User:Risker/Scratchpad

Rambling thoughts about how we all can work together

A very significant portion of the active (5+ edits/month) and very active (100+ edits/month) community can remember the days when the Wikimedia Foundation had only a handful of employees, almost all of the $3 million budget really did go to keeping the servers running, and the entire enterprise was held together with hacks and fixes and the sheer willpower of tens of thousands of volunteers. Most of us are glad that those days are far behind us, too: the rapid growth of visitors and the editing community, as compared to the engineering corps and the available hardware, meant that the focus of the entire organization was on preventing things from blowing up rather than making significant improvements.

Today, the WMF itself has almost 200 employees; there are about 100 paid employees of chapters, and another 50 or so paid Wikipedians in residence in GLAM institutions around the world. Revenues for 2013-14 are expected to be $50 million for the WMF itself, with $21 million of that dedicated to "Product and Engineering". The plan calls for 115 employees in the Product and Engineering departments by the end of the fiscal year.

These are, for the most part, very positive developments. The dramatic success of the WMF fundraising program over the last several years has permitted long-overdue structural, software, staffing and process improvements, such as the introduction of the "Vector" skin, addition of the Virginia data centre, and a full-time operations team. Software improvements are more carefully reviewed in advance, and deployed on a regular basis. The mobile site was developed and deployed, and is responsible for an ever-increasing segment of visits to our projects. Until about 18 months ago, however, most of these changes happened in the background, or had minimal impact on the active editing community.

English Wikipedia as a testing ground
As the largest WMF English-language project, English Wikipedia is the primary testing ground for all new WMF software and extensions. It's not hard to figure out why: anything intended for all-sites deployment isn't ready unless it can work on this project, so they might just as well start here. For a long time, this wasn't considered a big deal; the "tests" often involved much-wanted enhancements or were so deeply into the background that the average editor didn't notice them. However, as the WMF Engineering team has started to move into "editor engagement" and user interface "enhancements", the tests have become more obvious and editors have increasingly felt like guinea pigs, never being quite sure what will change each time they log in. "Why does my watchlist suddenly have bright new colours?" "What am I supposed to do with this feedback, and why does everyone think this article with 15 photos needs more?" "Why did that tab move, it was over there yesterday!"....and these are just minor changes that were initiated by WMF staff.

The elephant in the room is the VisualEditor extension, which was peacefully working its way through early testing when, to the shock of a lot of editors, it was deployed as the default editing interface for articles and user pages, without any testing of two major features (templates and referencing).