[08:47:09] Jonas_WMDE: Hey, one question regarding WDQS. http://tinyurl.com/je6a6nt [08:47:27] this is an example. Is there a way to change var name for label to something else? [08:47:49] It seems it automatically assign "var"+Label to it [09:14:37] the wikibase label service uses this pattern [09:14:52] you could also use rdf label [09:15:08] or ( x as y ) in the select part of your query [09:15:40] Okay, thanks [12:54:46] Jonas_WMDE: https://gerrit.wikimedia.org/r/298755 [16:43:16] DanielK_WMDE: hm… re WikiPageEntityMetaDataAccessor::loadRevisionInformationByRevisionId [16:43:45] why should the second parameter default to 0? [16:54:32] hoo: because 0 means "current revision" [16:55:08] hm, but we explicitly changed that in the past [16:55:25] if I want the current revision, I still want to be able to tell how that should be determined [16:55:41] yes. now we have a separate parameter for that, no? [16:55:47] hm [16:55:54] if i can specify the mode in two places, that's really confusing... [16:56:07] what happens if the two contradict? what does it mean? [16:56:22] Well, as the documentation says, the second is ignored, if the first is given [16:56:58] "Will be ignored if a mode has already been given as second parameter." [16:57:30] hoo: i find that really confusing. why not make it simple and use the second param only for the rev id (or 0), and the third only for the mode? [16:57:57] (we should have an Assert that makes sure the second param isn't a string) [16:58:19] I could do that [20:51:27] Why is poor sourcing such a persistent problem on Wikidata? [20:51:57] The lack of property suggestions imo [20:52:02] People don't know how to begin. [20:52:17] And all the Wikipedia imports didn't make stuff beter. [20:52:33] I think it's going to bite us, and soon [20:53:18] Because integration with English Wikipedia, which is very proud of verifiability, is not going to happen [20:53:23] https://phabricator.wikimedia.org/T77972 could use more support [20:53:35] if it means flooding Wikipedia articles with unsourced information from Wikidata [20:54:19] e.g. I feel compelled to say this is a bad idea: [20:54:19] https://en.wikipedia.org/wiki/Wikipedia:Village_pump_(technical)#.5BIdea.5D_Wikidata_descriptions_to_help_disambiguate_article_topic_on_mobile_web [20:54:43] True, there is a lot of unpatrolled vandalism. [20:56:11] not even vandalism, but the long legacy of importing statements from Wikipedia and then feeding them back ... sources got lost in translation [20:58:34] Yeah, trying hard to improve that though. [21:00:04] I'm not very active on Wikidata. How is this being addressed? [21:00:47] Well I've been just adding sources to items I encouter. [21:02:22] Manually? I guess about 90-95 percent of statements are unsourced. There needs to be some automatic process if Wikidata expects to come anywhere near the level English Wikipedia requires it to be [21:03:51] 62.64% of our statements contain references, 27.46% is Wikipedia though. [21:05:30] Yeah, that's the problem. [21:08:15] The unreferenced ones that are imported from Wikipedia without indication is the huge problem though. You can filter the ones that indicate that something was imported from Wikipedia out with LUA. [21:10:38] note though that authority statements usually don't need references since they are self-referential [21:10:57] so you need to exclude everything with external ID basically [21:11:49] e.g. if you have VIAF identifiier for an item, VIAF page itself is a source for that fact usually [21:12:28] Ah yes. [21:13:41] I think Wikidata should be restricted to statements like that. Authority control and interlanguage links are probably the only thing that works on Wikidata. [21:17:14] 2824 of the 3613 statements for sexual orientation doesn't contain any source. https://www.wikidata.org/wiki/Wikidata:Database_reports/Complex_constraint_violations/P91 [21:19:58] WD doesn't have a Biographies of Living people policy? [21:20:05] sjoerddebruin: that's not good and potential trouble. I'm not sure if wikidata has policies like WP:BLP but having claims as touchy as sexual orientation on entries about actual persons without sourcing is inviting trolls and asking for trouble [21:20:37] https://www.wikidata.org/wiki/Wikidata:Requests_for_comment/Verifiability_and_living_persons [21:21:24] ouch [21:24:08] It's like people don't care at all. [21:24:15] looks like it went too far. But I then proposed marking properties with "must be referenced" [21:24:23] https://www.wikidata.org/wiki/Property:P91 [21:24:27] "use IF AND ONLY IF they have stated it themselves, unambiguously, or it has been widely agreed upon by historians after their death [21:24:28] " [21:25:24] Why don't you replicate the vague wording on Wikipedia? "Contentious material, whether negative, positive, or neutral, about living persons" must be sourced [21:25:25] i.e. while it may be too much to stomach for all properties, maybe for some? [21:25:35] BTW: do not forget https://www.wikidata.org/wiki/Property:P2559 [21:25:48] "No source for heterosexual people." [21:25:49] wtf [21:26:06] SMalyshev: oh, I had no idea that existed, nice [21:26:15] no source - no claim, for this property it's completely fine I think [21:27:17] Finnusertop: this was going too far as we have a lot of unsourced claims and defining "contentious" may prove.... well... contentious? [21:32:45] SMalyshev it defacto means: anything that is said about a living person excluding truisms ("the Pope is Catholic") [21:35:20] Finnusertop: really? would a date of birth be contentious? [21:35:31] I wouldn’t call it a truism, but usually there’s not a lot of debate about it either [21:38:24] I don't know how low the bar objectively is, but I don't think a date of birth without a source would survive for very long in a biography of a living person [21:38:51] We're almost at 20 million Wikidata entries! Exciting [21:40:59] WikidataFacts: One of the oldest disputes on Wikipedia is Jimmy Wales disputing the date of birth in his article. [21:41:04] It's come up a lot. [21:41:28] huh, I never heard of that, sounds like something I should’ve heard of :D [21:41:33] SMalyshev: Are you familiar, in a general sense, with what causes WDQS to not update with new information? [21:41:52] Most of the time it updates just fine; then there are these sticky things. [21:42:17] hare: example [21:42:18] ? [21:43:14] Hmm, looks like my one example eventually fixed itself. [21:44:06] Maybe everything eventually fixes itself. [21:44:51] Finnusertop: My general perspective on this is that Wikidata is a baby and will probably, in the long term, be more usable. Wikidata is now where Wikipedia was in 2004; Wikipedia sucked in 2004. [21:46:50] True, and given that, we should move considerably slower in terms of Wikidata Wikipedia integration [22:15:07] hare: good :)