This month:
Development of the extension for setting up WikiProjects, as described in the last issue of this newsletter, is currently underway. No terribly exciting news on this front.
In the meantime, we are working on a prototype for a new service we hope to announce soon. The problem: there are requests scattered all across Wikipedia, including requests for new articles and requests for improvements to existing articles. We Wikipedians are very good at coming up with lists of things to do. But once we write these lists, where do they end up? How can we make them useful for all editors—even those who do not browse the missing articles lists, or the particular WikiProjects that have lists?
Introducing Wikipedia Requests, a new tool to centralize the various lists of requests around Wikipedia. Requests will be tagged by category and WikiProject, making it easier to find requests based on what your interests are. Accompanying this service will be a bot that will let you generate reports from this database on any wiki page, including WikiProjects. This means that once a request is filed centrally, it can syndicated all throughout Wikipedia, and once it is fulfilled, it will be marked as "complete" throughout Wikipedia. The idea for this service came about when I saw that it was easy to put together to-do lists based on database queries, but it was harder to do this for human-generated requests when those requests are scattered throughout the wiki, siloed throughout several pages. This should especially be useful for WikiProjects that have overlapping interests.
The newsletter this month is fairly brief; not a lot of news, just checking in to say that we are hard at work and hope to have more for you soon.
Until next time,
I would like to call to enable user talks' Flow function in English Wikipedia as early as possible, thank you.
Village Pump (proposals) would be the page for running that RFC. Actually is a Flow RFC currently being drafted by a few editors. A series of Flow pages have received consensus-to-delete, one page got consensus to end the Flow trial and convert back to Talk. The only two remaining Flow pages on EnWiki are this one, and the Flow Testing page. The current draft RFC is a proposal to completely *disable* Flow on EnWiki.
You're certainly welcome to initiate an RFC proposing broad activation of Flow, but I expected it would get SNOWed with Opposes.
Shwangtianyuan, I don't think this is the right place to discuss it. A more proper venue might be Wikipedia:Village pump (technical). Also a fair warning: Flow is a fairly contentious topic, so you will need to make a very persuasive argument.
Hello there! Happy to be writing this newsletter once more. This month:
Some good news: the Wikimedia Foundation has renewed WikiProject X. This means we can continue focusing on making WikiProjects better.
During our first round of work, we created a prototype WikiProject based on two ideas: (1) WikiProjects should clearly present things for people to do, and (2) The content of WikiProjects should be automated as much as possible. We launched pilots, and for the most part it works. But this approach will not work for the long term. While it makes certain aspects of running a WikiProject easier, it makes the maintenance aspects harder.
We are working on a major overhaul that will address these issues. New features will include:
- Creating WikiProjects by simply filling out a form, choosing which reports you want to generate for your project. This will work with existing bots in addition to the Reports Bot reports. (Of course, you can also have sections curated by humans.)
- One-click button to join a WikiProject, with optional notifications.
- Be able to define your WikiProject's scope within the WikiProject itself by listing relevant pages and categories, eliminating the need to tag every talk page with a banner. (You will still be allowed to do that, of course. It just won't be required.)
The end goal is a collaboration tool that can be used by WikiProjects but also by any edit-a-thon or group of people that want to coordinate on improving articles. Though implemented as an extension, the underlying content will be wikitext, meaning that you can continue to use categories, templates, and other features as you normally would.
This will take a lot of work, and we are just getting started. What would you like to see? I invite you to discuss on our talk page.
Until next time,
Have a good one all!
To let everyone know, I will be making a change to the {{Load WikiProject Modules}} Lua module, used on this project to render the different sections. This change involves, among other things, moving the different section headers out of the subpages and onto the main WikiProject page. This change will make the WikiProject easier to use and maintain. During this change, there may be a brief time window where the WikiProject page looks broken. Once the changes are finished, everything should be restored to normal. If not, purge your cache and it will work. If there are any bugs, please let me know promptly on my talk page. Thank you.
This work should now be done. Please let me know if there are any bugs.
Looking good :)
Just a heads-up that there's a new requested move for this article. This time I find myself on the other side of the fence to last time!
I love the new project page layout and the idea of automatically updated content thanks to WikiProject X and Reports Bot. But so much of the content Reports Bot provides is completely irrelevant to this project, or in some cases just not working.
Discussions: currently I can see 4 listings, all of which are identical (a link to the "Dream Factory" discussion on Talk:Television Centre, Southampton). Aside from the annoying repetition, there are more recent discussions on other talk pages that are tagged with our WikiProject banner, yet these aren't showing up (examples include Shirley Parish Church, Farnborough, Hampshire, Mims Davies to name just a few). Hitting the Refresh button doesn't seem to have any effect.
Assess for quality / not tagged by the WikiProject seem to be ok.
All the other Tasks sections are populated mainly with articles that are not tagged by this WikiProject so shouldn't be showing up on our project page. Examples that I can see on the front page right now include:
- Modern immigration to the United Kingdom (Expand)
- Honda CBR900RR (Unencyclopaedic)
- Powers of the police in England and Wales (Merge)
- Guildford College (Merge)
- Canon PowerShot G5 X (Wikify)
- Asylet, Oslo (Orphan)
- Sinai Synagogue, Leeds (Stub)
- Finchley Reform Synagogue (Stub)
Mentioning @Harej as he operates Reports Bot and is involved in WikiProject X.
If these can't be easily/quickly fixed we'll need to decide what to do. Updating all these sections routinely by hand doesn't seem viable as we don't have enough active editors to manage that, but displaying incorrect information on our front page doesn't make us look too clever.
Thank you for the feedback, Waggers.
The discussions list bug you point out is a known issue. We hope to figure out the cause of it soon.
The other sections are based on SuggestBot. Nettrom should have some insight as to why irrelevant content is being recommended.
Hi @Waggers and @Harej! Sorry about the late response to this, it's been a busy week.
If I understand you correctly, the problem here is that the suggestions include articles that are not a part of WikiProject Hampshire? This is something I've anticipated might be a problem, but I wanted to send out a round of suggestions and see if it actually came up before doing anything about it. The thing is, solving it might not be straightforward, but as I don't run a project I'm happy to hear thoughts on this.
The articles that SuggestBot suggests are all from "task categories" (documented here) because otherwise it will suggest articles that aren't in need of improvement. Due to this restriction, it sometimes suggests articles that are not clearly similar to its input set (in our case, articles assessed by the project). It will also not suggest the same article twice, which in turn means it sometimes have to pick an article at random from a task category.
I went through our logs and checked the articles Waggers listed, and they're either weakly related to the project, or randomly selected (which might be obvious, but it's good to know what went on behind the scenes).
One way to solve this is to only suggest articles from the project that are also in specific task categories. This has the benefit that it'll never suggest wrong articles. A couple of drawbacks are that it'll require the project to tag articles (e.g. if it needs more sources use {{Refimprove}}), as otherwise the bot has no articles to choose from. It'll also never identify articles that might be within the scope of the project, but which the project hasn't tagged yet, although IIRC, Harej has a program that looks for those already?
Curious to hear your thoughts on this.
First of all, I would like to express my appreciation for WikiProject Hampshire for testing the new WikiProject interface and tools. Thank you.
WikiProject X is up for renewal at the Wikimedia Foundation. We would like to continue working to make our existing tools better: to make them easier to use, and to integrate them with other Wikimedia projects, including Wikimedia Commons and Wikidata. Please review our renewal proposal and leave feedback.
As always, if you have any questions or feedback, please let me know on this talk page or the WikiProject X talk page.
Thank you again, Harej.
Hello there! Happy to be writing this newsletter once more. This month:
In July, we launched five pilot WikiProjects: WikiProjects Cannabis, Evolutionary Biology, Ghana, Hampshire, and Women's Health. We also use the new design, named "WPX UI," on WikiProject Women in Technology, Women in Red, WikiProject Occupational Safety and Health. We are currently looking for projects for the next round of testing. If you are interested, please sign up on the Pilots page.
Shortly after our launch we presented at Wikimania 2015. Our slides are on Wikimedia Commons.
Then after all that work, we went through the process of figuring out whether we accomplished our goal. We reached out to participants on the redesigned WikiProjects, and we asked them to complete a survey. (If you filled out your survey—thank you!) While there are still some issues with the WikiProject tools and the new design, there appears to be general satisfaction (at least among those who responded). The results of the survey and more are documented in our grant report filed with the Wikimedia Foundation.
There is more work that needs to be done, so we have applied for a renewal of our grant. Comments on the proposal are welcome. We would like to improve what we have already started on the English Wikipedia and to also expand to Wikimedia Commons and Wikidata. Why those? Because they are multilingual projects and because there needs to be better coordination across Wikimedia projects. More details are available in the renewal proposal.
The Wikimedia Developer Summit will be held in San Francisco in January 2016. The recently established Community Tech team at the Wikimedia Foundation is interested in investigating what technical support they can provide for WikiProjects, i.e., support beyond just templates and bots. I have plenty of opinions myself, but I want to hear what you think. The session is being planned on Phabricator, the Wikimedia bug tracker. If you are not familiar with Phabricator, you can log in with your Wikipedia username and password through the "Login or Register: MediaWiki" button on the login page. Your feedback can help make editing Wikipedia a better experience.
Until next time,
For this month's issue...
Making sense of a lot of data.
Work on our prototype will begin imminently. In the meantime, we have to understand what exactly we're working with. To this end, we generated a list of 71 WikiProjects, based on those brought up on our Stories page and those who had signed up for pilot testing. For those projects where people told stories, we coded statements within those stories to figure out what trends there were in these stories. This approach allowed us to figure out what Wikipedians thought of WikiProjects in a very organic way, with very little by way of a structure. (Compare this to a structured interview, where specific questions are asked and answered.) This analysis was done on 29 stories. Codes were generally classified as "benefits" (positive contributions made by a WikiProject to the editing experience) and "obstacles" (issues posed by WikiProjects, broadly speaking). Codes were generated as I went along, ensuring that codes were as close to the original data as possible. Duplicate appearances of a code for a given WikiProject were removed.
We found 52 "benefit" statements encoded and 34 "obstacle" statements. The most common benefit statement referring to the project's active discussion and participation, followed by statements referring to a project's capacity to guide editor activity, while the most common obstacles made reference to low participation and significant burdens on the part of the project maintainers and leaders. This gives us a sense of WikiProjects' big strength: they bring people together, and can be frustrating to editors when they fail to do so. Meanwhile, it is indeed very difficult to bring editors together on a common interest; in the absence of a highly motivated core of organizers, the technical infrastructure simply isn't there.
We wanted to pair this qualitative study with quantitative analysis of a WikiProject and its "universe" of pages, discussions, templates, and categories. To this end I wrote a script called ProjAnalysis which will, for a given WikiProject page (e.g. Wikipedia:WikiProject Star Trek) and WikiProject talk-page tag (e.g. Template:WikiProject Star Trek), will give you a list of usernames of people who edited within the WikiProject's space (the project page itself, its talk page, and subpages), and within the WikiProject's scope (the pages tagged by that WikiProject, excluding the WikiProject space pages). The output is an exhaustive list of usernames. We ran the script to analyze our test batch of WikiProjects for edits between March 1, 2014 and February 28, 2015, and we subjected them to further analysis to only include those who made 10+ edits to pages in the projects' scope, those who made 4+ edits to the projects' space, and those who made 10+ edits to pages in scope but not 4+ edits to pages in the projects' space. This latter metric gives us an idea of who is active in a certain subject area of Wikipedia, yet who isn't actively engaging on the WikiProject's pages. This information will help us prioritize WikiProjects for pilot testing, and the ProjAnalysis script in general may have future life as an application that can be used by Wikipedians to learn about who is in their community.
Complementing the above two studies are a design analysis, which summarizes the structure of the different WikiProject spaces in our test batch, and the comprehensive census of bots and tools used to maintain WikiProjects, which will be finished soon. With all of this information, we will have a game plan in place! We hope to begin working with specific WikiProjects soon.
As a couple of asides...
- Database Reports has existed for several years on Wikipedia to the satisfaction of many, but many of the reports stopped running when the Toolserver was shut off in 2014. However, there is good news: the weekly New WikiProjects and WikiProjects by Changes reports are back, with potential future reports in the future.
- WikiProject X has an outpost on Wikidata! Check it out. It's not widely publicized, but we are interested in using Wikidata as a potential repository for metadata about WikiProjects, especially for WikiProjects that exist on multiple Wikimedia projects and language editions.
That's all for now. Thank you for subscribing! If you have any questions or comments, please share them with us.