There are many different elements to comment on. Le 13/11/2019 à 23:14, Ingo Koll a écrit : Ingo, I think it was very probably an issue with the tools.wmflabs. For some reasons, it is often "not working so well". In which cases, the tools hosted there show 404 or 405. It is unrelated to the tool. When you see such errors, better wait... and try again later. Unfortunately nothing we can do. All good proposals. But... It is probably good to explain that the tool is a pilot tool to support the Structured Data effort on Wikimedia Commons. It is not a tool to "do lot's of stuff on Commons directly from an app". I agree it would be cool to a certain extent :) But this was not the goal of the tool. There were three different ideas in our minds when we started discussing this tool in Cap Town in summer 2018. One idea from the WMF Structured Data team was to run several pilot projects to help adoption of structured data by community members. So their goal was outreach... community involvement... promotion of structured data etc. One idea was to improve the description of images uploaded during the Wiki Loves Africa contest. Initially, we thought the tool would be only for WLA. Then we realize it could be for... all categories on Wikimedia Commons One idea was to provide a tool that would allow running small contests, and that would be as simple to organize as possible. Practically speaking, to run a little challenge, all you need is to set up a campaign (takes litteraly 3 mn), then to advertise it (might take more time), then to plan a gift for winners. And that's it !!! Being a lazy person for example, I absolutely wanted that at the end of a challenge, there would be participation data available in cvs format, so that I could download and tweek data as much as I could to analyse participation. But I also wanted a visual "report", with all relevant information (name of participants, figures, licence, credits...) so that the program leader could simply do a screenshot at the end and taaaadaaaaa..... report done !!!! Example for WikiIndaba campaign : https://commons.wikimedia.org/wiki/File:Results_ISA_Campaign_WikiIndaba.png Those were the three goals in our mind all along. Could the tool do other things ? Of course it could. But the more complete, the more complex. In particular, we did NOT want to make categories updatable because the goal in our mind is to somewhat run away from categories to embrace structured data. So allowing category editing... would actually go against our goals :) I remember we had this discussion over time. This is irrelevant to the ISA tool itself. It touches to the role of Wikimedia Commons. Is Wikimedia Commons a place to host media, that could be used (or not) by anyone, in any types of circonstances ? Or is that a service to Wikipedia ? If you see Commons as being primarily a resource to illustrate Wikipedia, then yes, Commons is full of crap. If you see Commons as a more general resource, there is still a lot of crap, but also interesting things that are irrelevant to Wikipedia. But still interesting for others and for other uses. Note that the discussion is irrelevant to Wiki Loves X to a large extent. Please also note that Wiki Loves Africa has actually a fairly good reuse rate generally. But beyond this... Let me tell you what *I* see as very important in Commons and in structured data. That makes me... feel hot all around. Sometimes... I want to prepare a conference. And I want a picture. And this picture... I want it to represent a boat. And I want the picture to also feature a woman. And I want the picture to be mostly in blue shades. And I want the picture to be geographically in Africa. And last, I want the picture to be of good quality. How on earth do I find this picture RIGHT NOW ? I can not. For sure, I can go to a category... and start looking, looking, looking... and I could spend hours just looking for such an image. And eventually find something that could do. Perhaps. So what if I want to have 10 such images, filling up those conditions I request, so that I can actually choose for my favorite ? Will I go manually through every single image on Commons ? How could I find those when those images when they are only described by categories, which are only string based text. It is impossible. I can not find manually those images. And I can not really ask a machine to search that for me either. Now, if I drop string based categories, and instead describe images with structured data, such as describing with items from WikiData (as is being done with depicts in the ISA tool)... then... I can do queries !!! I can actually ask to a machine to look for images that represent a boat, a women, in blue, in good quality and located in Africa. And the machine will give me the answer. And that... is magic. Do you understand what I mean ? And if what makes your heart tick is ONLY Wikipedia... consider... how will you illustrate an article about... say... fishing. You want a GOOD image to illustrate fishing, but only fishing at night. How do you look for such an image ? You go to https://commons.wikimedia.org/wiki/Category:Fishing ? 1448 images to look at. Good luck ! Consider... if you could do a query... that would select for your own service "All the featured images related to fishing at night". Then maybe you will get 10 images. You select the best. You illustrate the article. Done ! In such cases... better have a big pool of images than a small one. This is what structured data will ultimately allow you to do... it will HELP you find the BEST and the RELEVANT. In French, we would say "trouver une aiguille dans une botte de foin". Cheers Florence