IRC office hours/Office hours 2018-02-13

From Meta, a Wikimedia project coordination wiki
Jump to navigation Jump to search

[18:00:06] <spinster> Welcome all!
[18:00:08] <pigsonthewing> Hi folks
[18:00:25] <Keegan> Hi all, welcome to this session for SDC
[18:00:38] <Keegan> Structured Data on Commons, SDoC, whatever you want to call it :)
[18:01:09] <Keegan> So, we're having this as a chat, we don't have any presentations unless people insist
[18:01:17] <Keegan> Is there anything up for talking about?
[18:02:02] <Keegan> We can always start with an update of where we are, if that's desired
[18:02:08] <pigsonthewing> Please do.
[18:02:21] <Lucas_WMDE> that would’ve been my question :D
[18:02:44] <Keegan> Okay, we'll begin there
[18:02:55] <Keegan> risler will be able to answer that
[18:03:21] <risler> We're currently in a Design and Prototype phase
[18:03:43] <Keegan> (don't worry, there's more!)
[18:03:56] <risler> Some of you have already seen our first, small, proposed changes to UploadWizard for multilingual file captions/descriptions
[18:04:15] <risler> We are also working on design updates to File pages and Search
[18:04:43] <Keegan> (captions are at https://commons.wikimedia.org/wiki/Commons:Structured_data/Get_involved/Feedback_requests/Multilingual_Captions )
[18:05:03] <risler> We will also be doing some Information Architecture design; soon we will be asking the community to help us with the new ontology for what data goes in what system in this new Structured Data paradigm on Commons
[18:05:44] <risler> that sums up where we are now :)
[18:05:55] <Keegan> That conversation should be starting in a couple of days. It's going to be a fun one.
[18:06:49] <Lucas_WMDE> “what data goes in what system” – as in, what’s on Commons and what’s on Wikidata?
[18:06:51] <VIGNERON> Keegan: where will be this conversation?
[18:07:29] <VIGNERON> and +1 with Lucas_WMDE question ;)
[18:07:32] <Keegan> VIGNERON: it will be a new feedback request subpage on Commons
[18:08:15] <Keegan> It should be up at some point on Thursday
[18:08:17] <risler> Lucas, the short answer to that question is "yes" :)
[18:08:18] <spinster> VIGNERON: if you are a member of the community focus group, you will be pinged about it
[18:09:04] <Lucas_WMDE> risler: thanks :)
[18:09:05] <risler> the longer version has to do with what gets stored on the new instance of Wikibase on Commons, what is referenced on Wikidata via Wikibase federation, and what remains on Commons as wikitext
[18:09:17] <Lucas_WMDE> ah, okay, more options :D
[18:09:34] <Lucas_WMDE> (I probably shouldn’t comment on that discussion with my WMDE hat on, just wanted to understand what it was about)
[18:09:47] <Keegan> Plug for the focus group sign-up: https://meta.wikimedia.org/wiki/Global_message_delivery/Targets/Structured_Commons_focus_group
[18:10:39] <spinster> This is a proposal / first throw of what we mean: https://commons.wikimedia.org/wiki/File:Structured_Data_on_Commons_-_which_information_goes_where_-_version_2017-10-31.png
[18:10:48] <Keegan> I think risler is writing more on this, is there another topic that someone would like to discuss, or question to ask?
[18:10:49] <spinster> To be refined and improved together.
[18:10:52] <Keegan> Ah, there's the link
[18:11:00] <sjoerddebruin> EXIF will be a lovely discussion.
[18:11:08] <Keegan> Wouldn't it though?
[18:11:29] <Keegan> I think for now, EXIF stays in Mediawiki, is that correct risler?
[18:11:34] <sjoerddebruin> (not related to where, but how)
[18:11:35] <Keegan> If so, is that how it should be?
[18:11:59] <YULdigitalpreser> I'd like to share that the International Image Interoperability Framework community is interested in Structured Data for Commons. https://groups.google.com/forum/#!topic/iiif-discuss/Nb4KNWSltlc
[18:12:18] <revi> https://commons.wikimedia.org/wiki/Commons:Structured_data/Get_involved/Feedback_requests/Multilingual_Captions TL;DR is "V2 is selected", right?
[18:12:22] <YULdigitalpreser> I let the IIIF group know I'd be willing to volunteer as such an ambassador.
[18:12:28] <yannf> how will be defined "notable contributors"?
[18:12:38] <revi> probably one with WD Qitem
[18:12:42] <risler> EXIF goes in Wikibase@Commons at the moment, but again, that's up for discussion :)
[18:12:48] <spinster> YULdigitalpreser: That would be awesome!
[18:13:04] <BrillLyle> It would be great to have EXIF metadata parsed into the structured data. Not sure if it is now or will be? If that EXIF information is searchable?
[18:13:12] <Keegan> revi: I wouldn't say "selected" so much as "that sounds like the design type that we should probably go with" :D
[18:13:15] <yannf> revi, anyone can have a WD Q item
[18:13:27] <revi> I don't think i can have a Q item?
[18:13:35] <revi> Keegan: sounds same to me lol
[18:13:54] <yannf> I do have one ;)
[18:14:06] <Keegan> revi: You know, leave a little bit of room to change if needed :)
[18:14:12] <revi> haha
[18:14:13] <revi> yeah
[18:14:17] <pigsonthewing> I'd be interested to know where using Wikidata QIDs to identify concepts lies on the roadmap: "this is a picture of a Q397447 at Q1799857"
[18:14:19] <revi> yannf: I don't, heh.
[18:14:21] <Keegan> BrillLyle: spinster has an answer for you I believe
[18:14:39] <yannf> anyone who has published anything can get a Q item as author
[18:14:43] <Keegan> pigsonthewing: Okay, we'll get to that in a moment
[18:14:59] <spinster> @BrillLyle This is what we are going to discuss in the upcoming community consultation indeed: the pros and cons of having EXIF data in Wikibase.
[18:15:23] <yannf> or is doing something in official capacity, that's a pretty easy
[18:15:29] <nathalie> what would be the cons of having EXIF data in Wikibase?
[18:15:38] <BrillLyle> Q items require some notability -- ugh I can't believe I said that. I don't know I am worried a Q item requirement might inhibit folks from Wikipedia, etc. sister projects who aren't comfortable with Wikidata (yet)
[18:16:16] <VIGNERON> nathalie: same here, I see only a lot of pros!
[18:16:22] <pigsonthewing> EXIF is often incorrect - especailly dates/ times, coordinates.
[18:16:22] <abittaker> pigsonthewing: that is indeed in the roadmap. we're calling it "depicts" there.
[18:16:27] <revi> sounds like 'defining "notability"' is community business in the end
[18:16:40] <revi> community@commons
[18:16:45] <Fae> spinster: nathalie: It could become a privacy issue if EXIF became easily searchable (in theory it is now, but only the tech savvy tend to do it). When people upload, they do not expect their camera to be searchable against all images they have ever uploaded, against all past accounts...
[18:16:52] <spinster> @nathalie Not sure - we have heard some pushback on it, but we don't know the exact arguments and would like to hear those.
[18:17:03] <pigsonthewing> Not just "depicts", but "collection", "owner", "location", etc.
[18:17:24] <spinster> Fae: That's one very valid objection indeed.
[18:17:52] <BrillLyle> maybe there could be an option to strip identifying metadata from images
[18:18:03] <BrillLyle> when people upload
[18:18:06] <tobias47n9e> If the structured data discussion is still going on, I was interested to hear if there would be a way to make image requests structured too.
[18:18:40] <addshore> tobias47n9e: it is indeed still going on :)
[18:19:02] <tobias47n9e> addshore: Oh hey! Nice to hear from you
[18:19:04] <Fae> BrillLyle: I'd rather keep all original EXIF. It's a question of how visible or searchable we make /all/ of the file metadata, rather than just the obvious date, GPS, camera settings.
[18:19:32] <yannf> BrillLyle, I was asking because there were some disagreements on Commons about creating Creator templates linked to WD
[18:19:36] <BrillLyle> I joined a few minutes late but I was wondering if the timeframe and status update of progress of the project has been covered? If there's an updated link to the project timeline?
[18:19:44] <marktraceur> tobias47n9e: Where would you want that data to be stored? It would show up as not having a "representative image" in Wikidata, but the Commons Wikibase instance wouldn't have any information on an image that doesn't exist...
[18:19:49] <risler> pigsonthewing, collection, owner, etc will also be available to Wikibase@Commons via "references" to Wikidata items through Wikibase Federation
[18:19:58] <revi> if they really care about the metadata privacy, they should take care to remove it before uploading IMO
[18:20:24] <BrillLyle> @Fae - agree, but some people might like privacy options for their EXIF meta, right? Just thinking about this because I have to strip metadata at my job often
[18:20:31] <nathalie> spinster: Fae: I did a little experiment with EXIF and privacy a couple of years ago, here is the account: https://www.academia.edu/29221627/Invisible_and_Public_Image_Embedded_Metadata_and_Contextual_Visibility
[18:20:32] <VIGNERON> Fae: maybe we can keep the complete original EXIF as now and just the common part and not private as data
[18:20:34] <pigsonthewing> Thank you; my question was /where/ in the roadmap - i.e. when? Or: sooner or later?
[18:21:02] <yannf> about privacy, I am more worried about GPS data than camera ID
[18:21:08] <Fae> VIGNERON: That's pretty much how Commons image pages work now. We should continue with that as a design choice.
[18:21:09] <risler> pigsonthewing, you will see most of this done this year
[18:21:10] <Keegan> These thoughts on EXIF are good to hear, please do post them to the consultation when it goes live later this week
[18:21:12] <spinster> @BrillLyle The roadmap has not significantly changed since we uploaded it - it's still quite correct as you can find it on https://commons.wikimedia.org/wiki/Commons:Structured_data/Development
[18:21:24] <BrillLyle> @yannf Yikes. I love the idea of the Creator templates. I wish they weren't manual. But I can't imagine there is a reason to not want them. I would like them to be more integrated in this project work maybe?
[18:21:28] <Keegan> Because we don't have answers on EXIF, we're looking for this kind of input and discussion
[18:21:36] <tobias47n9e> marktraceur: Maybe a request namespace / i.e. wikibase installation? I think there would be some interest from app developers to write sparql queries to find very specific image requests.
[18:21:51] <yannf> BrillLyle, I agree with that, but some others don't
[18:22:08] <BrillLyle> @spinster - Thanks so much
[18:22:32] <marktraceur> tobias47n9e: I'd suggest that would be far better served by another system, but I'm open to the idea
[18:22:39] <jheald> sorry to have missed the start -- where are we?
[18:23:16] <Keegan> Hi jheald, we were talking around EXIF data and where to put it. That conversation will go into the next consultation of ontology.
[18:23:18] <BrillLyle> @yannf I hope it doesn't become something similar to the Infobox wars on En Wiki
[18:23:25] <addshore> jheald: you can catch up with the logs @ https://wm-bot.wmflabs.org/logs/%23wikimedia-office/20180213.txt
[18:23:35] <Keegan> jheald: Was there something you had in mind for today?
[18:23:59] <BrillLyle> Will there be an improvement in the categories situation on WikiCommons when Structured Commons is rolled out?
[18:24:00] <pigsonthewing> risler: cool!
[18:24:01] <yannf> BrillLyle, yes, that's why we should look ahead for this issue
[18:24:26] <jheald> thx addshore, let me just catch up
[18:24:59] <marktraceur> BrillLyle: In that many categories will become mostly obviated by the more complex and structured data stored in Wikibase, I'd say so...
[18:25:17] <tobias47n9e> marktraceur: Sounds good. Probably a good idea for the Commons app to get a Wikibase installation on Labs for that
[18:25:26] <jheald> marktraceur: I don't believe that
[18:25:35] <BrillLyle> @marktraceur -- oh good, so categories (and hopefully tags) will become three dimensional, vs. current hierarchical model?
[18:26:01] <pigsonthewing> Maybe, once we demonstrate how categories can be done better using structred data, on Commons, Wikipedias will follow suit?
[18:26:02] <addshore> tobias47n9e: I might be able to help out if you need something there ;)
[18:26:11] <marktraceur> jheald, BrillLyle e.g. instead of having a category for pictures of men and dogs, there would be a property that separately indicated the picture included a man and a dog.
[18:26:17] <jheald> marktraceur: but I do believe that the project is proceeding on that assumption, and that worries me
[18:26:21] <BrillLyle> Q on Roadmap: From what I can see, we should expect some sort of official document on March 1, 2018?
[18:26:31] <yannf> BrillLyle, example: https://commons.wikimedia.org/wiki/Creator:Diego_Delso not "notable" outside of Commons
[18:26:44] <marktraceur> jheald: Well, it's still open for discussion, that's going to be part of the much larger ontology discussion
[18:27:00] <tobias47n9e> addshore: Awesome. If you are at the hackathon we could talk with Vivek about it. Maybe some other Commons app people will come too.
[18:27:07] <addshore> tobias47n9e: I will be :)
[18:27:14] <jheald> would be very very useful IMO if each category had a CommonsData item
[18:27:18] <BrillLyle> Will this be similar to the "alt=" text that is descriptive to people with visual impairments? I am not stating this very clearly....
[18:27:23] <sjoerddebruin> I'm still waiting for scholarships decisions for the hackathon ;)
[18:27:38] <tobias47n9e> addshore: Awesome!
[18:27:38] <jheald> 1) would enormously help with the transition
[18:27:40] <addshore> I /might/will/could/maybe/ have something coolish to demo at the hackathon tobias47n9e ;)
[18:28:06] <jheald> 2) categories likely to be around for a LONG time as curated spaces
[18:28:09] <pigsonthewing> Yannf: Did you try creating a Wikidata item for him?
[18:28:21] <sjoerddebruin> There is a Wikidata item. https://www.wikidata.org/wiki/Q28147777
[18:28:22] <tobias47n9e> addshore: Me too. Secret secret project. Def. better than center core sinking a drone ship :)
[18:28:23] <BrillLyle> Speaking of ontologies --- What is the status of using various / existing ontologies? Is there a comprehensive list of what will be utilized and cross-walked?
[18:28:41] <jheald> so marktraceur: when and where will this ontology discussion take place?
[18:28:41] <Keegan> BrillLyle: so that "official document" you are referencing will be the technical decisions of what we're going to put where (the ontology). That should be more toward the end of March
[18:28:52] <sjoerddebruin> But I think he wants to know how we can keep incrowd people from notable people.
[18:29:00] <Keegan> The community conversation starts this week and runs through March 1, so we're not going to turn around and decide that day :)
[18:29:03] <yannf> pigsonthewing, he has one
[18:29:28] <BrillLyle> @Keegan Ah, thanks so much. Will there be community feedback on drafts of that or is it non-public facing? Assume we can also provide feedback when the document is released?
[18:29:47] <sjoerddebruin> Opening up the Wikidata nobility for this would create a huge loophole, so we need indeed something for that.
[18:30:34] <BrillLyle> @sjoerddebruin -- concerning to hear "incrowd" reference. I think the concept of contributor or creator (a la subject) would be more helpful
[18:30:41] <spinster> @BrillLyle re: existing ontologies: I'm preparing that discussion at this moment. You'll hear more from that in the upcoming months too. Ideally, the Wikidata and Commons communities would work together in mapping the most common externally used ontologies to Wikidata items and properties
[18:31:12] <spinster> Some of it is already happening (e.g. the Art and Architecture Thesaurus is already in Mix'n'match and mapped gradually)
[18:31:23] <Keegan> BrillLyle: ^. Basically, we have two ontology conversations. The What Goes Where of Commons, and then the deeper structuring of the actual data
[18:31:25] <BrillLyle> @spinster Thanks. Very important to include stakeholders like folks from librarian and archive worlds -- WikiCite too
[18:31:47] <spinster> @BrillLyle That's definitely what we want to do :-)
[18:31:59] <pigsonthewing> If, in the end, we have to have items on Commons-Wikibase about photographers, we will need to replicate (or "trasnclude", in some sense) a lot of Wikidata properties
[18:32:14] <BrillLyle> @sjoerddenbruin -- Wikidata notability needs to be more flexible if Structured Commons is to work. Wikidata is building block/skeleton for much content
[18:32:16] <VIGNERON> +1 for including librarians (and wikisorcerors), books have specific needs
[18:32:16] <BrillLyle> IMO
[18:32:39] <Keegan> pigsonthewing: that sounds probably correct :)
[18:32:56] <Lucas_WMDE> pigsonthewing, Keegan: I thought there weren’t going to be structured pages on Commons for creators?
[18:33:25] <Lucas_WMDE> only “smart URIs” to their user page (regular wiki page)?
[18:33:32] <Lucas_WMDE> or did I misunderstand something?
[18:33:41] <BrillLyle> @spinster re: existing ontologies -- I hope the Wikidata identifiers will be the priority, with VIAF and other identifier systems that provide Authority control. I'm probably preaching to the choir here :-)
[18:33:43] <risler> Lucas_WMDE; you are correct
[18:33:50] <risler> smart URIs
[18:33:52] <pigsonthewing> "Structured user pages"?
[18:34:24] <NotASpy> https://commons.wikimedia.org/wiki/File:Structured_Data_on_Commons_-_which_information_goes_where_-_version_2017-10-31.png this is shit, nobody can ever find when a property on Wikidata has been edited and it's screwing up a page on English Wikipedia. It'll be absolute chaos if it's rolled out to Commons without being fixed.
[18:34:46] <addshore> sjoerddebruin: BrillLyle I mean, the "structural integrity / use" part of wikidata WD:N is pretty open already
[18:35:01] <NotASpy> edits to linked Wikidata items need to display in the history of the page on Commons, either merged or side by side, or some other way of being clearly visible when you hit History on Commons (or any other local project).
[18:35:02] <risler> but I think pigsonthewing was referring to something else - having media relating to photographers and having to include a lot of data from Wikidata that references that Photographer
[18:35:08] <spinster> @BrillLyle Nothing wrong with that :-) Yes, identifiers are one way to go. And there is also a way to map properties to external ontologies - let me find an example
[18:35:51] <pigsonthewing> Addshore: and yet items about Wikimana speakers, with links to videos of their talks on Commons, were deleted without discussion, on Wikidata
[18:35:58] <BrillLyle> @addshore - yes... for now it seems to be but recent discussions impinge on this :-(
[18:36:06] <yannf> https://commons.wikimedia.org/wiki/Commons:Village_pump/Proposals/Archive/2017/08#Finalize_Commons:Creator_and_approve_as_policy
[18:36:15] <Keegan> NotASpy: the plan is to address those issues so that they're not issues.
[18:36:16] <spinster> https://www.wikidata.org/wiki/Q1656682 - equivalent class property - you see this one is mapped to the CIDOC CRM, a museum ontology
[18:36:26] <yannf> the discussion about creating Creator templates ^
[18:36:33] <addshore> BrillLyle: pigsonthewing sjoerddebruin sounds like the community need to have some more discussion around WD:N
[18:36:35] <NotASpy> pigsonthewing: yeah, the standard of administration on Wikidata is woeful. It's one of those projects where CiR cases go to pass an RfA, sadly.
[18:36:41] <Keegan> NotASpy: Content and their histories will be accessible on Commons
[18:36:44] <jheald> NotASpy: Would you flag every single image of every painting by a creator, whenever somebody added a day and month to their date of birth?
[18:37:16] <NotASpy> jheald: I'd want to see that edit on Wikidata by clicking History on Commons, yes.
[18:37:23] <BrillLyle> @addshore -- no! :-) laughing here. it is not necessarily a good thing this discussion, IMO
[18:37:33] <pigsonthewing> NotASpy "CiR"?
[18:37:37] <sjoerddebruin> ^
[18:37:44] <NotASpy> jheald: it's about making changes on Wikidata which impact Commons (or en.wp etc) visible on the local project.
[18:37:50] <geniice> competence is required
[18:37:57] <BrillLyle> @NotASpy - YES
[18:38:02] <BrillLyle> Data is a good thing
[18:38:10] <BrillLyle> Curated, structured metadata even better
[18:38:15] <Keegan> Let's not go down the road of project v project, there's too much to discuss in the topic at hand :)
[18:38:19] <jheald> spinster: ideally we should have an external ID property for CIDOC, not equivalent class
[18:38:40] <sjoerddebruin> Please note that the amount of administrators on Wikidata compared to the size of the project is pretty low. (53)
[18:38:58] <BrillLyle> Is the persistent URI similar to or equivalent to the User ID that all registered users have? Could they be conflated?
[18:39:00] <jheald> 'equivalent class' statements ought to be cleaned out and turned into external IDs wherever possible
[18:39:23] <spinster> jheald: I heard otherwise from people who are into Linked Open Data; some seem to prefer equivalent class. It's one of the things I'd like to see refined / discussed more thoroughly if possible
[18:39:45] <NotASpy> no objections to the data being stored on Wikidata, just as long as we can work out and track changes on Wikidata which affect how that data is displayed on Commons. I actually support it, because when you spend time and get it right, it's bloody powerful. It's just too difficult to work out why something has broken locally when there's not been an edit to a page in weeks, and then having to trace it on Wikidata.
[18:40:10] <Keegan> NotASpy: /me nods
[18:40:13] <jheald> spinster: well yes, that's how they do things in Linked Open Data. But it's not how Wikidata does things
[18:40:22] <BrillLyle> like Name : BrillLyle User ID : 27454434 from https://tools.wmflabs.org/quentinv57-tools/tools/sulinfo.php?username=BrillLyle
[18:40:25] <Lucas_WMDE> NotASpy: FWIW, Wikidata edits should already show up on the watchlist and recent changes of relevant pages on other wikis
[18:40:45] <Lucas_WMDE> (except that it’s currently disabled for commons for performance reasons, https://phabricator.wikimedia.org/T179010)
[18:40:50] <Lucas_WMDE> (temporarily)
[18:41:27] <BrillLyle> Speaking of performance issues -- will structured commons impact and/or improve performance?
[18:41:29] <Lucas_WMDE> I’m not sure if showing them on history pages as well has been discussed before
[18:41:50] <jheald> spinster: we simply have too many external sites we track to for equivalent class to be efficient, and filtering by http prefix is monumentally inefficient
[18:41:52] <NotASpy> Lucas_WMDE: it's still a bit haphazard if there's a few changes to review.
[18:42:00] <addshore> BrillLyle: What parts of performance? =o
[18:42:04] <Keegan> BrillLyle: Could you be more specific? The performance of what?
[18:42:33] <pigsonthewing> I recently raised the issue of lack of Wikidata admins: https://www.wikidata.org/wiki/Wikidata:Project_chat/Archive/2018/01#Lack_of_admin_attention
[18:42:39] <BrillLyle> The performance of Wikimedia Commons, its response time. Also if Wikidata is slow will that impact Commons?
[18:42:39] <spinster> jheald: Sounds like a great argument pro identifiers indeed :-)
[18:43:21] <jheald> spinster: 'equivalent class' is now legacy, with only about 1000 uses
[18:43:34] <risler> BrillLyle, which response time specifically? Image load time? Search performance? Meatadata load?
[18:43:39] <NotASpy> Lucas_WMDE: personally, I'd like 'related change' edits to Wikdata to slot into the local page history dynamically and be highlighted in yellow or similar, so I can trace it back more easily. Just as it does with related changes, but clearer and in history.
[18:43:40] <pigsonthewing> I asked "What can we as a community do, to ensure that matters needing admin attention are addressed in a timely manner? ". There was only one response :-(
[18:44:24] <NotASpy> pigsonthewing: the only real response is 'nothing'. Wikidata like all projects has more administrative actions needed than admin 'time' available to carry out the actions.
[18:44:24] <sjoerddebruin> That's the "problem" on Wikidata: people like to edit and be productive instead of going into discussions.
[18:44:32] <BrillLyle> @risler -- all of the above. I get a lot of lag when I edit Wikidata, Wikimedia Commons, sometimes En Wiki. I know there are performance issues. Just wondering if Structured Commons will improve and/or create additional
[18:45:10] <BrillLyle> @sjoerddenbruin Ha!
[18:45:10] <jheald> So presumably, with ontology, the most important thing is to make it extensible, so people can propose new properties as with Wikidata; will that be the case?
[18:45:46] <pigsonthewing> I see Lydia is lurking - *waves* - Maybe we can disciuss that in the next Wikidata office hours?
[18:45:50] <BrillLyle> I can't imagine WHY no one responded to you Pigs
[18:46:11] <addshore> pigsonthewing: more eyes
[18:46:12] <pigsonthewing> ?
[18:46:27] <Keegan> BrillLyle: There are not concrete plans for SDC related to performance other than to not make it worse :)
[18:46:29] <jheald> Then, it will be a question of properties being created, as people modify templates to be able to draw data from CommonsData ?
[18:46:32] <NotASpy> there's too many properties on Wikidata, there's really a need to make sure it has all that's needed and nothing extra. I'd quite like Commons related fields to need agreement from Commons first, so we don't have thousands of new fields for Commons users.
[18:46:36] <sjoerddebruin> More specific project chats :P
[18:46:37] <risler> jheald; yes, that is the most important thing (but not the only thing, we also want everything in the optimal place too)
[18:46:49] <sjoerddebruin> (but enough taking hostage of the SDOC discussion)
[18:46:56] <Keegan> If things wind up faster, that's great, but it's not a part of the grant
[18:47:27] <risler> we will try very hard not to make things slower :)
[18:47:30] <BrillLyle> @Keegan - interesting
[18:47:37] <jheald> risler: so what spaces of items are in consideration for CommonsData items, apart from items for images?
[18:48:18] <risler> jheald: all types of media that Commons supports
[18:48:36] <jheald> risler: yes, but other than files?
[18:49:12] <BrillLyle> Going back to the multilingual support, I am curious how that works a bit more. I am going to be uploading public domain images that might have multiple languages in their titles and definitely descriptions
[18:49:38] <BrillLyle> Like an image could have a name in Spanish = English
[18:49:45] <risler> jheald: our focus right now is files. if the opportunity presents itself, we can explore other types of items, but it's not the focus
[18:50:24] <jheald> risler: so, the ontology is basically decided then. It's wikidata items + properties on files.
[18:50:51] <Keegan> BrillLyle: Could you provide a use case?
[18:50:57] <pigsonthewing> Other than (possibly) creator, it's hard to imagene concepts that should be items on Commons, but not Wikidata. Any counter examples?
[18:51:00] <BrillLyle> Oh wait, the captions would be able to be in multiple languages. I am confused as this seems to overlap with Wikidata. Too bad there's no Wikidata popup where we could update the description in both places at once
[18:51:15] <VIGNERON> for multilingual support, there is also multilingual SVG like https://commons.wikimedia.org/wiki/File:Wikidata_nodes_in_white.svg (where all translations are stored in the same file)
[18:51:23] <risler> jheald: that's just the very high-level ontology. there's a lot of detail that needs to be filled in :)
[18:51:30] <BrillLyle> https://en.wikipedia.org/wiki/List_of_Colecci%C3%B3n_Patricia_Phelps_de_Cisneros_publications#Conversaciones/Conversations_series
[18:51:51] <jheald> pigsonthewing: the big one is categories of course, but you also mentioned non-notable creators
[18:51:53] <BrillLyle> It's books here but images might have the same multi-lingual convention
[18:52:00] <NotASpy> yeah, per pigsonthewing, I don't see that there's going to be a massive number of properties needed for Commons. We need to keep it as minimal as possible.
[18:52:13] <Lucas_WMDE> VIGNERON: O_O I had no idea that was possible in SVG
[18:52:21] <Lucas_WMDE> *nice*
[18:52:29] <NotASpy> I'd like to see keywords deployed as per Flickr, so search can be considerably less shit
[18:52:35] <pigsonthewing> NotASpy: I'm talking about items, not properties
[18:53:00] <yannf> my point is that for Commons, we can't separate notable and not notable authors
[18:53:01] <VIGNERON> Lucas_WMDE: it's possible, but not well known on Commons (sadly since it's quite powerful and efficient)
[18:53:12] <risler> NotASpy: we'll have something similar to keywords with depicts properties, which will be a major thing for Search in the future
[18:53:14] <Keegan> We have less than ten minutes to go, if there are any further questions or need for clarification, or anything asked we haven't answered yet
[18:53:29] <pigsonthewing> jheals: but most categories are unions of two parent concepts
[18:53:31] <yannf> we need a solution which include all
[18:53:42] <pigsonthewing> "jheald", sorry
[18:53:42] <Lucas_WMDE> VIGNERON: I wonder if MediaWiki handles it correctly? giving you a PNG fallback in the right language depending on ?uselang ?
[18:53:44] <jheald> And btw, what is supposed to be the difference between 'filename' and 'title' and 'caption' and 'description' ? Do we have use-cases up anywhere
[18:53:54] <Keegan> I'll also take a moment to promote signing up for a SDC focus group, which receives short notifications when we have new requests for feedback posted https://meta.wikimedia.org/wiki/Global_message_delivery/Targets/Structured_Commons_focus_group
[18:54:33] <sjoerddebruin> Wikimania plans, even while the conference is focused on gaps?
[18:54:45] <jheald> pigsonthewing: that may be critical information to store in structured format, for import of information into Commons
[18:54:47] <VIGNERON> Lucas_WMDE: yes, it works quite well (not sure about the ?uselang it's probably override by the "lang=" in the code)
[18:55:27] <pigsonthewing> ...still wondering what "CiR" means
[18:55:29] <Keegan> jheald: a filename is the the name that a file is stored under. A title is the proper title of a work. The description is text related to all information that may be needed to describe a file. A caption is a short version of a description, with only the most relevant details to the work
[18:55:32] <spinster> BrillLyle: technically you will be able to enter a title in two or more languages; and the same for descriptions.
[18:55:33] <Keegan> Basically.
[18:55:41] <NotASpy> pigsonthewing: but Commons will use a vast array of items. You're going to have millions of things like https://www.wikidata.org/wiki/Q17771802 (in fact, we already do)
[18:55:47] <jheald> pigsonthewing: but there are other nice add-ons that could be done for categories as well, with structured data -- eg default image sort order, & meaningful alternative sort orders
[18:55:54] <risler> CiR = Competence is Required?
[18:55:57] <Keegan> (five minutes remain)
[18:56:41] <BrillLyle> @spinster -- Oh that's great!
[18:56:44] <pigsonthewing> risler: TY.
[18:56:58] <pigsonthewing> jheald: I'll need time to ponder that one.
[18:57:53] <yannf> BrillLyle, example: https://www.wikidata.org/wiki/Q34376297 notable or not?
[18:58:13] <pigsonthewing> jheald: could be more linke a page, with a listeria ruleset
[18:58:18] <jheald> spinster: have you had feedback back from GLAMs about eg metadata for engravings, including book engravings ?
[18:58:22] <yannf> he is a WP contributor
[18:58:23] <pigsonthewing> s/linke/like
[18:58:49] <BrillLyle> @yannf I would err on the side of inclusion but that's not really the discussion here
[18:58:58] <NotASpy> looking at https://www.wikidata.org/wiki/Q17771802 as an example, there's going to have to be a better way of handling the image statements for items, as you're going to potentially be linking every file for that building to the Wikidata item, or is it just going to loop back round to the relevant Commons category ?
[18:59:05] <spinster> jheald: not yet. I'm aware of how complex that is... will be interesting to model
[18:59:10] <NotASpy> but what if there's not a Commons category ?
[18:59:10] <Keegan> Okay, we will wrap things up here for the "official" hour. People are welcome to continue their discussions here or (preferably) on-wiki.
[18:59:34] <VIGNERON> yannf: obviously notable (but not sure if it WP identity is really public, even if it's well known in the community, privacy again...)
[18:59:41] <pigsonthewing> While we're here: Shoutout for Template:Wikidata Infobox on Commons
[18:59:46] <Fae> As cat sorting was mentioned, I'll mention "curation". Maybe someone can raise that at the right time...
[18:59:47] <VIGNERON> thank you Keegan and spinster !
[18:59:52] <Keegan> Thanks to all for the discussion, good stuff to think about as this project continues to move forward. We'll have another one of these in a month or two when there are fun new things to discuss
[18:59:56] <pigsonthewing> https://commons.wikimedia.org/wiki/Template:Wikidata_Infobox
[18:59:58] <spinster> Thanks, everyone, for your questions, input and feedback :-)
[19:00:02] <jheald> pigsonthewing: the listeria ruleset for a category would be nice, but it may not be practical to generate on the fly -- excluding the images in subcategories can be a lot of very big join operations
[19:00:12] <spinster> And for the great input on community consultations on-wiki so far
[19:00:20] <Keegan> #endmeeting