Give A Little, Get A Lot ... But How Much?

The term community contribution is one of those phrases in the open source world that's gone from being shopworn to downright fly-blown. Everyone talks about it, but less often do we dig under the skin of those words to extract a little true meaning from them. Given that Joe "Zonker" Brockmeier is moderating a panel on the subject of community contributions at the Linux Foundation's Collaboration Summit this week, I thought I'd chew Joe's ear about this subject and ask him a few questions.

Q: When you measure community contributions, what criteria do you have for measuring? For instance, if you're tracking a specific project, do you count only contributions that have been merged into the mainline for the project, or do you also count in unofficial contributions / forks?

A: "I'd look at all contributions that are offered to a project, regardless of whether they're actually merged into upstream or not.

"Forks can be a bit tricky -- is a company maintaining a 'fork' and then contributing patches back at some point? I'd consider that a contribution. Companies maintaining a fork (like, say, Apple with some of its use of BSDed code) that doesn't go back to the originating project, not so much.

"Note that contributions can take forms other than code as well. Companies also contribute things like legal services, infrastructure (hosting, bandwidth, etc.), sponsoring developers to attend conferences, and many other methods of improving a project that aren't necessarily code-related."

(This last point generally goes unmentioned in most of the discussions about being a project contributor that I've heard. I suspect that's partly because such contributions tend to be doubly invisible: if you're a code contributor yourself, for instance, you don't tend to think a lot about who's supplying the bandwidth or the infrastructure for your Subversion or git repository.)

Q: Which contributors are the most surprising sources, either for the quantity or quality of their work? (And the opposite as well: who are the ones who contribute the least, despite appearances to the contrary?)

A: "I'm not sure I can give a fair answer to this, really. I'd say there are some companies that are very well regarded in the community that do a lot for code contributions to some projects, but then if you look at their support for community infrastructure they may fall down a bit. There are others that aren't quite as well-regarded, but that actually tend to donate rather a lot of developer time and money to community infrastructure. On the whole, though, I think it evens out.

"The only company I might pick on specifically is Oracle. I think the 'Unbreakable Linux' offerings from Oracle are very poor community hygiene.

"In general, though, I try to avoid pointing fingers and judging contributions from others -- there are a lot of factors that go into a company's decisions to support specific open source projects. What may seem right from the community perspective may not make any sense at all from the perspective of a company's customers or shareholders."

(Another important point. It's easy to forget, in all of the mutual back-patting that goes on, that most open source of importance is produced by companies with a bottom line and a body of shareholders -- many of whom are not "enthusiasts", but simply see a company that turns a profit and maybe pays dividends, and expect that to continue to be the case.)