Wikipedia talk:Requests for comment/Deployment of Vector (2022)

RfC on the draft of the RfC :D

Hello @Certes, Ganesha811, Izno, Pelagic, Terasail, TheDJ, and Xaosflux: thank you for your continued help for the project of Desktop Improvements and the Vector (2022) skin. We've prepared the RfC, as suggested on the WP:VPR page. It's ready for you to review. Do you perhaps have answers to the questions in the ombox, or any other thoughts? Olga and I would like to launch the RfC in a week (next Monday), if possible. SGrabarczuk (WMF) (talk) 15:23, 5 September 2022 (UTC)

@SGrabarczuk (WMF) depending on how much feedback has already been gathered, integration options/issues with popular scripts/gadgets may want to be included in one of the lower sections. It's hard to tell from which perspective contributors will respond to the RFC in: If it is from a "how will this change my experience", that may be important to them. If they look at it from "how will this change the average (logged out) reader's experience" - they might not care about that; if it's "how will the effect other (especially newer or logged out) editors experience" it may be somewhere between. — xaosflux ^Talk 15:38, 5 September 2022 (UTC)

Mentioning current gadgets is a great idea. Enterprisey (talk!) 22:39, 6 September 2022 (UTC)

To be brutally honest, I strongly prefer current ("legacy") Vector and will opt out of the change. The only effect on me is that I'll now have to log in to Wikipedia on other devices, to apply my preference to avoid the new skin. I don't think I'm the best person to advise on its finer details. Certes (talk) 20:27, 5 September 2022 (UTC)

From a quick read the RFC page does fail to mention the User menu until Findings and results, which is quite noticeable when looking at the image comparison in the New features section. Since this is how all users would access talk/contributions and logged-in users quicky access everything else. The image side by side instantly shows the lack of these buttons but doesn't have a bullet point on what has happened to these elements. So should probably be mentioned in New features aswell. Terasail^[✉️] 00:44, 6 September 2022 (UTC)

Good catch @Terasail! This was accidental. I've now added the user menu to the feature's section as well. OVasileva (WMF) (talk) 09:30, 21 September 2022 (UTC)

RfC Question

For bot reasons there needs to be a succinct question (followed by a signature) at the top of the RfC. What blocks the deployment of the new Vector (2022) skin (preview) from becoming the default for desktop users on English Wikipedia? works but then the sections below aren't answering that question (The Vector (2022) skin can be deployed, The Vector (2022) skin cannot be deployed until additional work, not mentioned in the section "Tasks which will be completed before the deployment", is completed.) So the bolded question should be moved to the top of the page and either the question changed to match the discussion sections, or the discussion sections changed to match the question. Best, Barkeep49 (talk) 15:46, 5 September 2022 (UTC)

I did not understand the question the first time I read it, nor is it very clear now even though I have grasped what they are trying to ask. I think the question should be rephrased to be more direct, along the lines of: "Should English Wikipedia adopt Vector 2022 as the default skin for desktop users, after the changes described below are deployed?" —Ganesha811 (talk) 05:41, 6 September 2022 (UTC)

Well this seems to be less of an RFC on it being default and more a request for delay. If the community wants to wait longer for this to change as the default or not But it is going to happen or something along those lines. Which I can see going down well... Terasail^[✉️] 12:38, 6 September 2022 (UTC)

Nutshell/Lead

There's a tricky balance here between giving all the necessary information and potentially overwhelming participants. It seems like some nutshell or summary (what would go in an article's lead) would be helpful. That is give people a paragraph length understanding of what this RfC is, what has changed, and the process to get there. Best, Barkeep49 (talk) 15:49, 5 September 2022 (UTC)

Comments from Enterprisey

In order, as I read through the page. Permalink to the version I read.

• "Vector (2022)" feels to me like it makes sentences weirder. Could "Vector 2022", without the parentheses, be used?
• Lead: The TLDR reads like you're asking what are the issues, when from the options given it looks like it's asking whether the deploy can happen. Suggest "Is this the full set of issues", "Does anything else block the deployment", etc etc.
• Lead, final paragraph: Unsure, from just this paragraph, what would be discussed in "a subsequent RfC".
• Section 1.1, paragraph 1: "welcoming" is loaded, recommend different word choice, perhaps "useful" or "simple" or something functional?
• Overall comment on Section 1: Given what the information is, this is the best way to present it, I think. Well done.
• Section 2.1: I would turn the dial closer to talking about the functional improvements and new features enabled by this change, such as the sticky TOC. I see they're in the "New Features" collapsed section. I would suggest summarizing the ones editors have been most enthusiastic about in a new paragraph. Maybe in addition to, or instead of, the current third paragraph.
• Section 3.2: I feel like there's an implication here that the default would be to not support the skin after deployment. I recommend changing the wording to prevent that reading. Perhaps "Next steps", "Further work", "More new features", etc etc.

Thanks for the work on this RfC so far. Enterprisey (talk!) 22:38, 6 September 2022 (UTC)

Outcomes

The RfC lists two possible outcomes – deploy now or deploy later. Should these be the only options, without a "don't deploy" option? If so, can we also have an RfC on whether the WMF will grant me $1m now or later? Certes (talk) 09:47, 7 September 2022 (UTC)

I don't think that is the intent - they also say However, we will not continue with this change without the explicit approval of the English Wikipedia community. However, it was my initial impression, and it will be the impression of most Wikipedians and that impression alone will be enough to turn the RfC into a train wreck and remove any chance of this gaining consensus.

If I am correct about the intent then I would suggest the drafters modify the second option to make it clear that it is not just a delay but a rejection, one that may change but will not be ignored.

I would suggest simplifying it down to something like:

Do you support making the Vector 2022 skin the default on enwiki?

Followed by two sections; "Support" and "Oppose". "Oppose" would be immediately followed with the note: "For editors opposing, if there are changes to the skin that would make it more likely that you would support its implementation in a future RfC, the team responsible for creating the skin would appreciate you detailing them".

BilledMammal (talk) 03:18, 9 September 2022 (UTC)

I'm strongly in support of this change - the current phrasing is confusing and makes it unclear what the possible results of the RfC would be. —Ganesha811 (talk) 13:32, 12 September 2022 (UTC)

I support this rephrasing. The current wording is either unclear or presupposes adpotion, with the community choosing only the timing, neither of which seems desirable. Certes (talk) 14:11, 12 September 2022 (UTC)

Yes, this change makes sense. Simple question without the implication that adoption is a fait accompli, and the community would only have a chance to determine some design details/timing. Femke (talk) 16:06, 12 September 2022 (UTC)

@Femke, thank you for the comments you've shared about Vector 2022. I appreciate that this is all constructive and friendly feedback. As I can see, just like me, you help people make informed decisions, which is why I hope you'll help with the following dilemma.

I'd like to address the part "the community would only have a chance to determine some design details/timing". We have provided a lot of details to make it clear that over the past three years, many communities and hundreds of their members (incl. English Wikipedians) have contributed to the current state of Vector 2022. If it seems that now, the enwiki community can only determine some details, then I'd be most grateful if you had any ideas on how to fix that. I understand that this is a tough nut to crack, and if you don't know any solutions, that's fine. I just want to make it emphasized that Vector 2022 is a product of collaboration with the community already. SGrabarczuk (WMF) (talk) 22:54, 16 September 2022 (UTC)

I think with the current wording this issue is solved. Just to clarify, I was playing devils advocate here. You know as well as me how delicate the WMF/enwiki relationship is, and I just wanted to make sure you wouldn't shoot yourself in the foot. Femke (talk) 09:01, 17 September 2022 (UTC)

Sure @Femke. Thanks. SGrabarczuk (WMF) (talk) 18:02, 18 September 2022 (UTC)

@SGrabarczuk (WMF): I see you have updated the draft, but these issues remain; are you planning to address them? In addition, I note WP:RFCNEUTRAL; you need to make sure your statement is neutral, rather than advocating for the change. Much of the text here might be better suited to be on a different page from the discussion itself. BilledMammal (talk) 03:42, 15 September 2022 (UTC)

Thanks @BilledMammal. We reframed the question so that it's now a clear yes/no question, to make sure that the community knows we will not be deploying without consensus. Is that still not coming through? Do you think potentially changing the language to support/oppose would help? OVasileva (WMF) (talk) 12:00, 16 September 2022 (UTC)

I would remove "at this time", as that doesn't help editors understand what is happening, and can give the implication that a "no" answer would be a delay, rather than a rejection. I would also remove "The following additional tasks must be completed before another RfC is started." and "Note: please do not state that the skin is not ready for deployment without adding context on what additional work is necessary", as editors in opposition need to be able to reject the skin without preconditions on that rejection in order for the RfC to be neutral.

This is related to the issue Certes raises, where they point out that this doesn't currently comply with the requirements for an RfC. BilledMammal (talk) 00:49, 19 September 2022 (UTC)

Update

Hey everyone (in particular: @Barkeep49, Enterprisey, Femke, Fgnievinski, Izno, L235, Pelagic, Sdkb, Shyamal, Sj, Terasail, and Xaosflux). Thanks for your feedback on the talk page, Village Pump, and the first RfC draft. Regarding the RfC, we went through the draft and made some changes based on the suggestions. We made the following changes:

• Clarify the question we were asking: "Can the Vector 2022 skin be deployed as the default to English Wikipedia at this time?"
• Make the language clearer and as neutral as possible
• Remove duplication of information
• Make the changes in the skin and the results clear and easy to understand

Do you think the changes we made help with the points above? Are there any other changes you would recommend? We plan on launching the RfC next Tuesday, September 20. Thanks! SGrabarczuk (WMF) (talk) 20:50, 16 September 2022 (UTC)

@SGrabarczuk (WMF) regarding the last point, a couple things that may help: (a) be very clear this is not for mobile view; (b) the "try it out" is overly complex for non-technical rfc contributors. I suggest a link to an article (a FA about a fairly non-controversial place, perhaps Marojejy National Park) that has the ?useskin populated; and directions to also click on it in "private/incognito" mode to see it from the perspective of a reader. — xaosflux ^Talk 21:08, 16 September 2022 (UTC)

For me, the amount of information is a bit overwhelming. I would consider copying the background section into a separate page. The truly important link to trying it out, and the planned changes before/after deployment can stand out more if that's done.

To better comply with WP:RFCNEUTRAL, you could perhaps include a heading with 'views from the webteam', and move less neutral information there. Alternatively, a small rewrite is needed.

• There are a few statements that aren't quite neutral
  • Though no software can ever be perfect, we believe that the new skin is a big improvement for readers on desktop already. We want them to start benefiting, even as we strive to make the skin better into the future.
  • so that more people will love reading Wikipedia (also a bit repetitive with first sentence)
  • bolded 1 billion pageviews
• The new skin does not remove any functionality currently available on the Vector skin. -> not quite true, right? The phab to unbreak the TOC limit template is scheduled for after deployment phab:T317818.

Good luck! Femke (talk) 08:56, 17 September 2022 (UTC)

A timely reminder from the wider world: giving people a choice between two forms of change, neither of which is wanted, can be unpopular. Certes (talk) 10:45, 18 September 2022 (UTC)

Avoid dramatic change; be considerate of existing workflows

The more dramatic the potential change at the end of the RFC, the more particular people will be about getting details right 'in advance'.

Thinking mainly about how this sentence is likely to be implemented...

"Once these tasks are completed, the team will set a deployment date, inform the community and the readers, and continue with the deployment process. The skin would be turned on for all logged-out users, and also all logged-in users who currently use Vector legacy (2010), the current default for desktop."

...the deployment sequence seems likely to provoke backlash, unless attention is paid to making the transition smooth, making switching in/out painless, and not automatically disrupting the workflows of superusers.

A crude example of a smoother sequence after a successful RFC:

1. Run Update the current banner language, telling people V-22 is being rolled out over the coming months, and inviting them to become early adopters and switch to it. For both logged-in and logged-out people.
   • While logged-in: this should include an interface at least as good as the French-WP approach: ideally a persistent one-click link in the sidebar to switch to/switch back away from V-22 [which tracks what your previous skin was to return you to that when you switch away]
   • While logged-out: for many this will be the first time they're aware of the pending change. This could set a session cookie or take them to a login page explaining that they should log in to customize their skin.
2. Turn on V-22 for 0.1% of logged-out sessions. Show messages welcoming them to the new skin and inviting feedback.
3. Run banners telling logged-in users using Vector legacy that those using "Vector legacy (default)" may soon be switched to a new "Default (Vector 2022)" skin pref, which is currently set to V-22 and will update as the default skin changes in the future. And those who don't want this can set their skins preemptively to a new "Vector legacy" option (no "default").
4. Deploy to ~everyone else
   • Turn on V-22 for all logged-out users and 10% of "Vector legacy (default)" users with a) fewer than 1000 edits and b) no custom JS.
   • Turn on V-22 for all "Vector legacy (default)" users with < 1000 edits + no custom JS

   You don't need to update the prefs of highly active editors, they regularly see the site logged-out and will know about the change; and are most likely to have custom workflows tied to current layouts and to be annoyed by being changed agains their will. Instead migrate them to the new non-default "Vector legacy" pref.

The above would require

• A new one-click button for skin-switching integrated temporarily into V-legacy and V-22 [useful for future changes too]
• Two new skin prefs ("Default" and non-default "Vector legacy") [useful for future changes]
• A longer series of banner communications [reaching many more people than these RFCs]

– SJ + 14:17, 19 September 2022 (UTC)

I'm not sure if we need that many banners. That seems like an overcomplication, especially if skin switching is made easier during the transition period. Femke (talk) 16:40, 19 September 2022 (UTC)

Hi Femke, agreed it shouldn't be complicated, just unambiguous. In the suggestion above, I meant just one new banner (for the logged-in Vector legacy users who are being auto-migrated without any action on their part). The first item above would just be a clarification of the language in the currently-running banner so people know the change is definitely happening. I believe some sort of message was already planned for display the first time people experience the new skin, so that wouldn't be new. – SJ + 16:18, 21 September 2022 (UTC)

Hi @Sj, thank you for your feedback! I will expand the section on the deployment cadence and add a note that we will be staging the rollout. This will be done on a shorter timeline than suggested here with logged-in users first, then followed by logged-out users, but we will want to make sure that as many as possible have a heads up that the change is coming as well as making sure that the deployment goes smoothly. In terms of staged rollouts for logged-out users - this is unfortunately technically impossible at this time (actually for the same reasons that were outlined on the village pump). It would require us to serve two different versions of the site. OVasileva (WMF) (talk) 09:28, 21 September 2022 (UTC)

Thanks Olga -- do you mean "logged-out" users in the latter sentence above? That seems like it could be useful for any rollout for the same reason that we cache in the first place: because those are the vast majority of users ;) Is it not challenging to have all rollout issues for logged-out users landing on the same day? Of course there is a cost for expanding (doubling?) the # of cached pages during a rollout, but I would hope that that translates to a decision on how long to expand the cache to support two configs, rather than deciding that "how long" should be "not at all". But perhaps you're saying current caching is not scalable in that way. – SJ + 16:18, 21 September 2022 (UTC)

@Sj - good catch, I did mean logged-out :) In terms of fragmenting the cache - you're right, unfortunately the current infrastructure is not capable of handling this. I agree that it would be immensely useful though for a variety of things we've wanted to do for a long time - better A/B testing for logged-out users, settings for logged-out users (such as dark mode) and more. That said, I don't think it would make sense for this particular rollout - for logged-out users it would mean that their experience changes on each session or pageview, which would be confusing and decrease usability overall. OVasileva (WMF) (talk) 19:30, 21 September 2022 (UTC)

Ok, thank you. ! I appreciated this recap, is it still being updated? I'd love details on where splitting is used for A/B testing now + its cost, and whether there's a plan for client-side testing in the future. Many designers / users of mediawiki would love that too :)

I realize the current deployment is settled, but commenting here b/c I hope every time we update things for readers – this skin glow-up being a fine example – the needed infra changes get moved up the priority stack. [To your last point: I thought there were fuzzy privacy-preserving solutions for persistent assignment of small groups of logged-out readers? At least good enough for smoke testing. But I may be misremembering / don't recall where this was hashed out.] – SJ + 22:50, 21 September 2022 (UTC)

Not an RfC

As a piece of sales literature, this page is ready to go. As an RfC in the sense of WP:RFC, it's completely on the wrong track. The opening statement is neither brief nor neutral. Advocacy such as more welcoming and comfortable, improve readability and usability, easier reading experience and these changes are improvements belongs in the discussion below as a signed opinion, not in the question as fact. Indeed, the bulk of the text should appear either in the discussion section or (due to its size) in a separate position document. Finally, seeding the No section with Note: please do not state that the skin is not ready for deployment without adding context on what additional work is necessary reads like an attempt to forbid responses such as "No, we shouldn't be changing this". Certes (talk) 00:00, 19 September 2022 (UTC)

To comply with the requirements for the RfC, I would suggest the format detailed at #Alternative format. BilledMammal (talk) 00:58, 19 September 2022 (UTC)

Hello @BilledMammal. Thanks for bringing up the issue of the neutrality. We've made some small changes that will hopefully help. Mainly - we've marked the team's opinion in its own section, and changed some of the phrasing to indicate that the improvements referred to are not the opinions of the team, but rather the results of the data itself. We would like to respect the format, and at the same time, help people make informed decisions based on hard data which we think we're providing in the description of the skin. Do you think this approach helped? If not, do you think there may be a way of presenting data and facts in a way compliant with the format, but also, not giving impression that these are our view? Would you like to point at any additional lines you think look too much like opinions or advocating? SGrabarczuk (WMF) (talk) 12:16, 19 September 2022 (UTC)

A brief introduction to an RfC is typically one paragraph, sometimes up to a few paragraph. A long introduction would be 500 words, this page is over 4,000. I'd like to resuggest that the background section is moved to a separate page, and prominently linked (bolded, with other bolding removed).

For neutrality, write the intro so that it's unclear what the web team thinks. The current introduction is clearly written by the webteam. Normally, the view of an intiator of an RfC puts their opinion in the first !vote, but given you will not participate in the !voting, I think it's fine to keep the 'views from the webteam' section like it is now.

This is an RfC written by the Web team at the Wikimedia Foundation, with help from a number of community members, after several weeks of preparation and discussion at the Village Pump.
TLDR
Should the Vector 2022 skin be deployed as the default to English Wikipedia on desktop at this time (pending completion of tasks already agreed upon by the community)?
Currently, the Vector 2022 skin is the second most-popular non-default skin on English Wikipedia. It is also in use as the default across over 20 community-led projects accounting for more than 1 billion pageviews per month, such as French Wikipedia, where it has been the default since 2020.

Current usage of non-default skins on English Wikipedia as of August 30

The objective for the Vector 2022 skin is to make the interface more welcoming and comfortable for readers and useful for advanced users on desktop. To do this, the skin introduces changes to the navigation and layout of the site, adds persistent elements such as a sticky header and Table of Contents to make frequently-used actions easier to access, and makes some changes to the overall styling of the page. The analysis of the data collected concluded that these changes improve readability and usability, and save time currently spent in scrolling, searching, and navigating – all of which can be interpreted to create an easier reading experience. The new skin does not remove any functionality currently available on the Vector skin.
Below, you'll find more details on how the skin was built over the past three years, what has been changed, and why these changes are considered to be improvements based on the completed research and data colelcted.
If there's consensus for deploying Vector 2022, the skin would be turned on for all logged-out users, and also all logged-in users who currently use Vector legacy (2010). Logged-in users can at any time switch to any other available skin. If the community decides against deploying the skin, no deployment will be made. The Web team will review the comments, propose further changes based on the feedback, and begin another RfC for discussion once the necessary changes are agreed upon.

This can be written neutrally and brief as:

This is an RfC to decide on the deployment of the Vector (2022) skin, following from the discussion at the Village Pump.

Should the Vector 2022 skin be deployed as the default to English Wikipedia on desktop at this time (pending completion of tasks already agreed upon by the community)?

The skin introduces changes to the navigation and layout of the site, adds persistent elements such as a sticky header and Table of Contents, and makes changes to the overall styling of the page. If there's consensus for deploying Vector 2022, the skin would be turned on for all logged-out users, and also all logged-in users who currently use Vector legacy (2010). Logged-in users can at any time switch to any other available skin. If the community decides against deploying the skin, no deployment will be made. The Web team will review the comments, propose further changes based on the feedback, and begin another RfC for discussion once the necessary changes are agreed upon.

To preview what the skin looks like, go to this article. The background and data analysis can be found on a separate page. Femke (talk) 16:37, 19 September 2022 (UTC)

@SGrabarczuk (WMF): In regards to improvements referred to are not the opinions of the team, but rather the results of the data itself, the issue is that you are presenting an interpretation of a result, and the presentation and the interpretation can be non-neutral. For example, you say Overall, the majority of logged-out users reported that they would view the changes either positively or neutrally. This is true, but is also not neutral; there are many other ways to present this data, depending on your point of view, such as Overall, the majority of logged-out users reported that they would view the changes either negatively or neutrally and Overall, a plurality of logged-out users reported that they would view the changes negatively.

To fix this, I agree with Femke that either you need to pull this summary out and present it as your interpretation of the data, or you need to replace it with the hard data and allow editors to interpret that data.

I would also emphasise the issue with the options; for this to be neutral it needs to be clear that editors can reject the change without detailing changes that need to be made. BilledMammal (talk) 00:44, 20 September 2022 (UTC)

I agree with BilledMammal that any sort of pitch (including the WMF's characterizations of shiny new features) should occur separately from the actual RfC prompt. The alternative format below would allow for a brief and neutral RfC. The originally proposed version is neither neutral nor brief and would rightfully earn a multitude of BADRFC !votes if launched in any format substantially similar to its original state. — Red-tailed hawk _(nest) 02:50, 21 September 2022 (UTC)

@Certes, @BilledMammal, @Femke - thank you all for your feedback, it has been super helpful. We've made a number of changes based on what you said. Mainly, we moved to a new format that is a mix between the alternative format suggested by BilledMammal below and Femke's suggestion above:

• The background, full data and results, list of features, process, and future sections are now hosted on a separate page linked from the main RfC page
• Opinions of the web team and responses to common questions are now in the discussion section
• We've removed all qualitative data from the short results list and are reporting the raw numbers only
• We've made it clearer (now bolded) that no deployment will be made if the community decides against the change

Thanks again! We didn't open today so we could make the changes, but plan on starting tomorrow. OVasileva (WMF) (talk) 19:18, 21 September 2022 (UTC)

Thanks. I think an rfc is appropriate for determining an appropriate decision for the redesign at this time. I know if it is not deployed right now it does not mean it will never be deployed. I am a happy user of the timeless skin which is mobile responsive and is very modern and accessible by newer standards. I just hope that the vector 2022 skin does not become like another Fandom desktop where it is only optimized for one mode only. Aasim - Herrscher of Wikis ❄️ 23:51, 21 September 2022 (UTC)

@OVasileva (WMF): Thank you for making those changes; there are a few more I would suggest, in #Alternative format two. The most important change is to fully move the results out, as they still don't comply with the neutrality requirement. This is both due to how you interpret the data with some of your interpretations being debatable, as can be seen by our extensive discussions on the sticky header, and due to the data you chose to present, with you excluding data that would be seen as negative, such as the survey of non-logged in editors.

The other changes aren't as important, but I believe they should be done in order to maximize participation and ease of use; these remove the introduction section, and move the discussion section below the survey section. The introduction section doesn't add anything to the discussion, and will be above the point where most editors arrive to the page. Removing it will allow editors to arrive above the table of contents for ease of navigation, as well as shortening the page. Moving the discussion section below the survey section is both better aligned with convention, and will make it easier for editors to reach the survey section as discussion sections often end up being disruptively long.

I would also suggest not launching the RfC tomorrow, but delaying until you have a solution to the coordinate problem, otherwise you will have editors opposing due the possibility that the solution won't be appropriate. BilledMammal (talk) 00:19, 22 September 2022 (UTC)

Thanks @BilledMammal for working through all this with us. We just made another round of smaller changes. I do think that the data is objectively crucial information for the RfC and should be included in the main text. That said, I agree that there could be a bit more done to focus on the numbers themselves and try to avoid conclusions. We've done another pass at the language to make sure that we're avoiding that. I also added a notice at the top to clearly mark that the section was compiled by the team for context.

In terms of the order - we had a couple of conversations early on in the process to highlight the background of the RfC and previous conversations with the community within the RfC, so we want to make sure that's still accounted for. We're going back and forth about the discussion versus voting coming first as well, but in the spirit of this being a Request for Comment, I think we're still leaning towards having the discussion and conversation first rather than the pure voting part of the RfC. OVasileva (WMF) (talk) 12:23, 22 September 2022 (UTC)

I concur with Certes. The launch of this RfC has been botched. For the record; I've never been a fan of using icons for menu items. I can't read icons. It's mystery meat navigation. We have a little bit of that with monobook. Vector makes that worse. Trying to solve some problems by introducing other problems isn't a way forward. --Hammersoft (talk) 23:13, 22 September 2022 (UTC)

Alternative format

Should the Vector 2022 skin be deployed as the default to English Wikipedia on desktop, pending the completion of a solution for coordinate alignment?

Support

1. Andre🚐 23:41, 21 September 2022 (UTC)

Oppose

For editors opposing, if there are changes to the skin that would make it more likely that you would support its implementation in a future RfC, the team responsible for creating the skin would appreciate you detailing them

1. Oppose I've never been a fan of using icons for menu items. I can't read icons. It's mystery meat navigation. We have a little bit of that with monobook. Vector makes that worse. Trying to solve some problems by introducing other problems isn't a way forward. --Hammersoft (talk) 23:14, 22 September 2022 (UTC)
2. Oppose - I showed the new view (on a HD desktop screen at 100% scaling) to my family a couple of months ago and they were all convinced that I was showing them a mock-up of Wikipedia with ads. Very few changes have been made to Vector 2022 since its initial deployment in 2020, and the way in which the WMF employees seem to consider this a rubber stamp for a done deal makes me sceptical that any problems brought up will be addressed to the satisfaction of the community. Daß Wölf 01:57, 23 September 2022 (UTC)

   @Andrevan, @Hammersoft and @Daß Wölf: this is the talk page of the RfC. You probably meant to comment on the RfC itself. Femke (talk) 02:00, 23 September 2022 (UTC)

   Yes, you are correct. I commented here before the RFC was opened, but I have since commented on the real RFC. Andre🚐 02:05, 23 September 2022 (UTC)

   @Femke: Thanks, I found my way there eventually. The opening statement proved hard to navigate and I was unsure where the RfC discussion was or if it existed yet, but then I saw your comment below at #RFC opening when? and decided this could be it ;) Daß Wölf 03:00, 23 September 2022 (UTC)

Webteams View

Put the content from Wikipedia:Requests for comment/Deployment of Vector (2022) here.

Alternative format two

Should the Vector 2022 skin be deployed as the default to English Wikipedia on desktop at this time (pending completion of tasks already agreed upon by the community)? OVasileva (WMF), SGrabarczuk (WMF) (talk) 15:02, 21 September 2022 (UTC)

The skin introduces changes to the navigation and layout of the site, adds persistent elements such as a sticky header and Table of Contents, and makes changes to the overall styling of the page. Currently, the skin is the default on more than 30 projects of various sizes, accounting for a bit more than 1 billion pageviews per month. To preview what the skin looks like, go to this article.

• Top of an article
• 
  Vector 2022
• 
  Vector legacy (current default)

• A section of an article
• 
  Vector 2022
• 
  Vector legacy (current default)

Analysis by the WMF

The background and details on the data analysis can be found on a separate page. Note: we're doing this to keep the opening statement as neutral as possible and to shorten the length and information, as per the recommendation of the community.

Next steps

If there's consensus for deploying Vector 2022, the skin would be turned on for all logged-out users, and also all logged-in users who currently use Vector legacy (2010) in a deployment with multiple stages to ensure sufficient time for testing. Logged-in users can at any time switch to any other available skin.

If the community decides against deploying the skin, no deployment will be made. The Web team will review the comments, propose further changes based on the feedback, and begin another RfC once the necessary changes are agreed upon.

Discussion

Voting discussion

Support

Yes, the Vector 2022 skin can be deployed. I may opt-out but I don't mind it becoming the default.

• --~~~~

Oppose

No, the Vector 2022 skin cannot be deployed. The following additional tasks must be completed before another RfC is started.

• --~~~~

Rationale discussion

Responses to common questions from the Web team

Why should we make the change now?

Though no interface can ever be perfect, we believe that the new skin is a big improvement for readers on desktop already. We want them to start benefiting, even as we strive to make the skin better into the future. We believe this change will be crucial to making the contents of the project more readable, the projects' interfaces more welcoming to less-technical contributors, and thus, to the overall growth of new readers and editors.

Why are you sure the new skin is an improvement over the old skin?

The results and analysis of A/B testing and qualitative testing confirmed the team's initial hypothesis that these changes make it easier to read and learn, navigate within the page, search, switch between languages, use page and user tools, and more, without negative effects to pageviews, account creation, or edit rates when compared to the Vector skin. The team has been working on the new skin for the past three years, ensuring that every change is tested and proven to work.

The current skin is good enough for me; why do we need to change?

The current skin, Vector, has been in use since 2010. When it was developed, it reflected the needs of the readers and editors of the Wikimedia sites in that year. (See the Wikimedia Usability Initiative wiki for more information.) Since then, vast new audiences have begun using the Internet and Wikimedia projects. Research done with these audiences showed that the current default skin doesn't meet their needs. The Vector 2022 skin aims to change the interface in ways which include the needs of all of the current audiences – both those who have been using the projects for a long time, as well as those who have joined more recently, or have yet to join.

What if I don't like a particular feature in the skin?

It is possible to configure and personalize the changes. The Web team offers support for volunteers with technical skills who would like to create new gadgets and user scripts. So far, many gadgets and user scripts have been built by community developers that customize different aspects of the new skin, including restoring full width, disabling sticky elements, restoring the old table of contents, and more. Check out the repository for a list of currently available customizations, or to add your own.

Can you just tell me how do I opt out from it? Do I need to do that if I'm using Monobook or Timeless?

If you're using the current default, go to your preferences and select Vector legacy (2010). You may also opt-out across all the wikis using global preferences. If you're using Monobook or Timeless, you will not notice the change.

What changes does the new skin bring?

The skin includes changes to the layout of the site, location and prominence of some features, the overall readability, and addition of sticky features. This improves the overall readability and usability of the site. Among the best-received by the communities, there are the new Table of Contents, sticky header, and the search widget. No existing features or tools were removed as a result of the new skin.

Will you support the skin in the future?

This is not a one-shot project, and we will continue working on the Vector 2022 skin. First, we will be working on the page tools feature, to be completed in October/November 2022. Then, we will collaborate with the Growth and Editing teams on making it easier to learn about how the wikis work and begin editing. For more details, see the sub-page.

Apart from that, we strongly encourage you to go to our FAQ page. OVasileva (WMF), SGrabarczuk (WMF) (talk) 15:02, 21 September 2022 (UTC)

Alternative format two discussion

When showing side-by-side comparison of Vector legacy and Vector 2022, I recommend putting old skin on the left and new one on the right. "Before" goes on the left, "After" goes on the right. —⁠andrybak (talk) 12:54, 22 September 2022 (UTC)

• The phrasing of "Voting discussion" seems a bit odd given that Consensus is ascertained by the quality of the arguments given on the various sides of an issue, as viewed through the lens of Wikipedia policy rather than being the result of a vote. Alternative format I is much simpler and it also doesn't include the nebulous language of pending completion of tasks already agreed upon by the community that will likely draw opposition about how hand-wavy it is. — Red-tailed hawk _(nest) 13:54, 22 September 2022 (UTC)

  Thanks for the feedback @Red-tailed hawk. We'll try to clarify the support/oppose section a bit further. In terms of the heading, I agree that "voting" is perhaps a misnomer here (we were inspired by the structure of this RfC, but agree the heading can be clearer. OVasileva (WMF) (talk) 14:04, 22 September 2022 (UTC)

Additionally, the prompt given in the "support" and "oppose" sections to railroad responses are simply not acceptable from an RfC perspective; editors are free to give whatever opinion that they desire without push polling prompts and for that reason is much more neutral. Alternative format 1 seems to be much better in that respect. — Red-tailed hawk _(nest) 13:58, 22 September 2022 (UTC)

Thanks @Andrybak - just did that! OVasileva (WMF) (talk) 13:58, 22 September 2022 (UTC)

Skin preview link

Just as a thought, would it not make more sense for the preview link to go to pluto? Since that is the article shown in the screenshots. It also has an infobox where the galaxy link doesn't. Terasail^[✉️] 11:32, 22 September 2022 (UTC)

@Terasail That's a good idea, thanks! OVasileva (WMF) (talk) 11:48, 22 September 2022 (UTC)

Just noting that Pluto with the skin currently has a formatting error (see the table here). Anarchyte (talk) 16:27, 22 September 2022 (UTC)

@Anarchyte I just added a change which should fix this. Terasail^[✉️] 17:43, 22 September 2022 (UTC)

Watchlist notification

This is quite an important change. I don't think CENT will be enough. Has there been any discussion about putting a watchlist notification for this RfC? Anarchyte (talk) 16:29, 22 September 2022 (UTC)

@Anarchyte, yes! MediaWiki talk:Watchlist-messages#RfC on Vector 2022. SGrabarczuk (WMF) (talk) 16:35, 22 September 2022 (UTC)

RFC opening when?

Hi, do you know when this RfC will be open? I think 1 October is a good deadline, but I do agree with the comments suggesting slight modifications to the RfC. Thanks. Aasim - Herrscher of Wikis ❄️ 19:32, 22 September 2022 (UTC)

It has opened already :). Femke (talk) 19:43, 22 September 2022 (UTC)

Why did it open, when there was still open discussion about the format above? In particular there was concern over the neutral phrasing of the question, given the "propaganda" of the web team's faqs above the voting section.  — Amakuru (talk) 21:38, 22 September 2022 (UTC)

While I agree the neutrality issues were not fully addressed, I don't think loaded language like 'propaganda' is appropriate here. I very much believe WMF staff should be granted the same friendliness as other editors new to the enwiki community. They've taken most of the feedback on board. Femke (talk) 10:43, 25 September 2022 (UTC)

Cookie preference

If ?useskin=vector-2022 is to be the new default, is there some reason a cookie couldn't be used to allow readers who are not logged in to switch to what ?useskin=vector-2022&withgadget=wide-vector-2022 or ?useskin=vector currently show? Johnuniq (talk) 04:26, 23 September 2022 (UTC)

Yes: please do this. I can use Tampermonkey to reload with ?useskin whenever I visit Wikipedia logged out, and I expect add-ons to reverse the change in popular browsers to appear very quickly, but a simple cookie and a prominent way of setting it would be more accessible for the non-technical user. Certes (talk) 11:01, 23 September 2022 (UTC)

That would significantly reduce the efficiency of pageloads for those anonymous users, or require doubling of the cache for anonymous requests (if not using redirects). —TheDJ (talk • contribs) 14:19, 23 September 2022 (UTC)

Table of contents visibility

I generally like the changes and there seems to be a warm reception from others for the new TOC. One clear drawback with the new TOC is I don't see it when I visit an article until I scroll down. Is this just me having too much junk loaded up in my sidebar? The TOC is often one of the first things I read when visiting a new article; not being able to see it immediately is a bummer. ~Kvng (talk) 19:59, 23 September 2022 (UTC)

Future (October/November) work will likely address this: the page tools are planned to be moved to the right side of the article, into the big void. I believe most experienced editors will collapse the remaining links by default, so that the TOC will be visible without the annoying scrolling one has to do now.

It may take a bit for user scripts with additional links to be adjusted probably, but this should be resolved relatively soon. Femke (talk) 07:44, 24 September 2022 (UTC)

Thanks. The plan to move to the big void assumes the big void survives; I see many here don't like it. I just discovered the ability to hide the tools. I'm an experienced user and the fact that you had to tell me this was possible indicates there's potentially a usability problem. I also got a surprise when I hid the TOC and it was not immediately clear how to get them back.

I've looked at the video demonstration and I see there are more options proposed than the big void. Some of this makes a lot of sense. It's not included in the current layout though so what still doesn't make sense is why we would deploy a layout where you don't see the TOC when you first visit a page. ~Kvng (talk) 14:59, 24 September 2022 (UTC)

Thanks for the feedback @Kvng. In terms of the timeline, we're currently working on the article tools work and hope to have it ready in October/early November. As you mentioned, this will make the main menu much shorter and allow for the ToC to appear on pageload and also allow for collapsing of the main menu for logged-in and logged-out users. OVasileva (WMF) (talk) 10:50, 27 September 2022 (UTC)

Awareness of alt proposal

I have started an alternative proposal in its own section. It seems to have gone mostly overlooked. I plan to ping a random subset of 1/10th of the current voters (support, oppose, and neutral) to get people to look at it. I will do this in 24 hours unless anyone objects. (Or maybe 48 hours if I stop feeling antsy.) Enterprisey (talk!) 18:54, 26 September 2022 (UTC)

I don't think pinging people makes much sense. RFCs remain discussions and fire-and-forget is almost always not appreciated. Izno (talk) 19:12, 26 September 2022 (UTC)

I doubt that will go over very well honestly. I think a followup RFC at a later time is going to be more sustainable. —TheDJ (talk • contribs) 11:29, 27 September 2022 (UTC)

My suggestion is not to ping any editors over this. This RfC is a mechanism for gathering feedback rather than one making a binding decision, since ultimately the WMF controls design decisions. I feel certain they have seen your suggestion and can take it under consideration. isaacl (talk) 02:58, 28 September 2022 (UTC)

Roger. I'll inquire with them about launching a followup at some point. Enterprisey (talk!) 06:14, 28 September 2022 (UTC)

Vector 2022 mobile view

I'm trying to test Vector 2022 in mobile view as displayed here, but can't figure out how to enable it. I have Vector 2022 set as default in desktop view, but when I switch to mobile view, I get the standard mobile skin (MinervaNeue, I believe?). When going to user preferences in mobile view, Vector 2022 is shown as being enabled, but the skin is clearly MinervaNeue. Throast ^{{{ping}} me!} (talk | contribs) 01:02, 28 September 2022 (UTC)

@Throast: MobileFrontEnd is not designed to support desktop skins. You can forcibly see what it comes up with by using a parameter such as in this link. That view is not supported, and is certainly not part of this RFC, the scope of which is only for users on desktop. — xaosflux ^Talk 01:28, 28 September 2022 (UTC)

When closing this RfC...

Premature, unlikely to produce further understanding. Enterprisey (talk!) 16:40, 4 October 2022 (UTC)
The following discussion has been closed. Please do not modify it.
I know I am a bit early to be saying this but I believe this is important. I have a feeling this RfC may close as "no consensus" given the mixed opinions and solutions and problems that Vector 2022 provides. I think when closing this there needs to be a focus on the needs of the WP:READER more than anything else. Obviously arguments that are just "I like this, I don't like this" should be disregarded and arguments that point out specific problems that make the skin difficult for readers and new editors to use should be disregarded also. It is tempting to count the number of votes but that defeats the purpose of consensus. Votes that are just plain "support" or "oppose" or just very weak arguments (like "I (don't) like X" without pointing out problems X can cause or the solutions X solves) can be thrown out as well. It is also important that WMF can (but not necessarily will) veto override our consensus per Wikipedia:OFFICEACTION, although that has previously resulted in controversy. I think the close needs to be detailed explaining the specific solutions the skin solves that editors have successfully identified as well as the specific problems the skin creates or does not solve that the editors have also identified. I think a "no consensus" should be interpreted best as "maybe give it a try for some logged out users but not all logged out users" and consensus to not deploy is best interpreted as "needs more work and community input". Consensus to deploy of course would mean Vector 2022 would immediately become the default. It will be difficult to close this technical RfC because even the closing admin or staff will have some opinion about the skin. ::Ningguang_Sighing:: I guess we just have to make do with whatever our current system offers. Aasim - Herrscher of Wikis ❄️ 03:36, 4 October 2022 (UTC) I disagree that arguments that point out specific problems that make the skin difficult for readers and new editors to use should be disregarded. Difficulty of use may be a legitimate argument against adopting a skin. Certes (talk) 12:42, 4 October 2022 (UTC) And I disagree with the first part of that statement, Obviously arguments that are just "I like this, I don't like this" should be disregarded. The question asked was, Should the Vector 2022 skin be deployed as the default to English Wikipedia on desktop at this time (pending completion of tasks already agreed upon by the community)? I do not see anything in the question that required more than a 'yes' or 'no' answer. This is not an XfD where we need to argue from policy. - Donald Albury 15:50, 4 October 2022 (UTC)

Is the RfC being publicized to non-registered users?

I noticed there are several recent support !votes from apparent SPAs, i.e., users that edited this page immediately after creating their account and haven't edited anywhere else. For reference, the users are Uxdiogenes (c), Kathleenwarner (c), Cattsmall (c), and Designseattle (c), who are currently 4 of the last 9 !votes under "Support". It certainly seems odd for so many users to create accounts just to comment here, but I don't want to just accuse them of any inappropriate behavior – maybe word about the discussion has spread and these editors are readers who created accounts to comment. I think that would be a good thing, since a key point of the RfC is to find what benefits readers the most. Just to satisfy my curiosity, is there somewhere where information has been shared with non-registered users inviting them to join the discussion? RunningTiger123 (talk) 02:38, 13 October 2022 (UTC)

Hi, I'm one of the people tagged. I've used Wikipedia before but never created an account because I'm extremely lazy and didn't need one until I wanted a cool signature like everyone else who signed. Hope that helps!! Cattsmall (talk) 03:50, 13 October 2022 (UTC)

@Cattsmall, what brought this page to your attention? — Qwerfjkltalk 20:21, 13 October 2022 (UTC)

Closing

I am writing for some feedback on whether I should act as a closer. I offered advice on the structure of the RfC (and stated that there should be an RfC) but have not offered any opinion on the change itself. I actually don't care what happens here beyond that whatever happens has the consensus of the community. So in that sense I think I exactly meet the question posed at Advice on closing discussions, The ideal closer for a discussion is one who honestly doesn’t care what the outcome is, only about the integrity of the encyclopedia and the RfC process. But obviously there are other editors who can close and so if people feel I am not the right closer I'm happy to defer to others. If people feel I am a suitable closer I'll spend some time starting to read the discussion in the hopes that it could be closed shortly after the 30 day mark (assuming that the conditions are right to close). Best, Barkeep49 (talk) 18:47, 13 October 2022 (UTC)

The optics would be poor if someone who "offered advice on the structure of the RfC" were to close the very same RfC. Besides, does the RfC actually need to be closed? It seems more like a community consultation for the WMF team to gather feedback, rather than a !vote to decide if the skin should be deployed. – SD0001 (talk) 12:28, 15 October 2022 (UTC)

There's an RfC question so I think it's very much a discussion about whether or not to deploy and definitely needs a closing. I posted here to get feedback so let me ask what about having offered advice on the structure suggests to you that I would not be able to assess the consensus reached? Best, Barkeep49 (talk) 12:41, 15 October 2022 (UTC)

I agree that the RFC veered off into "community consultation" land, but yes, there was a straightforward question at the beginning which deserves a definitive answer so there needs to be a formal close. As for who should do the close, my general rule for big complicated contentious things is "try not to touch it more than once". If you were instrumental in getting this going, better to let somebody else wrap it up. -- RoySmith (talk) 12:50, 15 October 2022 (UTC)

The optics of pretending you would respect the choice of the users and then ignoring that choice were always going to be awful. You can go ahead and have whoever you like be the person who puts their stamp on it. — LlywelynII 07:57, 19 January 2023 (UTC)

I don't see any problem with your closing. Your contributions were technical advice on RfCs in general, such as how to make the format bot-friendly, rather than opinions on the proposed change. I'd say wait for a completely uninvolved party but, if no one comes forward, close when appropriate. Although we may not be ready for it today, a closure will be important because it gives the community's answer to the question: "Should the Vector 2022 skin be deployed as the default to English Wikipedia on desktop at this time"? Certes (talk) 12:49, 15 October 2022 (UTC)

The Web team has posted a section highlighting the team's main takeaways from the discussion and the next steps the team will undertake. Has this RfC effectively been "closed" by its proposer, or are we still expecting a more conventional close by an uninvolved party? Certes (talk) 17:13, 21 October 2022 (UTC)

We wrote that our intention was to "provide a quick summary of our take on the RfC and next steps prior to closure". We didn't mean to close it ourselves. SGrabarczuk (WMF) (talk) 22:45, 21 October 2022 (UTC)

Vector 2022 update (more info on VPT)

In the technical section, the Web team has posted an update on the Vector 2022 skin based on the communities' requests and RfC closure. We will also be having an open meeting for anyone that has questions this Thursday at 19:30 UTC on Discord. Thank you! SGrabarczuk (WMF) (talk) 20:38, 12 December 2022 (UTC)

Vector 2022 deployment update

Cross-posting for a better visibility: in the technical section of the Village Pump, the Web team has posted an update on the deployment of Vector 2022. We will also be having three open meetings for anyone wanting to ask questions. Thanks! SGrabarczuk (WMF) (talk) 14:22, 11 January 2023 (UTC)

Crossposting for better visibility: The new format is godawful.

Even worse that apparently there was no consensus (with a majority opposed) and it got pushed through regardless. Now that we're stuck with it, see above for changes that needed to have been made before this went live but certainly needs to happen now. No, the numbers presented here for "only a fifth switched out" are completely meaningless if you made it this difficult to find out HOW to switch in the first place. That one in five of general users managed to figure it out for themselves speaks to a large majority of annoyance and opposition. — LlywelynII 08:00, 19 January 2023 (UTC)

Just curious

Given how awful this is for UX, what are the benefits on the backend that made forcing this through against majority opposition such a priority? — LlywelynII 08:01, 19 January 2023 (UTC)

Still curious about the above. In the meantime, as a new part of the change, the URL formatting is now broken because some genius stuck the "uglyolddumbformat" cruft between the page and section names. — LlywelynII 15:19, 29 January 2023 (UTC)

@LlywelynII, this isn't V22. — Qwerfjkltalk 22:27, 29 January 2023 (UTC)

It most certainly is. The relevant part of the URL wouldn't've even existed except for this rollout which—if you're in any way involved—again shouldn't've happened, assuming the project actually meant that it intended to solicit community feedback and planned to respect it. — LlywelynII 00:27, 30 January 2023 (UTC)

Awful close

Just because you knew the WMF would foist it upon us anyway was no reason to close this discussion as consensus to implement when it was blatantly obvious there was anything but. And another typical awful tech implementation by the WMF whose effectiveness as an organisation is almost directly inverse to the amount of donations they receive. Looks awful on desktop, the whitespace just for starters is an abomination. 2403:5802:19ED:0:6885:A0AB:DB97:DC2C (talk) 10:17, 19 January 2023 (UTC)

I've logged into my Wikipedia account for the first time in 7 months simply so I can turn the redesign off Hewkii999 (talk) 00:21, 20 January 2023 (UTC)

Wikipedia:Vector 2022 has an RFC

There is an RfC on whether Vector legacy should be restored as the default skin on the English Wikipedia. If you would like to participate in the discussion, you are invited to add your comments at Wikipedia:Village pump (proposals)#RfC: Should Wikipedia return to Vector 2010 as the default skin?. Thank you. InfiniteNexus (talk) 01:26, 20 January 2023 (UTC)

Raw results of studies?

Where can I see the raw results of the studies comparing the 2 skins? E.g. regarding influence of the text width? I can only see vague phrases like the majority” or “a significant minority” etc… A11w1ss3nd (talk) 16:48, 10 June 2023 (UTC)