Post in Wikipedia Weekly Facebook group by Erik Möller

Erik Möller - February 20, 2016

For the record, the desire to "hit the deadline" for the VisualEditor release extended beyond any grant agreement. (If that had been all there was to it, I would have pushed back.) The Board independently had repeatedly pushed to meet the arbitrary schedule, and even the team itself was motivated at the time to finally go in front of a larger audience, as I think James would attest. The project had already been delayed repeatedly; there was even impatience in parts of the community and the press.

So there was a general, shared feeling that we needed to do better. I take responsibility for not putting on the brakes; it was due to my own lack of experience and focus at the time.

My takeaway is that we simply didn't yet have mature processes in place for a release of this scope and complexity. For instance, even the community liaison support was conceived at the last minute. We made a lot of changes in the years that followed, some under Sue (e.g., addition of a "Beta Features" program, improved testing infrastructure, QA support), some under Lila (focus on performance & analytics), some after I left. I'm sure in some respects there's still lots of room for improvement in engineering processes.

I agree with Ori's point on the list, however, that most of this continuous improvement has been going on in spite of, not because of, what's been happening at the top. That's in many ways how it should be -- WMF's engineering organization has the capacity for independent self-improvement in all areas. But of course the drama that's going on right now is entirely avoidable and depressing, and if it continues, will damage existing capabilities and lead to regressions in important areas as key people leave.

I don't have regrets about leaving -- I was going to stick around for another 1-2 years at most; I was never cut out to be a lifer, and I left voluntarily because it was clear things were going to just continue to deteriorate at the top. But if some of the key folks in engineering left, that would really really suck. You don't want that to happen, trust me. These are good, super-talented people, and the institutional/technical memory that would leave with them would set the org back severely.