IRC office hours/Office hours 2019-07-16

From Meta, a Wikimedia project coordination wiki

Wikidata IRC Office hour[edit]

Office hour on #wikimedia-officeconnect on July 16, 2019.

[16:01:11] <Lydia_WMDE> #startmeeting
[16:01:28] <Lydia_WMDE> ah this bot and me are never going to like each other :D
[16:01:36] <Lydia_WMDE> Hey everyone!
[16:01:46] * Lucas_WMDE starts wm-labs-meetbot`with a trout
[16:01:47] <Lydia_WMDE> Who's here for the Wikidata office hour?
[16:01:49] <Lucas_WMDE> hi! o/
[16:01:53] <melderick> \o/ Lydia
[16:01:53] <abian> o/ :)
[16:02:00] <Lydia_WMDE> yay :)
[16:02:28] <Lydia_WMDE> Good to have you here :)
[16:03:11] <Lydia_WMDE> As usual I'd start with an overview of what's been going on over the past 3 months and then what's coming up. Ask away if you have questions.
[16:03:33] <Lydia_WMDE> Quite a few things happened on the tech side of things.
[16:03:47] <Lydia_WMDE> We now have support for schemas on Wikidata with ShEx \o/
[16:04:40] <Lydia_WMDE> Now you can store a machine-readable description o f how certain parts of the data should look like and then test a set of Items against it to see which ones need attention.
[16:05:08] <Lydia_WMDE> I hope this will help us quite a bit with getting to more consistent modelling beyond what's currently possible with constraints.
[16:05:41] <Lydia_WMDE> As with the constraints the schemas are not enforcing so you can still enter data that doesn't fit.
[16:06:02] <Lydia_WMDE> Has anyone tried testing some Items against a schema already?
[16:07:11] <melderick> not yet ^^
[16:07:18] <Lydia_WMDE> If not I highly recommend this one: https://www.wikidata.org/wiki/EntitySchema:E33 :D
[16:07:24] <Lydia_WMDE> They are <3
[16:08:00] <Vigneron> I've looked at them but didn't find a use yet
[16:08:22] <Lydia_WMDE> Vigneron: specify all the artwork modelling!!!!111eleven :D
[16:08:56] <Vigneron> I was thinking about monuments fro WLM ;)
[16:09:05] <abian> https://www.wikidata.org/wiki/Help:P131
[16:09:14] <Lydia_WMDE> Or that
[16:09:19] <abian> And this is another use case :)
[16:09:25] <Lydia_WMDE> ohhh yeah
[16:09:27] <abian> We should have one Schema for each country
[16:09:59] <abian> That diagram is too twisted and useless :)
[16:10:14] <Lydia_WMDE> The other big thing that happened is that the development of the Wikidata Bridge has started for real. The first lines of code are written.
[16:10:57] <Lydia_WMDE> That'll get us the ability to edit Wikidata from Wikipedia once it's ready.
[16:11:09] <Lydia_WMDE> You can see the start of the documentation here: https://www.wikidata.org/wiki/Wikidata:Wikidata_Bridge
[16:11:18] <Lydia_WMDE> and here https://www.mediawiki.org/wiki/Wikidata_Bridge
[16:12:15] <Lydia_WMDE> It'll still take some time but our plan is to have something ready to test very early on so you can give feedback as it develops.
[16:13:11] <Lydia_WMDE> And finally Wikimedia has it's own URL shortener \o/ That means you can finally paste shortlinks from the query service on Wikimedia wikis.
[16:13:54] <Lydia_WMDE> We also have something for the eyes: a new dashboard showing the landscape of external identifiers that Wikidata has: https://wmdeanalytics.wmflabs.org/WD_ExternalIdentifiersDashboard/
[16:14:57] <Lydia_WMDE> The constraints also got some love: There is now a third level called suggestions for constraints that are used mainly to make a recommendation to the editor as opposed to alerting them to something that they did wrong.
[16:15:47] <Lydia_WMDE> Think of something like "if someone has a date of birth statement you probably also want to add a place of birth statement"
[16:16:57] <Lydia_WMDE> And we're now also using the constraints to give better suggestions when making statements that have a limited value set. One example is driving side. When making a new statement for driving side you now get left and right as suggestions based on the one-of constraint definition.
[16:17:15] <James_F> Nice.
[16:17:55] <Lydia_WMDE> More behind the scenes we worked more on migrating away from the wb_terms table to keep up with growth.
[16:18:25] <Lydia_WMDE> And we continued pushing for getting the new termbox (the box with labels, descriptions, aliases) for mobile out. Any day now. Hopefully.
[16:18:56] <Lydia_WMDE> Oh and there is a new dashboard where you can see how many edits are done in which namespace: https://grafana.wikimedia.org/d/000000170/wikidata-edits?refresh=1m&orgId=1&panelId=13&fullscreen
[16:19:21] <Lydia_WMDE> Alright. On to non-dev things then.
[16:19:33] <Lydia_WMDE> We have a new admin: BRPever \o/
[16:20:03] <Lydia_WMDE> There were plenty of Wikidata workshops and meetups all around the world. <3
[16:20:30] <Lydia_WMDE> If you are organizing one please don't forget to add it to the weekly summary so others can know about it as well.
[16:20:51] <Lydia_WMDE> There is a link to the page where you can add things to the next weekly summary at the end of every weekly summary.
[16:21:33] <Lydia_WMDE> While we're on the topic of events: WikidataCon is shaping up nicely and the program committee is working on putting together the program.
[16:21:45] <Lydia_WMDE> https://www.wikidata.org/wiki/Wikidata:WikidataCon_2019 has info about timelines and everything.
[16:22:11] <Vigneron> \o/
[16:22:56] <Lydia_WMDE> And of course cool things happened outside Wikimedia as well. For example the Library of Congress is not integrating Wikidata IDs in their authority files: https://blogs.loc.gov/thesignal/2019/05/integrating-wikidata-at-the-library-of-congress/
[16:23:15] <Lydia_WMDE> THis whitepaper was published: https://www.arl.org/resources/arl-whitepaper-on-wikidata/
[16:23:46] <Lydia_WMDE> There is an article about Wikidata in Mozilla's Internet Health Report: https://internethealthreport.org/2019/wikidata-gives-wings-to-open-knowledge/
[16:24:17] <Lydia_WMDE> https://blog.wikimedia.de/2019/05/09/new-testing-ground-for-wikibase-a-federal-agency-goes-on-an-expedition-in-the-wiki-universe/ is a nice overview of the work at the german national library with Wikibase
[16:24:31] <Envlh> LoC: s/not/now
[16:24:50] <Lydia_WMDE> heh yes, sorry :D
[16:24:59] <Envlh>  ;)
[16:25:56] <Lydia_WMDE> And there are two essays related to Wikidata currently in review for a book for Wikipedia's 20th anniversary. They are by Denny and me. I'd love to have your feedback: https://wikipedia20.pubpub.org/
[16:26:44] <Lydia_WMDE> There are also a number of cool new tools that are worth mentioning:
[16:26:58] <Lydia_WMDE> For example https://opentapioca.org which does entity linking with Wikidata entities
[16:27:26] <Lydia_WMDE> https://tools.dicare.org/projects/ looks at how "related" the Wikimedia projects are to each other
[16:27:52] <Lydia_WMDE> https://qanswer-frontend.univ-st-etienne.fr/ now also answers questions related to Lexemes
[16:28:17] <Lydia_WMDE> https://tools.wmflabs.org/hauki/ lets you browse lexicographical data
[16:28:40] <Vigneron> Lexemes for the win \o/
[16:28:53] <Lydia_WMDE> indeed :P
[16:28:55] <Lydia_WMDE> https://www.wikidata.org/wiki/User:Teester/CheckShex.js let's you check if the Item you are on fits a certain Schema
[16:29:09] <Lydia_WMDE> Schemas for the win!
[16:29:49] <Lydia_WMDE> And there is a new python package to work with Wikidata entities and dumps:
https://lists.wikimedia.org/pipermail/wikidata/2019-April/012929.html
[16:30:46] <Lydia_WMDE> Quite a few cool things also came out of the last hackathon in Prague. For example https://tools.wmflabs.org/integraality/ which lets you build nice dashboard to get an overview of how complete a certain part of the data is
[16:31:05] <Lydia_WMDE> Like this one:https://www.wikidata.org/wiki/Wikidata:WikiProject_sum_of_all_paintings/Property_statistics/Sandbox
[16:31:40] <Lydia_WMDE> The complete project list is at https://www.wikidata.org/wiki/Wikidata:Status_updates/2019_05_20#Other_Noteworthy_Stuff
[16:32:03] <Lydia_WMDE> Alright that brings us to what's coming next
[16:32:38] <Lydia_WMDE> We'll continue coding on the Wikidata Bridge and you'll see first prototypes on test wikis to give feedback.
[16:32:52] <Lydia_WMDE> We'll finally publish the new termbox for mobile users.
[16:33:27] <Lydia_WMDE> We'll continue reseraching/thinking about Signed Statements and what suitable next steps for that are.
[16:33:41] <Lydia_WMDE> And last but not least Wikimania is coming up \o/
[16:34:24] <Lydia_WMDE> There will be quite a few Wikidata things. A number of talks have been accepted and we'll have a meetup and there will be a pink pony session again to express all your special pink pony wishes :D
[16:34:33] <Lydia_WMDE> Who's coming to Wikimania?
[16:34:39] <Envlh> \o
[16:34:53] <abian> o/
[16:35:03] <Vigneron> \o/
[16:35:04] <Lydia_WMDE> Yay :) See you there!
[16:35:31] <Lydia_WMDE> Alright. That was it from my side. Do you have questions? Things we should discuss/talk about?
[16:36:28] <Vigneron> random question: does Shex works on Lexemes?
[16:36:49] * Lydia_WMDE looks @ Lucas_WMDE but thinks they do
[16:38:15] <Lydia_WMDE> Ah I have a question!
[16:38:42] <Lydia_WMDE> What are you working on atm around Wikidata? Time to brag.
[16:40:14] <melderick> just started to have a look on Noble Titles : it is a mess :D
[16:40:14] <Vigneron> nothing special, preparing WLM (checking mostly) and some Lexemes
[16:40:42] <Vigneron> melderick yes it is! thank you for looking into it ;)
[16:40:44] <Lydia_WMDE> melderick: uhh that does sound like an area with a lot of mess-potential ;-) Good luck!
[16:41:20] <Lydia_WMDE> Vigneron: Wiki Loves Monuments is really cool. What are you preparing on Wikidata for it?
[16:41:22] <Lucas_WMDE> Vigneron: it should work, you’ll just need to work with the lexeme RDF model
[16:41:54] <melderick> first mess is : i find things like "count of XXX" and "county of XXX" both in the same hierarchy of classes ^^;
[16:42:12] <Vigneron> Lydia_WMDE checking that all monuments does have an item with some minimal informations, mainly adding the 2018 "newly protected" monuments
[16:42:23] <Lydia_WMDE> nice
[16:44:00] <melderick> Vigneron: this seems exactly what schemas are good at, if i understood them correctly
[16:44:21] <Lydia_WMDE> jep
[16:44:49] <Vigneron> indeed, that's why I need to look into Schema! (but meanwhile SPARQL query are good enough for our basic needs ;) )
[16:45:19] <Lucas_WMDE> …I think I just learned where the word “county” comes from :D
[16:46:42] <James_F> Lucas_WMDE: Ha. :-)
[16:47:12] <melderick> Lucas_WMDE : really ? :)
[16:47:32] <Lydia_WMDE> Oh speaking of quality... I forgot something. We're trying to make ORES' judgements better. This way the coloring in recent changes and watchlist for example that tells you which edits are likely vandalism will be much more accurate. In order to do that we need to train the system some more. If you have some minutes please help make the machines smarter by teaching them here: https://www.wikidata.org/wiki/Wikidata:Edit_labels
[16:47:32] <Lydia_WMDE> :D
[16:48:03] <Lucas_WMDE> I just never made the connection! (though „Grafschaft“ exists in German as well as a fairly literal translation)
[16:48:10] <abian> Okay :)
[16:48:34] <abian> A question: do we have any way to know if an entity violates any applicable Schemas (without having to know all the Schemas)?
[16:48:35] <Lydia_WMDE> https://labels.wmflabs.org/stats/wikidatawiki/ shows statistics of how much is done. Currently at 61%
[16:48:52] <Lydia_WMDE> abian: I think not yet.
[16:50:05] <abian> "yet" :D
[16:50:39] <Lydia_WMDE>  ;-)
[16:50:45] <abian> And about the Bridge... what kinds of changes to Wikipedias are expected when it's available?
[16:51:20] <melderick> Lydia_WMDE : when faced with a live vandalism, is there some way to mark it (like with a special comment) so that it helps ORES as well ?
[16:51:29] <halfak> o/
[16:51:38] <halfak> Hey! I'm the engineer/researcher behind ORES.
[16:51:47] <halfak> The answer to melderick's question is: soon.
[16:52:06] <halfak> We're working on a collaboratively editable system to allow you to train ORES while you work. It'll be called Jade.
[16:52:15] <Lydia_WMDE> Right now a lot of the infoboxes that make use of Wikidata add a little edit pen that leads to Wikidata. With the Wikidata Bridge this will open a window directly on the Wikipedia and then people can edit there without having to go to Wikidata and without having to understand all the things there.
[16:52:20] <melderick> halfak : ok :) i'll continue adding a comment such as "obvious vandalism" :)
[16:52:31] <halfak> It's designed to allow normal Wiki mechanisms deal with "label damage" where someone falsely flags something as bad/good to mess with ORES.
[16:52:36] <Lydia_WMDE> The template editors will need to make some small changes to the template to enable it for that template
[16:52:49] <halfak> melderick, sounds good. We do look for those kind of edit comments in post processing and analysis.
[16:53:32] <abian> Lydia_WMDE: Cool, thanks!
[16:53:33] <melderick> halfak : perfect. thx :)
[16:53:39] <halfak> FYI: https://www.mediawiki.org/wiki/JADE
[16:54:54] <melderick> Lydia_WMDE: will there be some enforcement that Wikipedia editors add a reference to any edit they do ? Since they always complained Wikidata is full of unreferenced data
[16:55:16] <melderick> through Wikidata Bridge I mean
[16:55:38] <Lydia_WMDE> melderick: we'll definitely look into nudging for references but the details are not clear yet. It'll also depend on feedback from people.
[16:55:40] <Vigneron> melderick oh no, please no...
[16:56:16] <Lydia_WMDE> I'd rather not make it absolutely mandatory because there is just a lot of data that's not really referenceable or worth referencing - like a lot of instance of statements.
[16:56:17] <melderick> Vigneron: why not ?
[16:57:19] <Vigneron> because most statements doesn't/can't have statements, it shouldn't be mandatory, and especially not in a "revenge" spirit
[16:58:01] <Vigneron> so "nudge" is ok of course (and welcome), but not "enforcement" ;)
[16:58:44] <melderick> Vigneron : ah not in revenge spirit :) but ok for a "nugde" :)
[16:58:44] <Lydia_WMDE> I think there is a lot we have to try and see how it works out in the end and then adjust based on feedback and data.
[16:58:55] <Lydia_WMDE> There is a lot of unknown territory
[16:59:47] <Lydia_WMDE> Alright. We're at the end of our hour. Does anyone have any more questions or should we wrap it up?
[16:59:48] <melderick> true and this is exciting :D
[16:59:53] <Vigneron> Lydia_WMDE maybe you can look into how people use WEF (Wikidata Editor Framework, which also allows to edit from wikimedia projects ;) )
[17:00:02] <Lydia_WMDE> Yeah definitely
[17:00:33] <abian> The Bridge can be an exciting expedition full of surprises O:)
[17:00:39] <Lydia_WMDE> haha
[17:00:50] <Lydia_WMDE> the team can already tell many stories about that...
[17:01:03] <Lydia_WMDE> We already had a lot of those. More to come i'm sure.
[17:02:26] <Lydia_WMDE> Ok if there are no more questions let's finish this office hour for today.
[17:02:30] <Lydia_WMDE> Thanks everyone for coming :)
[17:02:34] <Lydia_WMDE> <3
[17:02:40] <melderick> thank you Lydia :)
[17:02:43] <abian> Thank you all :)
[17:02:50] <Lydia_WMDE> \o
[17:03:10] * Lucas_WMDE waves
[17:03:24] <Lucas_WMDE> Lydia_WMDE: do you want to risk an #endmeeting? :P
[17:03:32] <Lydia_WMDE> haha
[17:03:32] <Lydia_WMDE> sure
[17:03:36] <Lydia_WMDE> #endmeeting
[17:04:27] <melderick> have a nice evening everyone :)