(This was originally written as a position statement for the Developer Summit, but then I ended up submitting a different one: Sister project incubator. There is also another non-submitted position: Turn MediaWiki core into an embeddable library.)
- We need to bootstrap a commercially viable MediaWiki offering to grow our software ecosystem.
The size and growth of a software ecosystem is limited by the influx of resources: money and volunteer time. The Wikimedia movement is currently not in a good position to sustain its own growth: volunteer developer participation is shrinking and the only significant money stream is donations to the Wikimedia Foundation, which can only be increased by more aggressive fundraising, resulting in more and more community pushback, while there are more and more non-development-related projects that derive funding from the same source. This poses a problem as the current amount of technical resources are ridiculously low, even compared to the current popularity and impact of the movement, not to mention the size of our ambitions.
Money is the easier (or at the very least more predictable) of the two resources to influence (also, volunteers tend to follow the money to a large extent; people prefer to learn technologies which get them hired). From the example of other players in the free software/content/knowledge movement, the main alternative income sources are reliance on a few big supporters via grants or deals, commercialization of content and commercialization of the software. The first two are extremely unappealing for us; a commercially viable MediaWiki would fit into our mission well, though (of course it would have to be managed by a separate, for-profit organization). MediaWiki is opinionated software that promotes freedom and openness through its fundamental design decisions; and more widespread use would increase familiarity of potential new editors with the user interface.
The current technology spending of the movement is between around 30-40 million dollars. The guesstimated annual revenue of the largest competitor (Atlassian Confluence) is $100M; even just cornering 10% of the market would mean a 30% increase in resources. Technologically, MediaWiki does not seem inferior to Confluence; while it has ignored enterprise features for a long time, with some dedicated development it's in a good position to catch up.
Such an effort would involve improvements to MediaWiki's "sysadmin usability" (installer, extension and configuration management, security updates, upgrades), bundling of popular services (e.g. Parsoid / VisualEditor), documentation, and outreach / marketing. These are not particularly difficult things; a few full-time employees worth of funding should be enough to kickstart such a commercial spinoff; even with Bay Area wages that's less than $1M, a fraction of the potential payoff.