User talk:BilledMammal/Vector2022 close appeal

Draft discussion

@Tvx1, Toa Nidhiki05, Cessaune, Compassionate727, Æo, and Ann Teak: Pinging users who I believe may be interested in participating in the drafting of this appeal; if I have missed anyone, please ping them.

Note that the first issue, regarding the canvassed voters, is based on an assumption, as so far the answers the closer has provided to the relevant question is unclear. It may change if they provide a clear answer contrary to my assumption. BilledMammal (talk) 02:01, 24 March 2023 (UTC)

One thing I'd note is how massive the 30 canvassed editors are. There were 605 total participants here, 581 if you discount neutrals. That's 4.9% or 5.2% of total voters - and with only 226 oppose votes in total, that's a whopping 13.3% of oppose votes being canvassed. The math works out further:

• Support percentage (including canvassed votes) - 355/605 (58.7% including neutral) or 355/581 (61.1% excluding neutral)
• Oppose percentage (including canvassed votes) - 226/605 (37.4% excluding neutral) or 226/581 (38.9% including neutral)
• Support percentage (excluding canvassed votes) - 355/575 (61.8% including neutral) or 355/551 (64.4% excluding neutral)
• Support percentage (excluding canvassed votes) - 196/575 (34.1% including neutral) or 196/551 (35.6% excluding neutral)

If you include canvassed voters, Support leads by 21.3% or 22.2%. If you exclude them, Support leads by 27.7% or 28.8%. Canvassed voters result in a statistically significant swing of either 6.4% or 6.6% in favor of Oppose votes, and it doesn't appear the closers factored this in at all. If you exclude canvassed voters, Oppose votes make up barely over a third of total votes, while Support votes make up just under two-thirds. In order to justify a no consensus, the Oppose votes need to be of a substantially higher quality than Support votes - and they simply aren't. I can't think of the last time an RfC with such an overwhelming numerical majority was closed as a no consensus, and that's probably because it's extremely rare for minority arguments to be that much stronger than Support ones. Toa Nidhiki05 02:20, 24 March 2023 (UTC)

I've undone the addition of the stats; while I think the information is useful, and should be mentioned in a !vote to overturn, I am concerned that increasing the prominence of counts will increase the likelihood the editors who join the close review will fixate on "consensus is not determined by counting votes", and ignore the fact that the !votes of canvassed editors shouldn't be considered for anything, including for strength of arguments. Of course, lets see what other editors think of including those stats. BilledMammal (talk) 02:36, 24 March 2023 (UTC)

@Certes, Chipmunkdavis, SmallJarsWithGreenLabels, and Silver seren: Ping a few others from a different discussion who may be interested in contributing to this draft. BilledMammal (talk) 02:43, 24 March 2023 (UTC)

@Yngvadottir: Seemed to be interested.--Æo (talk) 12:25, 24 March 2023 (UTC)

Thanks for the ping. Despite the departure of the dev responsible for this unwanted change, it's obvious that the WMF are hellbent on making Wikipedia look like every other website which has to leave room for ad sidebars. They know best and won't respect the wishes of readers or editors, however clearly expressed, so there seems little point in continuing the fight. I was tempted to write that the WMF have lost any respect that may have remained after other arrogant impositions such as Framgate and UCoC, but the truth may be that the organisation has become so diverse and overgrown that the good people get lost amongst those whose agenda has gone off track. Certes (talk) 09:37, 24 March 2023 (UTC)

I don't agree with this at all.

This is simply getting mad at an organization that gives us essentially infinite power compared to any other organization (they literally let us run the website, and generally leave us alone). The times they do something incorrectly (Framgate, UCoC) we call them on it. I hesitate to even call this change incorrect—they were doing something they deemed warranted. I agree with the idea of wanting to change itself, even if I don't agree with some of the changes wrapped inside the change.

The so-called "wishes of readers or editors" that WMF isn't respecting—we haven't even given them a wish to respect yet. The wishes are, at worst, evenly split (no consensus) or at best, barely leaning towards rollback (rough consensus).

WMF are hellbent on making Wikipedia look like every other website which has to leave room for ad sidebars: 1) Why? Are they going to eventually introduce ads? No. There's less than a snowball's chance in hell. The backlash, the outrage. Editors would walk away in the thousands. 2) Thinning the amount of reading space is 'proven' (according to WMF) to make reading easier. I don't know how valid that actually is, but they say it. I haven't researched it.

Also, Despite the departure of the dev responsible for this unwanted change—there are so, so many potential reasons why AHollender no longer works there. We have no idea what Alex's role in the actual pushing out of V2022 was. Let's not speculate, or treat Alex like he's the cause of anything.

Also also, "good people" also implies that there are 'bad people', which isn't true. Nothing WMF has done (in this specific circumstance) has forced me to make a character judgement on any specific person. You should refrain from doing that, unless you have a valid reason to. Cessaune [talk] 11:59, 24 March 2023 (UTC)

I honestly agree with Certes that the white space trend in webdesign comes from corporate interests, in that it forces reading into a more linear pattern where everything the company wants the user to see takes up more scroll time before the user reaches the thing they're actually looking for (this becomes obvious on social media sites, which have the narrowest frame of all). But I don't think this is actually what's behind the WMF's insitence on fixed width. It's more of a "first as tragedy, then as farce" kind of thing. small jars tc 12:27, 24 March 2023 (UTC)

Thanks for the ping. I endorse the conclusions of the current close, though I'm obviously sad about the first one. My main complaint is that it just treats complaints about individual features of v22 as contributions towards the case for rollback, when they can be the road to a satisfying compromise; or as bug's that will be fixed, when they are real features that no one likes. I wouldn't want a new close to encourage any more debate about the core questions, but it should definitely recommend the further discussion of smaller modifications that gained support in the RfC, rather than deferring these problems to a potential third RfC in six months. small jars tc 10:17, 24 March 2023 (UTC)

• A couple of specific concerns I had:
  1. The closers cited in-progress Phabricator tickets as evidence that supporters' concerns have been or will be addressed. Literally nobody argued this at any point in the discussion. For anyone familiar with MediaWiki development, the reason is obvious: most Phabricator tickets remain "in progress" indefinitely, without any fixes being deployed. This is not a common sense argument and should not have been raised for the first time by the closers, much less been a substantial reason for their ultimate finding.
  2. They discounted arguments that are not based on policy. This would be sensible if we had policies on software development/UI design. We do not. Policy represents the accumulated, collective knowledge of the project, and such a strict application of it to areas for which it is not designed is imprudent. Moreover, it can only lead, as it did here, to the conclusion that Wikipedia editors cannot have meaningful opinions on the software they use. I think this is an unwise finding and precedent.
  3. They focused on how most Wikipedia editors are not UI designers and therefore only have "opinions," not legitimate, thoughtful beliefs, on UI. This is contrary to the spirit of the Wikimedia movement, which says that you do not need to be an expert in an area to meaningfully contribute to it. In fact, most of our strongest content editors are not credentialed experts in the areas they edit, and this encyclopedia would be much poorer without them. Also, we did have non-WMF UI designers comment here, and if memory serves, they all said the redesign was bad.

I may have more thoughts tomorrow. —Compassionate727 ^(T·C) 02:56, 24 March 2023 (UTC)

Those are very important points. I think #1 can be combined into the current second point, and #2 and #3 can be joined together and merged with the third point. We'll need to work out a way to do so concisely. BilledMammal (talk) 03:04, 24 March 2023 (UTC)

I've included your first point, although further wordsmithing is needed. I still need to consider how best to do #2 and #3. BilledMammal (talk) 03:35, 24 March 2023 (UTC)

Reviewing the discussion, I see that the phrabricator tickets were raised a few times; most by WMF employees, but twice by editors (As a (younger and more tech-savvy?) user who's been using the skin for more than half a year, I've gotten used to it at this point. The whitespace margin issues that are plaguing many others I've solved for myself by enlarging text for the domain, and the only real gripe that I have at this point is not having a persistent hidden table of contents across pages, which is being tracked on Phabricator. and ...Instead of reverting the entire redesign to get things like full content width and non-contracted menus back, the new skin could be updated so that the hamburger menu button stays when you scroll down, there's even currently a Phabricator request to make the content width switch stay instead of always going back to default when going to new pages, for logged-out users..., although neither of the users posted links.

The links the WMF posted were T317818 and T321498; the first covers configuration options for the table of contents, and the second covers the full width toggle. It still wasn't appropriate to base such a large part of the close on such a minor point, particularly when there were more than two issues raised, but the appeal will need to somehow reflect that it wasn't entirely original. I'll think on how to do that. BilledMammal (talk) 03:54, 24 March 2023 (UTC)

It's literally only 13-8. That could reasonably be closed as no consensus, I think. Cessaune [talk] 12:01, 24 March 2023 (UTC)

13-8? Toa Nidhiki05 12:31, 24 March 2023 (UTC)

More typical RfC size with the same percentage as this one. The overwhelming numerical majority you talk about isn't as overwhelming as you make it seem. Cessaune [talk] 12:32, 24 March 2023 (UTC)

Well, this wasn't a typical RfC. Hundreds of people voted, and 64% of them voted to rollback; that's numerically overwhelming, nearly a 2:1 advantage. You don't get to say "well if less people voted it could be no consensus" - that's not what happened. The quality of oppose votes was simply nowhere near enough to justify going against the clear, overwhelming consensus. Toa Nidhiki05 12:43, 24 March 2023 (UTC)

I agree with your points, but, again, "overwhelming consensus" is subjective. IMO, I don't think that 3:2 or 2:1 is overwhelming. If you do, then that's fine, but it will fail as an argument based on how it has been previously received. Cessaune [talk] 12:51, 24 March 2023 (UTC)

It’s an absolute definite majority. You’re never going to get a bigger difference in a discussion featuring so many people. With the idiotic bar you’re applying every large discussion ever is going to be a no consensus. Consensus is NOT unanimity. Tvx1 15:27, 24 March 2023 (UTC)

I don't think it's idiotic. Clearly the "Oppose" side are going to yell not a vote regardless of if it's 2:1 or even 3:1. I think that's mostly a bad faith argument from them, but frankly they're so outnumbered it's the best they have. However, I do think it's - at minimum - worth noting that 13% of opposed voters were canvassed, which is abnormally large. Toa Nidhiki05 15:34, 24 March 2023 (UTC)

"Idiotic"? I didn't even say that it wasn't a large majority. I simply said that I thought that "overwhelming" was giving the numbers too much credit, and that a no consensus outcome is reasonable IMO based on those numbers. Disagree, but don't call it idiotic. Cessaune [talk] 16:47, 24 March 2023 (UTC)

And maintain that if a no consensus is reasonable even for such a considerable majority, especially since arguments were overall of equal (good and bad) strength, you’ll never get any consensus in any such discussion. The bar people like you apply is simply idiotic. It’s utterly unreasonable.Tvx1 01:55, 25 March 2023 (UTC)

I find it reasonable. You don't have to. That's the beauty of conflicting viewpoints: we can both disagree. I don't see why you find it "idiotic" (it just seems like an unnecessarily strong word for something so small IMO) but that's your opinion. There's nothing more to say on this; let's worry about the actual close and the review. Cessaune [talk] 03:10, 25 March 2023 (UTC)

• My opinions about the closure are expressed here and here. Major points:

1. The strongest arguments are those based on our policies and guidelines, while the weakest are those based on subjective opinion, ... many !votes were based on personal opinions about how Vector 2022 was better or worse than Vector 2010. While these are usually considered weaker arguments, they were not entirely discarded but were not given as much weight as other points. > The RfC was about Wikipedia's interface, it was primarily about aesthetics and immediate visual/use impression, and therefore arguments based on personal opinion and taste should have been given the same weight as arguments making reference to policies and guidelines.
2. Those in favor of maintaining Vector 2022 also commented on the fact that the skin has been active on a number of smaller Wikipedias with varying degrees of acceptance. > This was a rather minor argument, raised by very few users, and in any case there were interventions from the French Wikipedia expressing strong disapproval and reporting that their voices had not been heard.
3. As some participants noted, it's likely that very few of those commenting in this RfC have any experience with UI design and, as such, the opinions presented here are only that, opinions. > Assuming that only commentators who have experience with UI design are entitled to express valid arguments is a very weak criterion. As I have written in point #1, the RfC was primarily about aesthetics and immediate visual/use impression, and opinions should have been weighed based on these criteria too.
4. At first glance, we have a clear numerical advantage for those supporting the rollback to the old skin, but many of these !votes were based exclusively on specific issues with Vector 2022, such as fixed text width, the large amount of whitespace, and the overuse of icons, as well as some accessibility issues. Others commented on bugs they encountered while using the skin. > There was a clear majority supporting rollback; not "at first glance". This majority was not considered as such and was instead considered a conglomerate of disparate opinions, while such an analysis was not made for the opposers' opinions.
5. While those in support of rolling back had a numerical majority, their arguments were relatively weak... > This is the very personal view of the closers; weak arguments were put forward on both sides.

Æo (talk) 16:09, 24 March 2023 (UTC)

• Thanks for pinging me, Æo. I don't think I can be of much use in strategizing or even considering a draft, however. I've an almost unbroken record of failure in political processes. I believe the closers believed they were constrained to close according to our norms for discerning consensus, including weighting arguments that referred to policy more heavily and regarding a majority of over 60% maintained throughout the discussion (so far as I saw) as not enough of a landslide to weigh raw numbers heavily against that principle. It is my view that they were wrong from the perspective of our mission here (the change of default skin affects those without accounts far more than it does registered editors, who can change skins once they figure out how, and hence is overwhelmingly a change affecting readers), but it didn't surprise me that they couldn't or wouldn't see that. What outrages me is the tone of the close, with its dismissal of actual reader accounts of the change degrading their ability to use the site—to read the encyclopedia—because the WMF and its supporters referred to research. At one of the noticeboards, someone referred to that as an argument from authority; yes, it is, and that's how we work when discussing article space. But in this context, it was at the very best tin-eared to put it in those terms. Also indicative, in my view, of an appalling failure to consider why we are here is the point in the close that the devs have now—at the time of closure of a sprawling discussion—fixed all or most of the problems with the new skin that the previous RfC was closed stating as necessary conditions for implementation of the skin. Flaws that were highlighted by several participants as not having been fixed when the new skin was imposed as default. As I say, the closers may have felt constrained to close the RfC as "no consensus". I get that. But "readers' experience cannot be what they say it is, we reject it as not based on research and not expressed in terms of PAG" and "the new skin now complies with all or most of the requirements established in the prior RfC, so it is inconsequential that none of those requirements was satisfied when it was rolled out" (my paraphrases) fall well below my expectations for administrators on this project. At least express sadness for having to devalue readers' testimony because of requirements for consensus, and take the WMF to task for once more utterly ignoring the community that it exists to assist. I've fulminated a couple of times at Wikipedia talk:Requests for comment/Rollback of Vector 2022, here with actual quotations from the close and here. I am saddened by the statistical argument about canvassing; I hadn't realized how many people came to the discussion from the canvassing. Otherwise, several people above are restating in a more effective way some of the points I tried to make. Realistically, I don't think there's any hope of getting the close overturned; it was foreordained by our processes. But I hope I'm wrong and I do believe pushing back is worth it. At a minimum, a new close statement would be nice. Yngvadottir (talk) 21:36, 24 March 2023 (UTC)

  with its dismissal of actual reader accounts of the change degrading their ability to use the site—to read the encyclopedia—because the WMF and its supporters referred to research. The issue here is that by making accounts to join the RFC, they are just as unrepresentative of readers as the rest of us, because everything you read on the internet is written by insane people. Snowmanonahoe (talk) 18:43, 26 March 2023 (UTC)

Draft update

I've believe I've incorporated most of the arguments presented here by various editors. I might do some minor copyediting, but assuming there are no objections or significant suggestions I will post it to WP:AN is 24 hours. BilledMammal (talk) 04:13, 25 March 2023 (UTC)

You should add something about how, in general, the refutations to the support !votes were often generally weak/not really argued strongly in the RfC, yet seem to be treated as if they were on par with the !votes. Cessaune [talk] 04:19, 25 March 2023 (UTC)

Do you have a suggestion for how? I think that is mostly addressed already in parts 2, 3, and 4, and I'm not sure what other aspect of the close we could address to do that. BilledMammal (talk) 04:21, 25 March 2023 (UTC)

As a general summary at the end, maybe? I think it's important to analyze how many times the closers did the exact same thing—stated a support argument and then refuted it with either valid but relatively unargued evidence, weak evidence, or evidence that contradicted the general RfC consensus. Cessaune [talk] 04:24, 25 March 2023 (UTC)

And that they did make any attempt to put oppose comments through the same scrutiny.Tvx1 16:53, 26 March 2023 (UTC)

I've made some prose edits to various parts of the draft, generally trying to tighten prose and clarify some of the arguments. See what you think.

One thing I'm questioning: you quote that "many users found [the presented evidence] compelling," which I cannot find in the close. What's up with that? —Compassionate727 ^(T·C) 14:57, 25 March 2023 (UTC)

Thank you, those look good to me. That sentence comes from a discussion about the close on their talk page; I've added a link to the diff, although I can remove the sentence entirely if you believe it would be appropriate? BilledMammal (talk) 03:00, 26 March 2023 (UTC)

My instinct is to remove or at least rework the sentence. I'm expecting this closure to be fiercely contested by people who just want this whole fiasco to be finished, so I think it would be better for the opening statement to stick to arguments that are as concrete and uncontestable as possible. My concern with this statement is that the evidence in favor of limited line widths actually was seen as convincing by many, and I don't want people to latch onto that as a knee-jerk reason to endorse.

Anyway, now that I think of it, we could probably just clarify that the statement is only true of the limited line widths and that would, IIRC, be an accurate summary of the discussion. —Compassionate727 ^(T·C) 07:40, 27 March 2023 (UTC)

Great to see there finally is some community effort to challenge this unacceptable close. It think the drafted appeal is pretty good, but I feel that point four should mention that the closers were not equally dismissive on opinion-based arguments on the oppose side.Tvx1 16:51, 26 March 2023 (UTC)

IMO, this argument leans heavily on interpretation (rather concrete facts) and would be better left for voters to make than included in the opening statement. —Compassionate727 ^(T·C) 07:41, 27 March 2023 (UTC)

How is this interpretation. They literally did not discuss any oppose arguments anywhere. Tvx1 15:19, 27 March 2023 (UTC)

I agree, but it's an assumption - better for the opening statement to not rely on assumptions, and leave such arguments for the responses. I think this is ready to post; I'll wait a couple of hours to see if there are any last minute comments and if there are none I'll do so. BilledMammal (talk) 23:09, 27 March 2023 (UTC)

Appeal posted

See here.

@Toa Nidhiki05, Yngvadottir, Æo, Certes, Cessaune, SmallJarsWithGreenLabels, Compassionate727, Snowmanonahoe, and Tvx1: Ping editors who commented here. BilledMammal (talk) 08:49, 28 March 2023 (UTC)