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)[reply]

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 (open in a private browser window to see what it will look like for a logged-out reader).

This is an RfC written by the Web team at the Wikimedia Foundation, with help from a number of community members after several months of preparation and discussion at the Village Pump.

The WMF Web team has been working on the Vector 2022 skin for three years. Since July 2022, there has been a discussion on the Village Pump on the needs of the English Wikipedia community. Within that discussion, the community decided on the changes necessary before deployment, which the Web team is addressing. The community also recommended that the next step would be this RfC.

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.

The Web team would like to thank the many Wikipedians who have worked on this skin and given their feedback and guidance. To name just about a dozen: Barkeep49, BilledMammal, Certes, Enterprisey, Femke, Ganesha811, Izno, L235, Pelagic, Sdkb, Sj, Terasail, TheDJ, WhatamIdoing, xaosflux, and Xeno. Thank you!

This section was written by the web team to highlight the main findings from the team's analysis of A/B tests and other quantitative data. The list is not exhaustive of all research, quantitative, and qualitative findings throughout the project. The full background on the skin 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.

The analysis of the data done by the Product Analytics team concluded that these changes improve readability and usability, and save time spent in scrolling, searching, and navigating – all of which was interpreted by the team to create an easier reading experience. The new skin does not remove any functionality currently available on the Vector skin.

Results at a glance

• On average, 87% of active editors across our pilot wikis (incl. French and Portuguese Wikipedias) continue to use the new skin once they try it.
• The sticky header decreases the amount of scrolling logged-in users have to do by giving access to tools that editors use most frequently. It decreases scrolling to the top of the page by 16%.
• The new table of contents increases navigation to different sections. Readers and editors jumped between sections 50% more than with the old table of contents.
• The new search bar was built to make it easier to find the correct search result from the list. This increased the amount of searches started by 30% on the wikis where tests were performed.
• PHP code in Wikimedia deployed skins has been reduced by 75%
• The skin does not negatively affect pageviews, edit rates, or account creation. There is observational evidence of increases in pageviews and account creation across partner communities.

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)[reply]

Yes, the Vector 2022 skin can be deployed. (This section may cover different kinds of support, like "I may opt-out but I don't mind it becoming the default.")

• Support - While I still think there are issues (width being the primary one), it's about time we updated the default skin. Vector 2010 is starting to show its age. Anarchyte (talk) 16:24, 22 September 2022 (UTC)[reply]
• Strong Support - the 2022 skin is not perfect, but the 2010 skin is much farther from perfect. The English Wikipedia community is conservative by nature, but I hope we will adopt this needed change to benefit readers and editors alike. —Ganesha811 (talk) 16:28, 22 September 2022 (UTC)[reply]
• Andre🚐 16:28, 22 September 2022 (UTC)[reply]
• Support not interested in using it myself, but I do think it is likely to be more user-friendly for beginners. (t · c) buidhe 16:44, 22 September 2022 (UTC)[reply]
• Strongly Support My only skin for all my accounts on all Wikimedia wikis. Zippybonzo | Talk (he|him) 17:01, 22 September 2022 (UTC)[reply]
• Support at this time. Will keep monitoring. --SarekOfVulcan (talk) 17:17, 22 September 2022 (UTC)[reply]
• Support: I have been using it for a while now, I think that it is an improvement overall and think it would benefit logged-out readers in particular to have this as default. Terasail^[✉️] 17:37, 22 September 2022 (UTC)[reply]
• Support. It's a more simple design than the previous skin. Agusbou2015 (talk) 18:41, 22 September 2022 (UTC)[reply]
• Support; I'm using it for almost a year now on all projects, and find it clearly fit for use as default skin for all users—including particularly unregistered readers. —MisterSynergy (talk) 19:02, 22 September 2022 (UTC)[reply]
• Support - Not anywhere near perfect, but good enough. Schierbecker (talk) 19:09, 22 September 2022 (UTC)[reply]

No, the Vector 2022 skin cannot be deployed. (For editors opposing, if there are changes to the skin that you would like to see completed before a future RfC on the skin, the team responsible for creating the skin would appreciate you detailing them)

• Oppose I can't get behind the fact that we're going to make a hard limit on the width of the screen for readers. Especially for image-heavy articles, this is going to cause a host of formatting issues when combined with moving the table of contents that will take substantial time to fix content-wise. If the skin change is going to create issues with the readability of existing readable text, then it is a net negative to our encyclopedia's readers. — Red-tailed hawk _(nest) 16:50, 22 September 2022 (UTC)[reply]
• The width of the screen is a deal-breaker for me. I have opted out of Vector 2022 globally because of this. --Rschen7754 18:04, 22 September 2022 (UTC)[reply]
  • I am also concerned that the screenshots at the top of the RFC do not adequately illustrate this problem, and this is not clearly disclosed in the RFC descriptions either. --Rschen7754 18:06, 22 September 2022 (UTC)[reply]

    @Rschen7754, could you elaborate if this is more about your personal preference (which is fine, that's why there's a gadget disabling the limited width for individual users) or anything broader? Note that it's been working on 30 different wikis, so I'd really like to figure out what issue related to the limited width you're most concerned about. SGrabarczuk (WMF) (talk) 18:17, 22 September 2022 (UTC)[reply]

    • Look at pages like this - the limited ability to use the screen width means a lot more scrolling to use the table. Not to mention that for editors, the limited screen width is a significant handicap. --Rschen7754 18:25, 22 September 2022 (UTC)[reply]

      @Rschen7754, there are at least two different issues to unpack: wide tables that get narrowed down, and the editing experience. We need a bit of time to address the first issue. Regarding the second one, in the editing mode, the limited width is disabled - both in VisualEditor and the wikitext editor. That seems to be an effective solution. What do you think? SGrabarczuk (WMF) (talk) 18:50, 22 September 2022 (UTC)[reply]

      @SGrabarczuk (WMF): in the live version, the right-side whitespace gutter is about the same size as the left side gutter on my screen, but in the screen shot File:Screenshot of English Wikipedia Pluto article in Vector 2022 skin.png the right-side gutter appears to be cropped to about half of that width. — xaosflux ^Talk 18:38, 22 September 2022 (UTC)[reply]

      Compare to File:Pluto full screen vector 2022 2022-09-22.png. — xaosflux ^Talk 18:41, 22 September 2022 (UTC)[reply]

      @Xaosflux - thanks for flagging this. We didn't crop the images - this is just due to the size of my window when I was taking the screenshot - should roughly correspond to the width of the right and left sidebars at 1400px (most 13 inch screens). We can potentially add some additional images for screens of different sizes or highlight/bold the "try it out" link to encourage people clicking on that. OVasileva (WMF) (talk) 19:07, 22 September 2022 (UTC)[reply]

      @Xaosflux I believe the screenshots are at two different screen widths. Does this image help to clarify?

      

      Also, as a reminder, we've just begun work on the article tools (phab task). Once this is done, in a month or so, you will be able to pin article tools (and other gadgets) to the right-side of the article. You can view the prototype of that here: https://vector-2022.web.app/Moth.

      

      AHollender (WMF) (talk) 19:21, 22 September 2022 (UTC)[reply]

• Oppose. First, the most important thing to consider is the experience of the reader, and per a survey conducted by the WMF they find new format skin harder to use than the current skin as can be seen on the graph below. This means that deploying the skin without an easy way for non-logged in readers to revert back semi-permanently to the current skin should not be considered as an option, and as no such option currently exists we need to reject this change.

The chosen width of the page is another issue; it is important to consider how the width affects the perception of Wikipedia, as we want to be perceived as a broadsheet, not as a tabloid, and this width change is unlikely to have a positive impact on this perception. It has also resulted in a unsightly gap between the left hand navigation bar and the content, while reducing the space available for content that need to utilize large amounts of space on the page, such as larger tables, charts, and panoramas.

Elsewhere on the page, the position of coordinates and icons have been usurped. While it is possible that this decision is an improvement, these content decisions should not be made without closer discussion with enwiki, and we should not approve the implementation of this skin until we are satisfied with the new location for coordinates, or until the decision to usurp their position is reversed.

Other issues also exist. For example, there has been no investigation of what highly used scripts will be broken, and the testing of the various features has been insufficient at times, with features such as the sticky header only being tested in such a way to see whether they are in use, rather than whether they improve the user experience. BilledMammal (talk) 19:12, 22 September 2022 (UTC)[reply]

• Oppose until the developers can assure us that the fixed width/width limitation will be removed in the near future and is not deeply baked into the layout or otherwise unfixable. I accept that the new skin cannot be perfect but this violation of a fundamental usability principle does not inspire confidence so I have to know that this is a problem that can and will be addressed. ElKevbo (talk) 19:28, 22 September 2022 (UTC)[reply]

• I'd like a commitment that the WMF will provide maintenance for the core V22 customisation gadgets/scripts (width, TOC positioning, etc) if the volunteer creators are unable to do so, including ensuring that they continue to work with future V22 changes. While the mediawiki documentation speaks of wanting a more consistent experience between readers and editors, I simply believe that's not going to happen. Our uses are just so different. So ultimately V22 needs to be able to handle both - and that agreement needs to be provided prior to becoming the status quo for readers because it'll be almost impossible to get afterwards. Nosebagbear (talk) 17:45, 22 September 2022 (UTC)[reply]
• One of the main issues I worry about, ironically, is accessibility of the link colours. These were changed to improve accessibility. I've changed the visited link colour in my CSS because I struggled to read them, especially on my watchlist. Given I have good eyesight, I worry about this being a wider issue. The colour can be made a tad bit more dark without failing the WCAG AA accessibility with the black prose (see phab:T213778), but that colour is still too light for me. If this is a problem more people experience, that would be a reason to oppose for me. Femke (talk) 18:50, 22 September 2022 (UTC)[reply]