whatever field they're working in (in this case, the broader "assistive
technology" space, which contains commercial firms, nonprofit
organizations, and government agencies).

Those I think are the main factors for successfully doing this. There
are some other items as well, like having a clear technical roadmap and
an overall vision and strategy that ties what you're doing back to the
main project and its goals. However I think to some extent these can
emerge organically as opposed to be planned out from the very beginning;
that was certainly the case for our accessibility work."

Thanks,
Donnie

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 252 bytes
Desc: OpenPGP digital signature
Url : http://foundation.x.org/pipermail/members/attachments/20061019/8fff051e/attachment.pgp