[08:40:06] Hi, I want to know how can I start creating a wikipedia page with detailed tables and links on them. I want to create a page for the company I work for. [09:08:02] Lydia_WMDE: https://en.wikipedia.org/wiki/Wikipedia:Stub [09:13:08] SMalyshev: am I missing something, or is there no simple “query” service in the new WDQS MWAPI? [09:13:41] as far as I can tell, I can find all pages in a certain category with the Generator service, but then I can’t get any information about those pages (e. g. revision info) [09:34:50] hello, can an admin delete the newly created item please Q30250227? shortly after creation, I noticed it existed already [14:13:46] What would be the right place and who could answer a couple of Wikibase installation question? [14:15:24] Thiemo_WMDE: did this newbie committer address your concerns? https://gerrit.wikimedia.org/r/#/c/336052/ [14:25:28] Specifically, we tried to install wikibase ourselves, and for some unknown reasons most of the edit buttons don't show up in http://88.99.39.75/wiki/Item:Q2 - no messages in logs which would give clue why [14:25:56] Oh, I'm logged in :) [14:27:59] Does any property exist yet? [14:28:09] Nemo_bis: yes susannaanas created one [14:28:13] P2 [14:28:24] http://88.99.39.75/wiki/Property:P2 [14:30:11] Hmm external identifier, dunno what such properties entail [14:33:39] That should not be relevant. The problem is that there is no edit button in the interface to link that property to the item. Are we missing some js, settings, privileges... [14:34:20] Nemo_bis: Thanks for asking. Unfortunately the answer is no. The lengthy comment contains a lot of arbitrary numbers, but no explanation why stuff must break at this breakpoint. The entire patch is nothing but that: a bandaid patch on top of bad CSS that is already there. We need to redo the CSS instead. [14:34:51] Lucie is working on this. [14:35:33] Thiemo_WMDE: ok, thanks; maybe worth telling the user so that they don't waste more time on that [14:40:39] In addition to the previous question I would like to find pointers - I remember reading new Wikibase properties could be made to take advantage of Wikidata properties. Where could I find informations about that? [17:28:31] DanielK_WMDE_: if you have a minute, would like to hear your opinion on T167759 [17:28:31] T167759: Reference hash is not stable - https://phabricator.wikimedia.org/T167759 [17:28:49] also if you know whether anything in the data model changed since April [17:50:42] SMalyshev: this is a known issue, and yes, we changed the hash. I think we also announced it. [17:51:00] I'd very much like to completely change how references work, but it's not likely to happen soon [17:51:36] actually, you can see when we messed with the hashes by looking at the history of the rdf test data [17:52:03] when we change hash calculation (e.g. of some value type), we have to update the test data [17:52:30] when we did this a few months ago, i also refactored the test data, so we don't have to update as many files [17:53:24] I see [17:53:46] but I am concerned that the hash depends on class names [17:54:01] and also order of properties in the class... [17:56:07] yeah I see commit 11b4a7e27ba70db64bb698da8a90d0382b78cde8... ok, I guess we will have to reload data for T131960 anyway [17:56:09] T131960: "_" character encoded as %20 in Wikidata URI RDF serialization - https://phabricator.wikimedia.org/T131960 [17:56:15] so that will be taken care of [17:56:51] DanielK_WMDE_: but I think we need to have some notification - e.g. create phab task for data reload - when such things happen. [17:57:33] I see I wasn't even on that ticket, so that probably explains why I missed it [18:05:51] SMalyshev: yes, you are right... I *though* we discussed the requirement for a reload internally... I wonder how we missed to ping you. Mayb we though an email to the list would be enough. [18:06:39] SMalyshev: i will try to pay better attention to this in the future. perhaps write an email about this too - but i'm not sure where to send it. wikidata-tech? [18:06:43] DanielK_WMDE_: sure, any means of ping is fine. But I want to know each time those tests are patched [18:07:09] I'll write to wikidata-tech [18:08:10] SMalyshev: would be nice to have a comment in these files - "if you change this, notify triple store owners of a breaking change!" [18:08:22] but n-triples doesn't have a syntax for comments... [18:08:27] we could invent one. hm... [18:08:30] DanielK_WMDE_: but the question of it depending on class names remains [18:08:38] not just class name [18:08:39] DanielK_WMDE_: maybe just README file in the directory? [18:08:41] order of private fields [18:08:44] all kinds of crap [18:08:45] right [18:09:07] even changing a field from private to protected will break the hash [18:09:27] nobody reads readmes :) [18:10:13] well, yeah, but better than nothing :) [18:10:30] also a good thing to point to and say "I told you so!" [18:10:37] make a patch? [18:10:43] ok I will add it [18:11:03] would we classify it as breaking change? [18:12:18] I think when we discussed it, we decided against that - hashes are not stable IDs, and should not be persisted externally by 3rd parties. [18:12:23] but it would be a "significant" change [18:12:48] ok, I'll add note to that effect [18:12:52] it is significant in that it requires a re-import of the data. [18:12:59] but it doesn't require a any code to be changed [18:43:08] SMalyshev: btw, i'm trying to push the testing of the cirrus integration stuff. it has been taking a back seat since we are trying to get the wiktionary demo ready for wikimania, but i really want to see it happen soon [18:43:19] ping lydia and amir about it every now and then :) [18:43:23] * DanielK_WMDE_ is goign home now [18:43:24] DanielK_WMDE_: same here :) [18:43:30] thanks! [19:11:58] Hi! I'd like to ask, are there any items that consist of both properties "instance-of" and "subclass-of"? [19:14:24] glorian_wd: yes. [19:14:38] lots of them: http://tinyurl.com/y7xcuysh [19:14:42] could you give me an example sjoerddebruin? [19:14:45] how many of them *should* have both is a different question ;) [19:14:55] Most occupations have one afaik [19:15:07] https://www.wikidata.org/wiki/Q386854 for example [19:15:38] Alright. Thanks WikidataFacts, sjoerddebruin [19:15:42] :) [19:15:42] there’s even items that are instances and subclasses of the same class, which is very dubious IMO: http://tinyurl.com/y8ct8jjw [19:16:35] WikidataFacts : i agree with sjoerddebruin, the question wasn't "which", but "are there any", so the answer is https://query.wikidata.org/#ASK{%0A%3Fitem%20wdt%3AP31[]%3B%0Awdt%3AP279[].%0A} => true :-p [19:17:10] indeed Alphos :) [19:17:16] "so is it a boy or a girl ?" - "yes" :-p [19:17:31] ooh, look at fancy-pants over there, with a non-SELECT query! :D [19:17:45] WikidataFacts eeeeeyup ! [19:20:39] WikidataFacts i'm a long-time avid fan https://phabricator.wikimedia.org/T145466 who initially found a weird behaviour when asking when Louis the 14th was right when he said "l'Etat c'est moi", "I'm the State" http://tinyurl.com/hmwd9dy (returned true, which is a blatant lie ^^ ) [19:20:57] s/asking when/asking whether/ [19:21:25] Alphos: lol [19:21:57] well I’m glad that’s fixed now, because we’re using a lot of ASK queries in the SPARQL constraints to be deployed soon… [19:22:04] would be awkward if they all returned true :D [19:22:06] besides, you call me a fancy-pants when you've been toying with regraphs for weeks now ! :p [19:23:11] regraphs? [19:24:11] WITH/INCLUDE [19:24:22] oh, those [19:24:26] ^^ [19:24:51] not my fault BlazeGraph’s optimizer is stupid ;) [19:25:46] haven't gotten to wrap my head around them just yet, and with summer coming it's unlikely to be done soon :-( [19:28:30] Q166907 is a Q878767. or, you know, ASK { Q166907 wdt:P31*/wdt:P279* Q878767 } #true [19:31:45] (plus the wd: where applicable :p ) [19:32:44] Alphos: you speak french, right? [19:32:49] yep [19:33:20] also english (duh), and bits of russian and german, if needed be [19:33:29] do you know if there's a reason why the most common french description for disambiguation pages doesn't match the label of https://www.wikidata.org/wiki/Q4167410? [19:34:22] (the most common description is "page d'homonymie d'un projet Wikimédia" but the label is "page d'homonymie de Wikimedia") [19:34:57] no idea whatsoever [19:38:25] aww, ok [19:42:38] there's a lot of variation in the descriptions used... somehow people have managed to say the same thing in over 50 ways [19:45:20] well, it's the human-readable parts, what did you expect ? :p [19:46:14] let people have their whim, they won't disappoint :D [19:47:59] there's *also* plenty with other stuff :P I'm just impressed at how many slight variations of the usual wording there are [19:52:18] I can't even see the difference between some of them [19:53:41] i'd check the apos ;p [19:55:15] (in short, mac ’ U+2019 vs windows ' U+0027) [19:55:31] those look different in the fonts my programs are using [19:55:45] I think it's just that my eyes are going funny looking at too many almost identical strings :P [19:56:12] levenshtein ! :p [19:56:24] or better yet, frame alignment ! :p [19:57:11] https://pastebin.com/dhpjPPas is the list of unique descriptions for disambiguation pages for french (plus the number of times they've been used) if you're curious [19:57:47] I'd link a sparql query but it doesn't like doing queries like that >_< [19:58:58] uc vs lc, wikimedia vs wikipedia, wikipédia vs wikipedia, wikimedia vs wikimédia [19:59:30] plus U+2019 vs U+0027 [20:00:22] yeah [20:00:55] (18 [20:01:00] (Sorry) :) [20:01:16] so that's 5 bits of entropy, 32 possible choices :p [20:01:42] add "de wiki(p|m)" vs "wiki(p|m), that's 6, so 64 [20:06:50] nikki : "page d(\u0027|\u2019)homonymie (|(|de|d(\u0027|\u2019)un projet) (W|w)iki(p|m)(é|e)dia)" give or take the spaces :p [20:09:43] * WikidataFacts throws https://tedclancy.wordpress.com/2015/06/03/which-unicode-character-should-represent-the-english-apostrophe-and-why-the-unicode-committee-is-very-wrong/ into the ring [20:10:44] Question. Every now and then, it seems that some properties are skipped while creating new ones. For example [[Property:P4049]] while Property:P4048 and [[Property:P4050]], Property:P4051 etc do exist. Why is that? [20:10:45] 10[1] 04https://www.wikidata.org/wiki/Property:P404913 => [20:10:48] 10[2] 10https://www.wikidata.org/wiki/Property:P4050 [20:10:59] WikidataFacts there is still no appropriate char whatsoever for the french typographic apos [20:11:12] Alphos: :( [20:11:28] yeah, there’s a bit about it in the comments [20:11:31] WikidataFacts on the bright sides, we're getting new emojis soon ! [20:11:41] the most important thing, obviously! [20:11:48] inorite ! [20:12:10] Rodejong because someone created a property, made a mistake, deleted it, and created a new one, i guess [20:12:52] If it was deleted, it would have shown. This seems like it never was created. [20:13:31] See [[Property:P7]] [20:13:32] 10[3] 04https://www.wikidata.org/wiki/Property:P7 [20:17:46] Alphos: https://youtu.be/5OPkGQoPeHk?t=5m22s [20:25:49] WikidataFacts ^^ [20:36:17] WikidataFacts: at least it's the right shape... I keep seeing signs with a german single closing quote for an apostrophe [20:36:31] bleh [20:36:54] but that’s the same as the English opening one, why do you blame German? ;) [20:37:16] because the signs are in germany :P [20:38:19] ah :D