Doldrums post-Wishlist survey

Hi colleagues!

a possible soundtrack for this discussion

The tremendous Community Wishlist Survey is over and the results have been published December 6th. It illustrates that the Wikisource community is the biggest one, or at least that the people reading English are more in the Wikisource community than in the others, as multilingual translations were not possible for the proposals. So, another year without any dedicate development for most of the projects.

The Community Tech Team made a bold choice this year by excluding Wikipedia, Wikidata, Commons and general proposals. Was it enough? I am not sure. The result is a Wikisourcian year, and that’s great for this project. Well, the 5th proposal is labeled Wiktionary but will also support Wikisource, and would be accessible in Wikipedia also.

One great result this year is the quantity of proposals written. This session leads more people to express needs. And there is plenty of those. Plenty that will stay unsolved for another year.

There will be a time to rethink about a strategy to seduce independent developers, to convince our projects are not small anymore, to found another source of funding and development outside of the WMF if it doesn’t come from there (Chapters? Google summer of code? Crowdfunding? NGO? In residence?).

Well, I am not ready for that yet. I am in the doldrums. First, I think we need to discuss about the process of the wishlist survey and how exhausting and depressing it could be for the participants. I think it is very important to talk about this process, to help the Community Tech team to build a better one in the future.


I see two ways in which small communities could benefit from the community wishlist

  • either we only vote on transversal proposals (fight against vandalism, anti-harassment, improvement of users’ rights, etc.). Thus, big communities (Wikipedia, Commons, …) could also give their opinion and take benefit of these developments
  • or, for each Wikimedia project (Wikipedia, Commons, Wiktionary, etc.), each community selects only one wish, the one that has obtained the most support. Since there are many Wikimedia projects, one could imagine satisfying 5 projects one year and the other 5 the following year.

@ifried and @DannyH might be interested in this conversation.

As someone who is advocating for different kinds of knowledge projects to be supported from the foundation, I have been thinking about this alot. This is all my personal opinion – my main role at the foundation is to support outreach programs so I am really focused on new communities to engage, like GLAMs, Educators, activists, etc…

Here are a couple of thoughts on why the smaller projects have a hard time getting bandwidth:

  • Most of the decisions at the foundation for technological work are being made with two different axis: scale of impact (either reading or contributing), and impact on knowledge equity. It would be highly unusual for any of the investments from the foundation to not be responding to some type of popular demand or demonstrated potential impact via scale metrics (in the case of the Community Tech Wishlist its scale of participants and advocates for the platform). It’s actually really refreshing that WikiSource, Commons and Wikidata are getting taken seriously after years of only Wikipedia getting attention because of the scale dimension.
  • You need more than passion for the project, but also a business case for both the Foundation and other communities to take the work seriously: what does the Movement or the broader Wikimedia ecosystem gain from investing in that project? This could be partners, new contributors, a new way to deliver content to users, etc.

I think the thing that the WikiSource community has done in the last couple years is create a well understood equation for the value it gains from being supported. This is visible in a few ways:

  • Actively organizing as a user group, with a bit of a strategy, and identifying community developers able to triage initial needs gaps.
  • Actively documenting activities around the world, especially in emerging communities like India and Indonesia which are strategic markets for expansion for Wikimedia.
  • Run pilot outreach, communication and partnership programs which demonstrate the ability of WikiSource to scale if given proper technological support.

WikiSource has a lot of folks outside of WikiSource contributors, i.e. GLAM and Education advocates, supporting it alongside a very, very persuasive alignment with open cultural projects more generally (digitizing public domain books, oral history, etc). It’s also very clear that transcription is something well supported by crowdsourcing.

I think, of the other sister projects, Wiktionary has a lot going for them: it has pretty decent readership numbers, its a reasonable competitor with the top in class commercial websites, the contribution strategy and method for collecting content is healthy and clear, and the wikiformat and multilingual community creates a lot of value add that other dictionaries don’t have. For Wiktionary in particular, I think there are a couple of things that aren’t really clear to me from a strategy perspective, that I think make it hard to communicate:

  • What is the potential contributor community that could be scaled up if contribution were easier to Wiktionary? There has been a little bit of outreach programs in Armenia around using Wiktionary for getting youth involved in editing, but I really don’t have good examples beyond that.
  • Who are the potential reusers that we are not getting? I think Lexeme data gestures towards a utility for Wiktionary in the Wikimedia space, but I have yet to see a serious partnership leveraging Wiktionary – Wikidata, Commons and WikiSource all have long histories of “give and take” relationships with GLAM organizations, governments and other kinds of partnerships.

As for other projects, again in my personal opinion/capacity, I think that most of the WikiVersity proposals were showing how little shared “purpose” the platform has right now, and its clear that many of the other communities aren’t “there” so to speak – in that the underlying community hasn’t fully articulated itself in global spaces. The exception might be WikiVoyage.


I kind of overpass the doldrums, so I can now participate to this discussion.

Thanks @Astinson_WMF for this long comment. I’ll try to answer to it, but I may miss some part, and I am sorry if you feel I missed the most important idea.

About the axis of evaluation for decision and the idea of shaping our idea as business cases, that’s a way of thinking we are not accustomed to. It is quite hard to think at the same time about the content, the editors, the audiences, the network, partners and ourselves. I mean, we are part of the process, part of the producers and part of the readers, but most of us (contributors in general, not only wiktionarians) have a narrow perspective on our projects.

We may need some relay or material to understand how to build projects and to “articulated itself in global spaces” as you said. Meta is an unknown territory for most of the contributors.

Well, I made a document to present and teach how to contribute to French Wiktionary, with 180+ slides, made for a 6hours training. One of the first comment on this document by the community was by a sysop who said he learned a lot. So, we have to consider that contributors have an incomplete and subjective knowledge of their own project. I think building this kind of document is a good step to build a better analysis, for the writer of it and for the readers. Also, we documented our knowledge and activities in a monthly newsletter since three years. We translated some issues in English but we had zero feedback so we stopped the translation. This is also a tool to have a better knowledge of local activities. Still, as part of a wikiverse, we could connect locally during some editathon for those lucky enough to live close to a wikipedian nest, but it is quite difficult to know the global trajectory or actuality.

We also built a tremendous user group that promote a common activity with little success and had led to some discussions about the strategy. Sadly, one of our common interested during the last two years was Wikidata Lexeme project and we tried to participate in the process in vain. It stayed as a trouble area, where a lot of people lost their time or mind. I hope we could build another activity on something more positive in the future.

Well, to challenge your questions:

The potential contributor community that may be interested in Wiktionary is the people speaking endangered languages. They may not have enough knowledge of their languages to write an encyclopedia but a dictionary is a great entry point. Lot of field linguists collect words and sentences as they study languages - that was my training in an amazonian fieldwork - and those community will be ecstatic to be able to save their languages in a collaborative project, adding words, meanings, sentences of examples, pictures and sounds. In 2015, I wrote an IEG proposal to support this idea, but it was not the right timing. Atikamekw in Canada the following year had a better success. And recently, the endangered language Shawiya was the first language that opened in first a project that is not Wikipedia, and it’s a Wiktionary. If contribution was easier, Wiktionary could be used for this large community.

We have very interesting outreaches, and some awesome partnerships are in course, despite being largely confidential out of French Wiktionary community. Four examples:

  • The Comité de cartographie français (French comity for cartography) offers their glossary of about 800 entries to Wiktionary. We included this content in 2018 and we are setting up a training to let them maintain the content themselves.
  • A publisher have printed a dictionary based on Wiktionary content with a proofreading and editorial process that included a Wiktionarian, Lyokoï. Part of the selling price will be donated to Wikimedia France. It went out in September 2019 and had a large press coverage.
  • France ministry of Culture is funding an ambitious project to build a RDF dictionary for French, named Dictionnaire des francophones. It includes French Wiktionary and other resources. It will be online in March 2020. They hired me for two years has the product manager to build the ontology and lead the development.
  • In the course of Dictionnaire des francophones project, a Wiktionarian in residence will have a one-year full-time position in a university, starting in January 2020. To be more specific, it will be at the Institut international pour la Francophonie, in Lyon 3 University, France. The scope of this residence will be to facilitate the reuse of the data by documenting some templates and helping contributors to be more consistent using lexicographic terminology, and also to support the activities of the Institute, with editathons about Francophonie and collection of data such as pronunciations of visiting student from Africa.

Wiktionaries may be more used in the future with those project and because the content is very often quoted in press and in socio-commercial networks such as Twitter. 2019-2020 will be a major turn for Wiktionary…but mostly for French Wiktionary as the other communities are not very much into this kind of initiatives.

Lexeme is not very interesting for Wiktionary. A separate Wikibase under CC BY-SA could have been, and the English Wiktionary community is still thinking about having a dedicated wikibase. Lexeme is mostly interesting for Wikidata and for reusers such as Google. Well, with the project I am building now, we will have a SPARQL endpoint to dig into French Wiktionary data, so we will not try to have another one in the wikiverse, but at some point, it maybe sounds odd that French Wiktionary rely on an external service rather than on a Wikibase.

This post is already very long, and sorry for that. I think it is very interesting to talk about projects difficulties and how to diffuse what we do, but the core of the conversation was initially on the Wishlist Survey. We may open a secondary thread at some point.


I’ll keep it short and brief. The way this year’s wishlist was structured was a hit and miss. While it recognized that sister project proposals often get dominated by Wikipedia/Commons/Wikidata proposals, it also failed in a sense that 4 out of 5 top proposals (and 5 out of top 6) belong to Wikisource. This year’s results is yet another evidence to forego the “top X proposal” format. Even if not, drawing the line at top 10 (instead of just top 5 this year) would have benefited and be more inclusive to more projects.

1 Like