[04:15:34] I really want to send this image in response to the recent email in wikidata-l: https://resources.stuff.co.nz/content/dam/images/1/8/s/4/z/8/image.related.StuffLandscapeSixteenByNine.620x349.18qta3.png/1450320174494.jpg [04:15:44] Can I? pleaseeee :D [09:58:40] Alright, https://www.wikidata.org/wiki/Property:P1559 seems to have entered the suggestions. [11:42:07] WDQS accepts character properties in REGEX() ! \o/ [11:42:43] #nerdgasm ^^ [11:44:10] http://tinyurl.com/zywmj2p 150 Results in 12363 ms [12:39:19] So I noticed the Press Coverage page doesn't seem to have any updates since 2014 - would anyone disapprove of me updating it? Or have any thoughts on why it hasn't been updated? https://www.wikidata.org/wiki/Wikidata:Press_coverage [12:40:23] I feel like I'm still getting the hang of how to contribute, thought this might be a place to start :) [12:44:50] argotechnica i'm saying this as any regular peon : please do :) [12:46:17] haha thanks Alphos [13:11:28] Aleksey_WMDE: https://gerrit.wikimedia.org/r/#/c/334162/5/repo/includes/Api/CreateClaim.php [14:01:05] Hi. What's the best route to report bugs in https://github.com/addwiki/wikibase-api/blob/master/README.md ? There is no issue tracker on that project [14:02:25] clemenstolboom: readme says: Issue tracker: https://phabricator.wikimedia.org/project/profile/1490/ [14:04:11] Thanks. Me suffer line blindness ... in particular #2 :p [14:11:55] i think they make glasses for that, but they cost a pretty penny ! :-p [14:48:20] some help needed, https://www.wikidata.org/w/index.php?title=Property_talk:P3414 "allowed values" parameter should allow all letters, numbers, and some other marks like "-" and "/". What would be the correct regex? [14:50:12] [0-9a-z] ? [14:50:17] Stryn [0-9a-zA-Z/-]+ [14:50:42] okay, great thanks :) [14:51:03] Stryn keep in mind that only allows plain latin letters [14:51:16] it should be fine AFAIK [14:51:28] diacritics are excluded, russian letters are excluded, ideograms are excluded, … [14:51:41] as the url doesn't contain ä and ö [14:52:30] Stryn but are you sure the format isn't the one specified ? [14:52:38] / in id seems weird [14:53:04] see https://www.wikidata.org/wiki/Q28263556 [14:53:44] ok then [15:17:10] leszek_wmde: Time to dump testwikidatawiki went from 1m21.433s to 0m34.786s with the prefetcher fix. [15:17:14] Thanks for that [15:17:29] hoo: slightly faster [15:46:44] o/ glorian_wd [15:47:01] o/ halfak [15:47:12] So yeah... I was saying that I don't think we want to have "appropriate qualifiers and ranking" in the quality criteria below Showcase class. [15:47:23] Maybe B class. [15:47:28] But certainly not C class. [15:47:56] okay why? [15:48:21] Because that's a pretty high level of curation. We need to make room in the lower classes. [15:48:47] You could miss out on a class because a single qualifier was missing or done wrong. [15:49:22] I'd rather say, "Some half-asses attempt at qualifiers and ranking" than specify that it must be good in the mid/low quality classes. [15:50:10] *thinking* [15:51:55] You haven't modified the scale right? [15:52:00] based on my feedback [15:54:50] I haven't. I'm working on splitting up the discussion. It's really hard to discuss a big list like the one in your post. [15:56:07] okay [15:57:11] I think we should have "appropriate qualifiers and ranks" on "B" class. But, maybe not for C class. So, there's a clear distinction between B and C class. Am I correctly understand you? [15:58:10] if we also have those criteria on both B and C, maybe it is a bit hard to distinguish because the difference of those two classes is only the amount (i.e. good number of statements vs moderate amount of statements) [16:02:40] glorian_wd, maybe just "some qualifiers and ranks" [16:03:00] glorian_wd, that's not the wording re. statements [16:05:13] halfak, sorry for the big list :( [16:07:22] halfak, so you think that scale C should be: [16:07:22] The critical properties for this type of item have statements with: [16:07:22] 1. External references for some non-trivial statements [16:07:23] 2. Some appropriate ranks [16:07:23] 3. Some qualifiers where applicable [16:11:23] Still working on splitting things up. [16:12:56] glorian_wd, what are you talking about with "unique properties"? [16:13:09] I don't even know how to respond to that. [16:13:18] What properties aren't unique? [16:14:18] halfak: one moment, I grab something for you [16:16:11] See new threads with responses here: https://www.wikidata.org/wiki/Wikidata_talk:Item_quality#Grading_scheme [16:21:24] halfak, that refers to multichill's comment: "The number of unique P statements (this would filter out items which have a lot of statements of one property type)" [16:21:24] suppose you have an item instance of: person. It has only one property: birth date, and there are many values on it (this person has many birth date for some reasons). So, many statements with one property. I agree there should be a quality criterion for this. Does it make sense for you? [16:21:24] But you are right. I guess I have to fix my wording, sorry [16:22:14] glorian_wd, we no longer have any language about # of statements [16:22:23] So I don't see the problem with unique properties. [16:23:11] If you have instance-of:person with 20 birthdates and no birth place, that'd fail "The critical properties for this type of item have statements" [16:23:13] for C class [16:24:08] halfak: *checking* [16:34:54] halfak: let's put it this way. How about if that item has all appropriate properties (e.g. occupation, birthplace, etc), but it has 20 birth date and 20 occupations. Would it be categorized as "A" grade? [16:35:16] glorian_wd, I don't think so because we don't even reference the number of properties. [16:35:24] I don't see what's confusing about this. [16:35:36] "All appropriate properties for this type of item have statements" [16:35:56] So unless the item had all of the other relevant properties, it couldn't be "A" class. [16:36:50] This is why I dropped reference to numbers of properties. [16:37:04] Because it has a host of problems that, it turns out, are not even relevant to quality. [16:47:11] halfak: what I meant was, the item met all your criteria for "A", but it has 20 birth date, 20 occupations. Well, I just think this may influence the quality [16:53:05] glorian_wd, Oh! So what you're saying is we should also say what "A" class is not? [16:53:18] E.g. "Doesn't have inappropriate property duplication" or something? [16:53:27] Is there a Manual of Style we can reference? [16:54:09] halfak: exactly. As far as I can tell from the discussion, there should not be inappropriate property duplication [16:54:14] unique property [16:54:44] glorian_wd, rather than saying that explicitly, I'd rather make reference to a policy or guideline. [16:55:48] weird, there’s a “single value” constraint on https://www.wikidata.org/wiki/Property_talk:P570 (date of death), but not on https://www.wikidata.org/wiki/Property_talk:P569 (date of birth) [16:58:23] WikidataFacts, might be related to Jimmy's date of birth silliness? [16:58:34] halfak: hm? [16:58:53] doesn’t ring a bell… Jimmy Wales? [16:59:17] https://www.theguardian.com/technology/2011/feb/25/jimmy-wales-date-of-birth [16:59:19] lol [17:00:05] Check out his birthday https://www.wikidata.org/wiki/Q181 [17:00:06] huh [17:00:37] so the difference is that dead people can’t spread confusion about their date of death? :D [17:00:38] It's either Aug 7 or Aug 8 depending on if you believe Jimmy or other references [17:00:43] lol [17:01:03] halfak: well, in some cases, there should be multiple values in certain property. That way, we can represent multiple truths. Hence, I dont think there's such guideline [17:01:42] glorian_wd, surely there are appropriate times to have multiple values and other times where it is inappropriate. [17:02:01] But if it is not important enough to have a guideline, then why should we include it in our quality criteria? [17:02:22] I mean, we don't have "description isn't just a bunch of racial slurs" in our "A" class criteria either. [17:02:58] halfak: aren’t those constraints on the property talk page such a guideline? [17:03:09] Oh sure! Good point. [17:03:13] Let's say something about that [17:03:25] "statements follow relevant property constraints" [17:05:55] Yup, I agree on that. Constraints some times need to improve (and are not complete) but better than nothing [17:21:02] halfak: okay do you want me to modify the criteria? or you will do that again? [17:43:14] hmm this is interesting https://www.reddit.com/r/wikipedia/comments/5srnfs/a_current_list_of_articles_that_cite_breitbart/ [18:00:08] glorian_wd, let's keep chatting in the sub-sections. [18:00:52] I think that, after we come to a conclusion about property constraints and how we'll talk about media, etc, we'll be ready to modify the scale again. [18:01:14] halfak: do you mean under https://www.wikidata.org/wiki/Wikidata_talk:Item_quality ? [18:01:30] Yeah. [18:02:01] Could you add a sub-section about Property constraints and link to an example of a constraint? (E.g. on date of death vs. date of birth) [18:02:11] I'm running off to lunch. Will help when I get back [22:37:49] DanielK_WMDE_: around? [22:39:12] SMalyshev: just barely... about to go to bed. what's up? [22:39:36] DanielK_WMDE_: we have a problem with content handlers - core assumes you can always create empty content, but Wikibase can't. Ideas? [22:39:57] oh damn, I meant to look into that! too much stuff going on... [22:40:19] I think I got rid of this assumption at some point. Or tried to... [22:40:22] well, I can make a patch if you have an idea which direction it should be solved [22:40:39] so, fix the core test? or just drop it? [22:40:42] Not sure whether this needs to be fixed in the one way, or the other [22:40:50] yeah me too :) [22:41:10] sorry. I'll try to look into it tomorrow. if i don't, poke me again [22:41:25] ok