User:BilledMammal/Vector2022 close appeal

Close review of Requests for comment/Rollback of Vector 2022

This close, while a significant effort and made in good faith, was flawed; it poorly reflected the results of the discussion and contained procedural errors.

First, the closers failed to exclude the approximately 30 editors who were canvassed by a WMF employee and who opposed reverting to Vector 2010. This contravenes convention and sets a problematic precedent that canvassing can contribute to a consensus or to preventing one. (Including the canvassed !votes changes the support-oppose ratio from ~2:1 to ~3:2).

Second, they discounted arguments focusing on problems with Vector 2022, arguing that those issues either had been or would be addressed. While this argument was raised by a couple of editors regarding a couple of problems for most problems it was not and could not be, as the WMF has either no plans to resolve them or plans that many editors consider inadequate. Already, the WMF has refused to implement unlimited width as default despite consensus. For the closers to conclude that solutions are forthcoming merely because Phabricator tickets exist is both novel and unsound, as Phabricator tickets often remain 'in progress' indefinitely, without any fixes being deployed.

Third, they scrutinized the arguments of supporters more heavily than opposers. For example, they considered the results of the user preference survey disputed, despite no editor arguing for any other position than that it showed opposition to deploying Vector2022. They also overestimated the number of editors who found the evidence the WMF presented for Vector 2022 compelling when they said that "many users found (the presented evidence) compelling"; only a small minority referenced the evidence at all, and a greater number argued that they didn't find the evidence compelling, typically on grounds that the evidence was not representative and that the WMF had already been proven to misrepresent the evidence it did choose to present.

Fourth, they discounted !votes based on user opinions, arguing that they were not based on policy and that the experiences of people who aren't UI designers are not concrete facts. There are three issues with this: first, we don't have policies on UI design, and where policy is silent we defer to editors' judgments; second, Wikipedia is built upon the belief that anyone can contribute valuably to discussions, not just experts; and third, this suggests that ordinary Wikipedia users cannot have valuable opinions about the interfaces they use.

Finally, they erred when they used their own opinion (Since we see the changes made by the WMF as compliance with the previous RfC, emphasis ours) to decide that the requirements of the previous close had been met. This was a decision that should have been deferred to the closers of the previous discussion, who believed that they had not been met, or to a consensus of the participants in this discussion, who also generally agreed that the requirements were not met.

This appeal was drafted with input from multiple editors. ~~~~