[Foundation-l] Development tasks and project needs

Erik Moeller erik_moeller at gmx.de
Tue Mar 29 11:22:00 UTC 2005


Sj-
> Nonsense.  Money is an excellent resource when properly applied, but
> there is no fundamental reason why we should have fewer volunteer
> developers than Linux has, for instance.  After all, what are we
> writing but the kernel for our collective intelligence?

An interesting comparison. However, it is worth noting that there is a  
billion dollar industry behind the Linux kernel, and many of its most  
active developers are directly employed by corporations like IBM, Novell  
and Red Hat. Hence, development tends to be driven by corporate interests.

It is no coincidence that the growth of desktop Linux, both on the kernel  
level and in the userspace, is depressingly slow (we're talking about a  
product which is almost 15 years old). It is facing many of the exact same  
problems that Wikimedia is facing, and the long-term solutions are likely  
to be identical: getting money from users to developers.

Open source, and volunteer activity in general, is not some kind of magic  
pixie dust which you can dispense to make everything just wonderful. This  
is especially true as the complexity of the project increases. We've had  
more offers for help on MediaWiki than we can count. Volunteers who say  
"I'm going to write some code" are easy to find. Those who actually do  
something are very rare indeed. Those whose code gets past Brion ...

Fundamentally, I also believe there is a moral obligation for each one of  
us, as individuals, to work on turning the free content world into a real  
economy, instead of trying to find ways to get ever more people to exploit  
themselves to an ever larger degree. We are not just dealing with  
"resources," we are dealing with human beings who deserve to be treated as  
such. I personally would encourage people to stop working on MediaWiki, or  
Wikimedia, if I can see that it harms their life, their children, their  
loved ones, their career.

That being said, here are some things we can do that will attract  
volunteers, short term and long term:

- provide an easier to use sandbox for people to work with the MediaWiki  
codebase, e.g. allow anyone to create a test-wiki from CVS HEAD, and  
perhaps even make the code wiki-editable. (Chroot jails, regular re- 
initialization, developer approval and the like help to keep the whole  
thing secure.)

- reach out to university CS departments. These give out projects to their  
students, and Wikimedia is highly loved in those circles, so it should be  
relatively easy to convince some of them to have their students work on  
MediaWiki.

- systematically improve the technical documentation and pages like "How  
to become a MediaWiki hacker". The parser documentation is especially bad  
to non-existent.

- do a better job of organizing the development-related pages. Meta is a  
mess. wikipedia.sourceforge.net is a static HTML page. It used to be a  
wiki, but SF.net was not suitable as a host. MediaWiki.org, in my opinion,  
should be a server outside our regular co-location facility, and be used  
for things like the above sandboxes, internationalization work, etc.

- use Subversion instead of CVS. It has a CVS-like syntax, so should be  
easy to learn for existing volunteers. It makes life a lot easier: atomic  
commits, very easy branching, moving files while preserving history is  
simple, etc. While we're at it, we could try to centralize extension  
development in one place.

- write and enforce a clear policy that new code needs to follow certain  
documentation standards before it is allowed into the HEAD branch.  
Undocumented code can be removed.

- long term: encourage the use of collaborative real-time editing tools  
like SubEthaEdit, use them in combination with Voice over IP to tutor new  
volunteers.

Needless to say, even the work that needs to be done to increase the  
efficiency of our volunteer efforts can be done faster if it is paid for,  
or at least organized under an official title. For the tasks for which you  
are qualified to do so, I highly encourage you to put ideas into action.  
Talk like above is cheap. Real work is worth paying for.

Regards,

Erik



More information about the wikimedia-l mailing list