Wikidata:Second Birthday/by Micru

Yet another year has passed and in appearance not much seems to have changed. Only by taking a look back to how things were back in October 2013, one can notice that there has been indeed a lot of positive change that sometimes is overlooked if not put into context. As Sven did last year, I wrote a short subjective summary about the state of the project.

What has changed

One year ago there were no quantities, there were no ranks, no statement ordering, and no monolingual text. So all those sorted population statements with historical data that are now present in items representing administrative units were simply not there.

Also this year the initial batch of sister projects got their sitelinks managed through Wikidata. Commons came first, then Wikisource, Wikiquote, Wikinews, and Wikidata itself. There is still a lot to be done regarding sister projects, but at least now some communities got a first contact with Wikidata, which is a crucial step for our diverse communities to make the most out of it.

Less visible changes but also terribly important have been the ability to merge and redirect items. There have been also other changes in how labels and sitelinks are edited, which brought some distress to existing tools, workflows, and stability, but that are a necessary step towards the planned UI redesign.

The community developer front has been very active too. Volunteers, students, and third parties have developed more functions and applications, like badges, property suggestion, "in other projects" sidebar, data edition from Wikipedia infoboxes, the Wikidata toolkit, The Game, and Histropedia, just to name a few.

There are now 500 properties more than last year, which is a 50% increase over the thousand properties created during the initial year. This enhanced expressivity poses a challenge to organize them in a comprehensive and automated way, and to present them to new contributors without overwhelming them. It is expected that the planned “statements on property pages” feature will help to classify them and to maintain them better and without so much overhead maintenance effort as now.

The number of statements, labels and sources is increasing steadily, as it can be seen in the statistics. The community has been also gaining a deeper understanding about how to represent more information about the real world, through countless discussions in different domains and WikiProjects. This deep insight building is a slow process, however it is the engine of change and data improvement, totally necessary if the Wikidata community wants to go beyond the mere accumulation of unconnected facts.

And of course the number of attempts to reach out, and to integrate with external databases keeps increasing both in number and in diversity. One of the latest has been the creation of items to represent all human genes.

During last Wikimania there was what can be considered the first community gathering. A great opportunity to put faces to nicks, to brainstorm ideas, to discuss issues in a more relaxed environment, and of course to share great moments.

What remains the same

Wikidata keeps gathering an awesome community that inherits the expertise accumulated during the first generation of Wikimedia projects. There have been few incidents, driven mostly by the inevitable friction generated by a mix of cultures from around the world, with each member used to different ways of working and communicating.

The partnership development-community has a good team spirit. There have been bumps on the road, the always inevitable bugs, but there is a mutual understanding and a good communication. When proposing changes one doesn’t have the impression that is talking in the void, there is always a dialogue, and a justification that some features might have priorities over others.

Massive support of infoboxes still looks far away on the horizon considering that the number with units datatype is not there, and it is not yet possible to use data from unconnected items. An added complexity is that sometimes the data is not easily importable as it was mixed with text on infoboxes, so adoption will be a slow process, that could be boosted if there was development support during the transition.

The performance of the site was improved at the beginning of the year, however there has been a regression, specially for large items. Slow reactivity is still a monster to defeat, or at least to tame.

What is expected to change

There is still a lot of expectation for the datatype with units, and for the ability to access data from unconnected items. The conversion of many infoboxes depend on those features for being successfully adapted to draw information from Wikidata, but specially they are necessary to convince Wikipedians to participate in the change.

Also Wikidata for Commons has just been set into motion. This project is being set forward by WMDE in partnership with the WMF, whose social capital suffered a setback with the last big projects that were perceived as “coming from outside the community”, and not grown from within. A good technical solution doesn’t mean that it will be embraced, specially if the incentive for change is not high enough. In the case of Commons, this incentive could be offered through complex transitive queries, or better cross-project metadata integration, however those candies are not there yet.

At the community there is also the challenge to accommodate users with a high volume of automatic edits (which bring some errors), and users that work manually but more precisely. Perhaps a way to define in which project an editor is working, and how, would help to increase mutual understanding, since the way of working in Wikidata is more transversal, reaching more items at once, as opposed to other sister projects like Wikipedia where the edit context is more readily visible.

A new challenge will be to transform property constraints into something both practical and in line with Semantic Web standards. There are limitations about how much information can be presented in the user interface, plus other limitations if Wikidata wants to keep being compatible with standards like OWL.

On the ontology level, compatibility with level external standards also plays a significant role, which sometimes clashes with the need of devising user-friendly properties and clear organization methods. So far there has been more efforts in gathering data than in defining basic concepts like “kitchen”, “freedom”, or “wish” in terms of relationships with other concepts. All advances in this area will help pave the way for Wiktionary, one of the still distant frontiers.

Conclusion

Wikidata is thriving and very alive. The number of challenges remains high on all fronts and the advances might not be as quick as they might have been predicted one year ago, but they are nevertheless happening. What makes a community strong is mutual understanding between all the persons involved, and so far we have enjoyed a lot of a good camaraderie that brings us all together forward as a team. If we continue like that we will see many more great Wikidata years in the future. Congratulations to everyone involved!