Topic on Talk:Growth/Personalized first day/Structured tasks/Add an image/Flow

Hi all. While I understand why you are approaching things this way, I am not sure it is the best thing to do. It's good to see that you're using Wikidata as part of the input for these suggestions, but a lot of Wikipedias now are heading towards using P18 values directly in infoboxes (e.g., on enwp, telescope articles already do that; lots of eswp and frwp articles also do so; and of course Commons infoboxes display everything from Wikidata).

I suggest instead to focus on adding P18 values to Wikidata items - and to also think about adding other image values, like P3451 (nighttime view), there are also interior views, coats of arms, etc. That way, you will significantly increase the impact of what you're doing (the images will be used in multiple languages from the start), while avoiding adding extra clutter (when converting an infobox to Wikidata, do you use the existing image on Wikidata or the one from the article), and you're more aligned with what others in the community are trying to do (expand Wikidata and increase its usage), while your current approach may easily bring you into conflict with editors on the different Wikipedias.

(Also, note that P373 will hopefully be deprecated soon, you're best using the sitelinks to Commons instead.)

Hi @Mike Peel -- thanks for joining in this discussion! It is definitely great to hear from people who are experienced on Commons. We also just heard this same thought from users in Czech Wikipedia, when we brought up the idea there, because they use Wikidata infoboxes extensively. I'm trying to think about the right way to approach this. I could imagine altering the task so that newcomers add P18 values, but it brings up a bunch of questions for me. I hope you (and anyone else watching the page) can reply with your thoughts on them:

1. In the case where the image would be recommended because it is used on another Wikipedia, should we be concerned about whether it would "clear the bar" to be the P18 for the item? In other words, my sense is that the bar for image relevance and quality is higher for P18 than for an image in an article -- because P18 is meant to be the image to represent the item. Taking a real example from the algorithm, perhaps the unillustrated article is 1991 World Championships in Athletics – Women's heptathlon, which has no P18. And the recommended image is this photo, which is used on the same article in German Wikipedia. The photo is of the stadium where the event took place, but taken when some other event was happening in it. I do feel like that image meets the bar of illustrating the article in an enriching way, but not the bar of P18 to represent the item on Wikidata. What's your take on situations like these?
2. What would you recommend for articles that have a P18, but don't have an infobox in them? On the one hand, we could enable users to add the P18 to the article's wikitext. Or do you think that the best thing is to add a WIkidata infobox so that the P18 is drawn in (which might be a task out of the reach of newcomers)?
3. When adding an image as P18, I imagine that might immediately make it show up in infoboxes in many Wikipedias. Because we're thinking about newcomers as the primary audience for this feature, I'm a little worried about weak decisions from newcomers reverberating across many Wikipedias. How do you think we should think about this?
4. Related to the previous question: as Wikipedias inherit information from Wikidata through these infoboxes, would that mean that Wikipedias are not able to patrol those changes, and they are only patrolled on Wikidata? Has that been working well for Wikipedias so far?
5. If we enable newcomers to make Wikidata edits through this feature, I would be concerned that there are fewer opportunities for them to make connections to their Wikipedia communities. Their Wikipedia edit count wouldn't increase, their work wouldn't be visible on their own Contributions list, and local editors wouldn't be seeing their work in watchlists or Recent Changes. What are your thoughts on this?
6. Is there anything I can read about the potential deprecation of P373? I hadn't heard about that, and I definitely want the algorithm's researcher to look into it.

Thank you for helping us think about this work!

1. If you would be comfortable adding an image to the top of an article, I think it would be okay for wikidata:Property:P18. Per discussion on the property's talk page, usage is not heavily regulated. I don't think the stadium photo appropriately depicts the 1991 Women's heptathlon either for a main image or a P18. Lower in the article perhaps—even then it's a stretch but falls in the range of mediocre edits that are unlikely to be challenged unless someone happens to be watching that article.
2. P18 but no infobox: Every community will have its own preference. If you'd be comfortable adding the image to the top of the article with the currently proposed tool, I think it's fine to start with the assumption that it would work in both P18 and as a thumbnail (captioned or not) at the top of the related article. I wouldn't necessarily add infoboxes automatically—that's at least a sensitive subject on ENWP.
3. P18 automatically shows in multiple places: Per the thread below, I wouldn't worry about this. Wikipedias that have opted into automatically pulling P18 from Wikidata are aware that the quality is variable. If anything, it makes it easier to undo that addition if the tool had instead hardcoded the image into multiple Wikipedias.
4. Wikidata quality has been a source of contention for ENWP and is a reason why its editors would rather hardcode their own short descriptions, mostly not use Wikidata metadata in infoboxes, and not want to outsource its Category feature to structured Wikidata. There is a way to view an article's related Wikidata changes within the WP's watchlist but I'm guilty of not really looked into it after turning it off when it was rolled out (and really noisy). Mike will have more to say about this but to my knowledge, it's something that's already in heavy consideration. I think it's widely agreed that Wikidata would need more quality checks and integrations before ENWP considers integrating more tightly. Again, if other language WPs have already bought into using Wikidata because they would lack the edits otherwise, I think they're already comfortable with using whatever Wikidata provides.
5. I would doubt that newcomers would automatically receive encouragement whether adding images, adding links, or adding Wikidata metadata. Those are entry points to heavier edits. I wouldn't expect much of a "community" for someone making these edits by mobile. For comparison, check out new editors who join and use some of the desktop web antivandalism or other semi-automated tools. They need to hit a high volume before regulars notice. At most, newcomers using the proposed mobile features will get a "welcome" template, based on current ENWP practices. Some local editors tracking Wikidata in their watchlists might see, but I'd figure that community connection will be its own problem to solve with any of these structured task proposals. If you invite/build a contingency of volunteer editors from a WP to be invested in this project, they might be able to figure this out on their end without product support.
6. Visit wikidata:Property talk:P373 to find wikidata:Wikidata:Properties for deletion#Property:P373—it hasn't been decided yet, it seems? Even if/when it is deprecated, it will just be linked through sitelinks when relevant.

Thanks, @Czar. Everything you've said pretty much makes sense to me (I'm sorry for asking similar questions on other threads before realizing you've already answered them here). This is all helping me understand what P18 and P373 are and are not for.

Also pinging @Mike Peel again -- I hope you have a chance to weigh in!

Sorry for the really slow reply @MMiller (WMF)! I completely agree with everything that @Czar says. The only extra point I'd make is on (4), you can see Wikidata changes in your watchlist, and it can be useful on occasion (but there is a lot of noise). However, compare that with the use of templates and images in articles, which can be changed dramatically without any notice appearing in a watchlist/article history. It's more about trust than anything else (and, for enwp, requiring referenced data only - but obviously not for things like images). (And on 2, I'm generally a fan of adding infoboxes, but others disagree.)

I do think it's worth exploring the other Wikidata properties like nighttime view - perhaps these aren't as immediately visible very widely, so newbie mistakes will make fewer problems there? Also, consider suggesting featured or quality images for adding to other parts of the article, rather than just in the infobox.