[Wikimedia-l] Post mortems (second attempt)

The discussion about post-mortems arose rather organically, not as a result of a decision to use a certain medium. The participants were: Jonathan Cardy, Erik Möller, Dariusz Jemielniak, myself, Ben Creasy, Asaf Bartov, Jon Beasley-Murray, Bence Damakos, Luis Villa, Eddie Erhart, Liam Wyatt, and Tisza Gergő. I think it is fair to say that we had a general consensus that: When something does not go well (for instance, various software releases), it would be highly valuable for the Wikimedia Foundation's senior leadership to prioritize creating a thoughtful and official post-mortem document and discussion. Post-mortems can support learning by the organization, and also by other people and organizations who might take on similar projects, and can make it possible for those who feel unheard to "move on" (as is so frequently requested), with the knowledge that their opinions have been heard and may be incorporated into future efforts. The one time the organization had such an executive-led post-mortem was about the Belfer Center Wikipedian in Residence; I think we all agreed the outcomes of this post-mortem were valuable:[1] In addition, Gergő mentioned his post-mortem on the Media Viewer, which I (and perhaps some others) had not been previously aware of.[2] The discussion concluded with the idea that perhaps the present crisis offers a good opportunity to instill a culture of reflecting on mistakes into the organization's ethos. Since Dariusz was involved in the discussion, I'm confident that idea will be brought back to the Board, and I view this as a positive outcome. Below, I'll paste Erik's initial comment, which began with the words "For the record..." (which I take as an indication he is willing to have the words republished), and which generated 29 "Likes" (far more than I'm used to seeing for any comment at Wikipedia Weekly). Erik Möller: 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. (end of Erik's comment) -Pete [[User:Peteforsyth]] [1] https://outreach.wikimedia.org/wiki/Assessment_of_Belfer_Center_Wikipedian_… [2] https://www.mediawiki.org/wiki/Multimedia/Media_Viewer/Retrospective On Sun, Feb 21, 2016 at 3:34 PM, Anthony Cole <ahcoleecu(a)gmail.com> wrote: