Wikidata:Requests for comment/Talk pages consultation 2019

From Wikidata
Jump to navigation Jump to search


THIS RFC IS CLOSED. Please do NOT vote nor add comments.

On 21 February 2019, the WMF informed various Wikimedia projects of the 2019 talk pages consultation.

The Talk pages consultation is a global consultation planned from February to June 2019, to bring Wikimedians and wiki-minded people together to define better tools for wiki communication. The consultation will seek input from as many different parts of the Wikimedia community as possible – on multiple projects, in multiple languages, and with multiple perspectives – to come up with a product direction for a set of communication features that a product team will be able to work on in the coming fiscal year.

An explicit objective of the consultation is to change communication on Wikimedia projects in some way, because the present wikitext communication system effectively forms a cultural barrier for new contributors, in spite of its flexibility and transparency. In enumerating various possible outcomes and solutions, the consultation page notes: "For this process to work, we need to be open to all kinds of directions."

In this stage of the consultation, the WMF suggests asking community members five questions. There are therefore five subsections for each of those questions (under § Suggested questions). It may also be appropriate for other issues to be considered on this page, in separate subsections (under § Other topics). Jc86035 (talk) 15:38, 23 February 2019 (UTC)[reply]

Bureaucratic stuff

[edit]
To allow for different types of Wikimedians to share their thoughts, we want everyone to be able to talk about wiki discussion systems in their primary language in an environment where they feel comfortable.

Wikidata at large is currently signed up as one "participant group" at mw:Talk pages consultation 2019/Participant group sign-up. WikiProjects, off-wiki groups and the like are also encouraged by the WMF to conduct their own discussions.

At the end of the discussion, one or more users are to summarize (i.e. formally close) the discussion and post the summary to mw:Talk:Talk pages consultation 2019. Since the WMF prefers that at least one primary contact be available, if you are interested in closing the discussion and will be able to do so, please add your signature to the next subsection. It may be appropriate for only administrators (or only experienced users) to close the discussion. Jc86035 (talk) 15:38, 23 February 2019 (UTC)[reply]

Users interested in closing the discussion

[edit]

Suggested questions

[edit]

When you want to discuss a topic with your community, what tools work for you, and what problems block you?

[edit]
  • Wikidata's "chat" page seems to work quite well, and the project is still small enough that centralizing most such discussion is entirely a plus. - Jmabel (talk) 17:15, 23 February 2019 (UTC)[reply]
    Hello Jmabel, and thank you for your feedback. I'm looking for details. When you see quite well, what works for good and what should be changed? Thanks, Trizek (WMF) (talk) 16:41, 13 March 2019 (UTC)[reply]
    • The only thing I would want to change is to enable watching individual sections. Other than that, I suspect most changes would be liabilities. (By the way "When you see quite well" seems very odd wording, was that a typo for something else?) - Jmabel (talk) 00:07, 14 March 2019 (UTC)[reply]
  • Broadly answering the question, since most WMF employees aren't Wikidata regulars: Wikidata is quite different to other WMF wikis, since most of its content is Wikibase-based, and it doesn't officially have a single language. Nevertheless, most of the important discussion fora (the project chat, and the other pages listed in {{Discussion navigation}}) are wikitext pages dominated by English speakers, although they do function reasonably well. Requests for comment can take an exceptionally long time (sometimes more than a year), and other discussions may not end satisfactorily due to the small number of users; the English Wikipedia, for comparison, invariably and inevitably has many more subject specialists in most areas. Talk pages for items appear to me to be dominated by spammers, and as noted by Jmabel, centralizing discussion is quite helpful considering the contributor–content ratio. A pertinent issue is that Wikidata is also newer than other projects, so many of the community processes may seem incomplete. Jc86035 (talk) 17:51, 23 February 2019 (UTC)[reply]
  • Talk pages for individual properties, and talk pages for wikiprojects, are also both forums that are significant and can work well. And user talk pages of course. I would like to note that I strongly prefer conventional talk pages to ones using Flow, because of the ability to easily read through previous conversations, and see the development of discussions presented in sequence and in context. Jheald (talk) 21:24, 23 February 2019 (UTC)[reply]
  • As others have said, Project Chat and talk pages for Projects and Properties seem to work very well. Good conversations also happen on Property Proposal pages and at Request a Query. I also use the Facebook group for off-the-wall questions. - PKM (talk) 23:11, 23 February 2019 (UTC)[reply]
    Hello PKM, and thank you for your feedback. I'm looking for details. When you see that you use Facebook, it if for which use? What is missing on the current discussion system(s) on the wikis that you would need and have on Facebook? Thanks, Trizek (WMF) (talk) 16:41, 13 March 2019 (UTC)[reply]
    The real-time aspect of Facebook groups and Telegram is crucial - particularly the ability to see a notification that someone has replied to a conversation on my phone, tablet, or laptop, regard;less of whether I have the application open or not. With Talk pages, all contributors have to be actively looking at a Wikimedia page to see a notification of a new message. That slows down conversations significantly compared to FB and Telegram. - PKM (talk) 21:02, 13 March 2019 (UTC)[reply]
  • I echo the comments above about Project Chat, the Properties and Project talk pages, property proposal discussions, RFC's, etc - those are generally working, though could be better. Discussions on item ('Q') talk pages - and I guess now Lexeme ('L') talk pages, almost never seem to be satisfactory. The central problem "blocking" progress on Wikidata I think is the difficulty of engaging the right community of users. Adding more pages to Watchlists by default (every time you edit an item?) and encouraging people to regularly check their Watchlists might help; the "alert" system when you are pinged right now does work but requires a "ping". And the "ping project" template is either broken right now or not doing what I would expect, it should be easier to contact the people you need to talk to! ArthurPSmith (talk) 14:02, 24 February 2019 (UTC)[reply]
    • On en-wiki a workaround for the problem of low-traffic article talk pages is to prominently have templates at the top of the talk page identifying the relevant WikiProjects for the subject. This gives people somewhere to go, if their initial post on an article talk page gains no response, or if further eyes are needed to break an impasse. It's not something we've ever I think taken up here. But perhaps here the expectation is that talk pages for individual Q-items are so unlikely to be looked at, that a WikiProject talk page or central ProjectChat is known to be the default venue for first response. Talk pages can be quite useful for reference information like the {{Item documentation}} template though. Jheald (talk) 10:22, 28 February 2019 (UTC)[reply]
    • I share the sentiment about item talk pages. As a relatively new contributor, I often find it desirable to consult those with knowledge or experience in a particular area (or even just be able to point them towards something that needs attention). Checking the item talk page is probably the first step a newer contributor like myself intuitively takes when seeking discussion/collaboration. Unfortunately right now that doesn't usually produce any valuable direction - my first port of call tends to be the project chat where I hope that someone interested in the same area sees my question. If WikiProjects could be associated to certain items and all instances and sub classes of those had links to the WikiProject on their talk page, I think that would massively improve the ease of collaboration (and discoverability of WikiProjects - I feel that currently you need to go out of your way to find them). --SilentSpike (talk) 11:32, 30 March 2019 (UTC)[reply]
  • Despite some bug, Flow is quite a good tool (no need to know the painfully unuseful wikisyntax, automatic signing and pinging, etc.). I'd love to see it more used (but - again - bugs need to be fixed first). Cheers, VIGNERON (talk) 16:43, 27 February 2019 (UTC)[reply]
  • Village pumps (or similar) and help desks (or Teahouses) typically work for me. "WikiProjects" are typically dead, or less than 3 volunteers try to be responsible for more than 1M pages, that's often the opposite of helpful. –84.46.52.11 00:39, 28 February 2019 (UTC)[reply]
  • I use Telegram groups, and sometimes Project Chat and WikiProject Italy. I prefer Telegram groups because it's more real-time --Sabas88 (talk) 17:38, 3 March 2019 (UTC)[reply]
    Hello Sabas88, and thank you for your feedback. I'm looking for details. What aspects of Telegram, beyond real-time, are missing on our wikis? What would you need on the wikis to satisfy your needs about talk pages? Thanks, Trizek (WMF) (talk) 16:41, 13 March 2019 (UTC)[reply]
    An usable mobile interface, notifications, threading (I imagine it as a proper forum in high traffic sections) --Sabas88 (talk) 08:58, 14 March 2019 (UTC)[reply]
  • It's worth noting that after the English project chat, the project chat that's most frequently used is the French one that user Structured Discussions (/Flow). While causation is always difficult to establish, this suggest that Structured Discussions encourages more people to participate in discussions. ChristianKl15:09, 8 March 2019 (UTC)[reply]
    Hello ChristianKl, and thank you for your feedback. I'm looking for details. Is there any aspects of the different chats make a difference for you? Why do they make a difference? Thanks, Trizek (WMF) (talk) 16:41, 13 March 2019 (UTC)[reply]
  • I try to avoid using FLOW as much as possible. If I am forced to use FLOW on my talk page, I will likely stop contributing.--Ymblanter (talk) 20:47, 9 March 2019 (UTC)[reply]
    Hello Ymblanter, and thank you for your feedback. I'm looking for details. What aspects of Flow you dislike and why? What aspects of "classical" talk pages you like and why? Thanks, Trizek (WMF) (talk) 16:41, 13 March 2019 (UTC)[reply]
    Infinite scrolling, a necessity to inspect every edit (I can not just grab a bunch of diffs and see what happened), pings for every section if I have a page watchlisted, to name a few. Generally, the discussions must be, well, structured, and it is necessary to see who replies whom, Whereas this can be achieved with the infinite number of subtopics, in this limit FLOW loses any advantages it was advertised for.--Ymblanter (talk) 16:53, 13 March 2019 (UTC)[reply]

What about talk pages works for newcomers, and what blocks them?

[edit]

I never had the slightest problem with them myself, but I'm a software developer, and used to markup languages. The one thing I see most is that people don't get the thing about signing your post. I suspect it might be useful that if someone's post does not include a signature, they would get a pop-up or such asking if they wanted one added. Logged-in users could have a one-time checkbox to dismiss this warning if they don't want to see it again. - Jmabel (talk) 17:18, 23 February 2019 (UTC)[reply]

What do others struggle with in your community about talk pages?

[edit]
  • A fair amount of Wikidata editors (1,145) use Structured Discussions (Flow) on their talk pages. As a result, those users might never actually have to interact with wikitext because of the nature of Wikidata content. I suspect Wikidata has a larger proportion of users who don't know how to use wikitext discussions than most other WMF wikis, although I don't really interact with new users much and almost never have to discuss anything with them. Jc86035 (talk) 17:58, 23 February 2019 (UTC)[reply]
  • The fact that Flow has no roleback feature was really annoying to me when my talk page was vandalized. Even if you manually delete vandalism it pushes the thread in which the vandalism was deleted to the top of the history of the talk page. If further resources are invested into Flow (which I would support), a proper roleback feature should be implemented. ChristianKl15:16, 8 March 2019 (UTC)[reply]

What do you wish you could do on talk pages, but can't due to the technical limitations?

[edit]
  • I wish there were a way to be automatically pinged for further comments in a particular section, especially so on wikis where I don't maintain a watchlist. - Jmabel (talk) 17:26, 23 February 2019 (UTC)[reply]
  • All the bugs with Flow, this one for instance phab:T106687. Cheers, VIGNERON (talk) 16:38, 27 February 2019 (UTC)[reply]
  • I would wish to be able to have a single discussion appear on multiple talk pages. {{ping project}} gets for example used quite frequently on Wikidata and it would be great to show all discussion that ping a particular Wikiproject on the talk page of that Wikiproject.
On the same token it would be great if it would be possible to have another tab to show all discussions that use {{Q|12345}} on Q12345. On Wikipedia, an analog feature might be to show discussions of pages that use a given template in another tab for pages of templates. ChristianKl15:22, 8 March 2019 (UTC)[reply]

What are the important aspects of a "wiki discussion"?

[edit]
  • A clear focus on a specific, actionable request or suggestion is helpful in moving conversations to some sort of conclusion. However, we should also allow for a modest amount of purely social interaction (thanks, checking in on other users, asking philosophical rather than procedural questions). - PKM (talk) 23:14, 23 February 2019 (UTC)[reply]
  • I think that a lot of discussions that should be centralised clearly aren't and oftentimes users aren't directed to the right venue, more often than not you have to be "in the know" with Wiki's, maybe it would be better to have a message appear when editing certain talk pages that links to more adequate forums are shown, for example when editing a talk page of an item this could direct people to the project chat or a local help desk, when editing the talk page of an administrator it should state "for more general enquiries related to administrators please go to the Administrators' noticeboard", Etc. If you've been editing for years but only edit content spaces you might not be as familiar with "community" spaces. Discussions tend to be non-centralised purely because users aren't directed to the right venues. -- Donald Trung/徵國單  (討論 🀄) (方孔錢 💴) 16:27, 25 February 2019 (UTC)[reply]
    Infinite scrolling, a necessity to inspect every edit (I can not just grab a bunch of diffs and see what happened), pings for every section if I have a page watchlisted, to name a few. Generally, the discussions must be, well, structured, and it is necessary to see who replies whom, Whereas this can be achieved with the infinite number of subtopics, in this limit FLOW loses any advantages it was advertised for.

Other discussions

[edit]

Before adding a new subsection, please check that the purpose of the section is not to address one of the non-goals. To create a new subsection, add a level 3 header, with your comments and signature below it, at the bottom of the page. Due to the open-ended nature of the consultation, subsections may be structured in any format, including (but not limited to) open discussion, support/oppose !vote, and multiple-choice !vote. ("!vote" is used on the English Wikipedia in referring to a poll, to emphasize that consensus is not based on simple numerical majorities.)

Orphaned talk pages

[edit]

How about after a sysop deletes a page its talk page gets automatically moved to "Wikidata:Orphaned talk page archives/26 February 2019" for future reference (unless vandalism or something is included), a system like this would probably be most beneficial for Wikipedia's but could be pioneered here. This way the deletion log will also automatically point to the archive and it will be moved to a sub-page named "Wikidata:Orphaned talk page archives/26 February 2019/DELETED PAGE" so it's quite easily accessible. -- Donald Trung/徵國單  (討論 🀄) (方孔錢 💴) 17:02, 27 February 2019 (UTC)[reply]

A discussion about talk pages, on a talk page, advertised on talk pages

[edit]

It concerns me that this is a discussion about talk pages, on a talk page, that has been advertised on talk pages. Possibly it might over-represent people that are comfortable with talk pages?

I do think we may have an issue, that although we are frequently assured that the number of contributing editors to Wikidata has become very large, nevertheless the community that participates in discussions (eg as to how properties should be used, how particular kinds of items should be modelled) often seems quite small.

I wonder if it would be an easy SQL query to run, to identify accounts that have made a lot of Wikidata edits (other than bots), but very few contributions to Wikidata talk pages? Perhaps those are the users we particularly need to be trying to reach for their opinions? Jheald (talk) 13:56, 28 February 2019 (UTC)[reply]

@Jheald: This would be impossible to satisfactorily resolve. I think it could be ameliorated by sending out a MassMessage, but you would still get an unknown skew of viewpoints because we don't know if talk pages themselves are driving away contributors. Nevertheless, sending out a MassMessage is probably a good idea. Jc86035 (talk) 14:13, 1 March 2019 (UTC)[reply]
Jheald and Jc86035, I share this concern. The team is making some (mostly small) efforts to geet those missing viewpoints (we're going to get a report from someone doing in-person observations at an edit-a-thon, and if anyone wants to talk to folks on Facebook or Twitter or whatever, that's perfectly fine), but we will be missing a lot. Most of the announcements are on wiki or in mailing lists for people already well-connected to the movement.
If you think it's a good idea, I could try to get an e-mail address created for people who can't post on this page. Then you could send a MassMessage with "post here or e-mail there", if you wanted. (Please ping me if you want me to look into that possibility.) It might also be possible to let people reply in Flow on Wikidata, if you think that would reduce the barrier enough. (I'm not sure whether this wiki's Flow defaults to visual or wikitext modes, so that would get automatic pings about replies for anyone who watches the discussion, but it might not hide wikitext, if that's the main concern.)
As a third option, I've talked to the team about using a survey, and if they decide to do that, then that's another thing that could be sent to users that you'd like to hear from. The biggest thing that I learned from mw:VisualEditor/Survey 2015 is that the commitment to on-wiki communication can be a lot lower than it sounds like. We got about 500 responses in Qualtrics and just two (2!) on wiki. I think this set of five questions is better suited for a discussion than for a simple survey. So perhaps while I've already got your attention, you would consider telling me if you have any ideas about what to ask in such a survey. I've been toying with questions like "Is it more important to improve access on mobile devices, or more important to keep things nearly the same?" (trade-offs and values) or "Which of the following ways have you used to communicate with other contributors?" (probably taking the list from mw:Talk pages consultation 2019/Tools in use). Whatamidoing (WMF) (talk) 19:51, 3 March 2019 (UTC)[reply]
@Whatamidoing (WMF): I think email might not be a good option (although my opinions here are not really based on any hard evidence, so take them with a grain of salt). The users might not have any assurance that their feedback would be received, and they might have to already have problems in mind in order to have something to write; i.e. "people don't know what they want until you show it to them". Furthermore, they would have to spend time re-stating the same problems that many experienced users have already mentioned.
Since the apparent main purpose of the consultation is to determine which things to improve, I think most of the survey questions would concern prioritization of issues. The discussion which has already occurred on the English Wikipedia (and on other Wikipedias) would be quite helpful in determining what to ask, since many contributors have exhaustively enumerated their problems and feature requests.
Maybe something like a "how much do you agree with these statements" 6-point scale page (e.g. "I have trouble indenting comments" or "I find talk pages [intimidating/inscrutable/weird/outdated]"), and then a text box for respondents to list other issues that weren't mentioned; and then a 6-point scale page for "how much would you want these improvements to be made" (e.g. "use of valid HTML on discussion pages" or "inline replies to comments"), with another text box; and then a 6-point scale page for "how important are these things to you" (e.g. "the ability to view all changes since the last visit to the page" or "the use of the same discussion software across all discussion pages").
I think identity questions would also be useful for analysis, as usual. In addition to the standard questions, I think field of work is something that hasn't been explored much in previous WMF surveys, and it could be important (e.g. in filtering software developers from analysis of how difficult talk pages are to use). Maybe things like gender identity, urban/rural, number of total edits, and number of months active (i.e. at least one edit made to any project) could also be requested.
I personally think that the "hub-and-spoke" system was not a very practical idea, even if it might be useful for constructing a survey. Aside from effectively excluding feedback from developers who usually only edit the MediaWiki wiki, users are essentially repeating the same things in each participant group; some users (e.g. myself) have to spend hours setting up groups and asking people to participate; the feedback is almost entirely dependent on the initiative of experienced users willing to put up with this stuff to set up pages on their own wikis; the feedback is also entirely dependent on having enough users who care about bureaucratic backwaters; the suggested questions are annoyingly vague; and there is very little assurance that the WMF is actually going to do anything with the feedback. (The landing page is also not getting a lot of page views for a global CentralNotice, or at least a large number of active editors are ignoring it and/or haven't bothered to find out what it is.) Jc86035 (talk) 10:02, 4 March 2019 (UTC)[reply]
I was reminded today that some previous work found that frequency of editing, rather than total volume, was the better predictor of struggle (at least on Wikipedia). The general idea is that daily editors (like us) will figure it out and have a lot of practice, but if you're on wiki once a year for Wiki Loves Monuments, or you drop in to fix a hundred typos every few months (the way some people binge-watch television shows), the oddities have to be figured out all over again.
I'm seeing some differences in the different consultations. This discussion and the one at the German Wikipedia have highlighted the difficulty of getting replies in the Talk: namespace. The German Wikipedia is concerned about archiving. (The English Wikipedia should be worried about that, given the mess that happened when we lost the Miszabots a while back, but I don't recall anyone mentioning it.) Multiple groups have mentioned signing as a problem for less experienced contributors. The English Wikipedia has talked about accessibility and indentation problems (one of my favorite discussions so far, by the way), but it hasn't really come up at the smaller wikis. The Chinese Wikipedia – I'm not sure that their concerns match anyone else's, actually. They have a lot more familiarity with Flow (more precisely, with the tiny fraction of Flow that has actually been implemented) on user talk pages, and their contributors haven't been around as long as us, and that really changes the perception. Whatamidoing (WMF) (talk) 21:59, 4 March 2019 (UTC)[reply]