[06:34:46] re: authorty control datas for http://test.wikipedia.org/?curid=41108#BEACON:_ZCZC::LN::Novershtern:Nowersztern::FN:Avraham::main.VIAF:19830420::included.VIAF::additional.VIAF:195245545:203037735:98091675::CZCZ [09:24:17] fyi: [11:21:49] hey :) is there a way to subscribe to a specific review request in gerrit similar to CC in bugzilla? [09:24:17] [11:22:37] Lydia_WMDE: just add an empty review with score 0 [11:45:33] hmm, english office hour ? yay [11:45:47] DarkoNeko: jep :) [11:45:50] in 15 minutes [11:49:48] :) [12:00:13] hey everyone :) [12:00:18] hi there [12:00:19] hello [12:00:20] hallo [12:00:26] let's start another office hour \o/ [12:00:33] who's here for the office hour? [12:00:44] * DarkoNeko raises paw [12:00:45] * jem- is [12:00:52] i am [12:00:55] fyi: Denny_WMDE is sitting next to me [12:00:57] first time attending [12:01:00] nice [12:01:02] well, I probably won't be too active, unfortunatly ^^; [12:01:12] alright [12:01:25] let's quickly collect some questions and then go to answering them [12:01:34] i'll keep a list that we can go through [12:02:12] no-one? come on - don't be shy ;-) [12:02:17] just a reminder, there is also a German speaking office hour scheduled today, starting at 18:30 CEST, i.e. in 4.5 hours [12:02:17] okay then [12:03:02] Denny_WMDE: is wikidata going to be a wikimedia project or project of WMDE? [12:03:03] I've heard remarks about how the wikidata system would render editing (it's data) more difficult for newbies. how wil lyou get around that ? [12:03:08] like, help bubble and the like ? [12:03:14] because I see so many WMDE people around here [12:03:30] good start :D [12:03:44] i am saving them and then we'll go through them [12:03:47] more questions? [12:04:18] Ok, another one focusing on Phase I-interwikis: What's the idea about interwikis in namespaces different from the main one? [12:04:31] k [12:04:41] petan|wk, the developpement is managed by WMDE, which should give it to the WMF witch will then manage and support its maintenance [12:04:56] [12:04:59] eh, right [12:05:01] :D [12:05:04] thx DarkoNeko [12:05:14] is it even open source project? [12:05:16] ok anything else or should we start with the answers? [12:05:18] The FAQ says "WMF witch"? I hope not :) [12:05:18] myu engliosh suck, as do my memory, so if I forgot anything.. ^^; [12:05:20] petan|wk: totally [12:06:20] for the API are a lot of the functions implemented? [12:06:22] ok since DarkoNeko already answered petan|wk's questions let's go on with: [12:06:23] [14:03:03] I've heard remarks about how the wikidata system would render editing (it's data) more difficult for newbies. how wil lyou get around that ? [12:06:23] [14:03:08] like, help bubble and the like ? [12:06:34] I am just catching up on the emails which I got [12:06:36] aniket: i'll put it on the list [12:06:42] http://www.mediawiki.org/wiki/Extension:Wikibase/API [12:06:50] I am looking at this list [12:07:04] thanks DarkoNeko for the question. In Wikidata it should be *easier* to edit the data than it used to be in Wikipedia [12:07:17] we will have forms and user guidance for entering data points [12:07:32] currently instantiating a template is far from trivial for newbies [12:07:45] but on wikidata's side. Will there be those from wikipedia/wiktionary have a direct link to them too ? [12:08:11] we expect on Wikipedia articles to have direct links to the respective Wikidata page [12:08:49] okay [12:09:18] we are even investigating about having direct editing features for data within Wikipedia, but this is not sufficiently advanced. The API would allow such an UI, but this really ties in to how Wikipedia editing may look like in say 9 months. [12:09:46] DarkoNeko, follow-up questions? [12:09:55] hmmm. not currently, thanks [12:10:42] petan|wk, i guess your question is answered by DarkoNeko ? [12:10:49] sort of [12:10:57] so, initial dev by WMDE, operations by WMF and it is all open source [12:11:15] if it's open source how come is development done only by WMDE? [12:11:17] our git is available from everyone to pull and send pull requests to [12:11:19] petan|wk: [[Wikidata]] has a lot of stuff on the development [12:11:20] 10[1] 10http://meta.wikimedia.org/wiki/Wikidata [12:11:25] ok [12:11:35] oh, i sure hope that WMDE will not be the only one developing this [12:11:55] that development is funded by WMF or someone? or it's purely open source [12:11:57] i hope that we will get many pull requests once we have a bit of a somewhat stable environment [12:12:05] petan|wk: wmde manages the development and has hired developers. that doesn't mean others can't contribute. the code is in git. [12:12:09] it is purely open source *and* it is funded [12:12:28] ok [12:12:30] though i think it'll be easier to contribute once the codebase has settled down a bit. it's all still very much in flux, currently [12:12:36] funded 50% by [[AI²]], 25% by google and 25% by...errr... well, the betty and gordon moor foundation [12:12:37] 10[2] 10http://meta.wikimedia.org/wiki/AI%C2%B2 [12:12:51] moore, yes [12:13:10] ok? [12:13:16] (please tell me if I shouldn't be answering questions :) [12:13:16] ok, does this answer your question, petan|wk? next question? [12:13:25] DarkoNeko, thank you for the support :) [12:13:29] right [12:13:42] ok next one is: [14:04:18] Ok, another one focusing on Phase I-interwikis: What's the idea about interwikis in namespaces different from the main one? [12:13:54] :) [12:13:56] jem-: good question [12:14:14] we are currently focusing on the main namespace, and actually have a bit of trouble with other namespaces [12:14:23] due to them being internationalized [12:14:55] I understand, but there sould be a numeric connection between them in all cases [12:14:56] we want to enable links to other namespaces, but it is not our main priority. it might be deferred a bit [12:14:59] if it's any consolation, even current interwiki bots have a lot of problems outside (main) [12:15:11] a numeric connection? [12:15:24] I know, DarkoNeko [12:15:39] Denny_WMDE: I mean: The Help namespace is number 4 in all projects [12:15:45] ah, ok, got it [12:16:05] (Well, number 6, after checking) :) [12:16:07] we will probably have a small dev note on links to other namespaces [12:16:12] and ask for input on the topic [12:16:26] we want to enable this, but it is not our top prio [12:16:39] i still think it will run for the wikimania demo [12:16:42] let's see :) [12:16:52] ah, right, the api allow getting a nsid to name conversion [12:16:52] Denny_WMDE is a huge optimist :D [12:17:04] jem-: the id is only the same for standard namespaces. but some wikis have special namespaces for Portal or Index. [12:17:12] does this answer the question, jem-? we will ask for input on this topic explicitly too [12:17:22] they may have completely different ids, even though they interlink between projects [12:17:30] Daniel_WMDE, yes, and es.wiki has "Anexos" beyond number 100 [12:17:31] if someone wants to start a note on this already, feel free to do so [12:17:44] jem-: so we can't rely on ids. [12:17:47] always easier than to start form an empty board [12:18:31] follow up questions, jem- ? [12:18:35] Denny_WMDE: we have this on the issues page: "Wikipedia-Links have to be normalized wrt namespace aliases etc (in the db and when querying it)" [12:18:46] Denny_WMDE: Ok, I'll be paying attention to the question in the future, I understand things have to be step by step [12:18:59] Daniel_WMDE : yep, we need to deal with that [12:19:08] thanks, jem- [12:19:13] let's go to the next question [12:19:17] But it's interesting because interwiki bots won't have to be retired until this is solved [12:19:21] sure, i just wondered if there's anytzhing to be added to this item, in the light of what jem- said [12:19:27] this would be: [14:06:20] for the API are a lot of the functions implemented? [12:19:27] [14:06:42] http://www.mediawiki.org/wiki/Extension:Wikibase/API [12:19:27] [14:06:50] I am looking at this list [12:19:55] Daniel_WMDE: i think that the normalization would pretty much resolve the question of jem- — let's pick this up later maybe [12:20:12] aniket, a number of the functions are already implemented [12:20:16] let me quickly check [12:20:23] thanks! [12:20:58] while he's checking, I4d like to remember people of https://meta.wikimedia.org/wiki/Wikidata/Queries ... I've been feeling a bit loenly on that page :D [12:21:10] aniket: just out of curiosity: why are you asking? do you have something spedific you want to do? [12:21:20] DarkoNeko: thanks! [12:21:37] aniket, so out of the 9 described modules, the first 7 are currently implemented [12:21:41] I was hoping to help implement some of the api but if it is done thats fine [12:21:50] 8 and 9 have an skeleton interface but do not do anything [12:21:54] ahh ok [12:22:25] where can I find information about module 8 and 9? [12:22:37] I am looking at the following link: http://www.mediawiki.org/wiki/Extension:Wikibase/API [12:22:48] but it doesn't break it down by module numbers [12:22:51] DarkoNeko: You have my word that I'll be adding stuff in that page :) [12:23:03] yay jem- :) [12:23:03] yay [12:23:06] aniket, i was thinking about the numbers in the table of content in the beginning [12:23:09] sorry, this was not clear [12:23:11] ah k setalias [12:23:16] and wbsearchbyname [12:23:22] so setalias and wbsearchbyname [12:23:24] right [12:23:30] thanks! [12:23:44] and i am not sure why this is not wbsetalias... [12:23:47] need to check that out [12:23:55] (scribble scribble) [12:24:33] from the team, it is especially john working on this aniket [12:24:45] in case you didnt get in touch with him already, but i think you did, on the mailing list? [12:24:48] yeap i am in the process of reading his email about it [12:24:54] good :) [12:25:02] thanks! [12:25:03] actually we are already out of questions? [12:25:10] we are.... [12:25:17] moaaaaaaaaaaaaaaar questions please [12:25:18] :D [12:25:43] so either we are sufficiently confusing in what wikidata is doing or sufficiently clear not too muster any more questions at the moment... [12:25:59] * Lydia_WMDE really hopes it is the latter [12:26:01] ;-) [12:26:25] heh [12:26:30] Well, I have more questions but more focused on phase II/III [12:26:38] jem- go ahead [12:26:39] do ask away jem- [12:26:43] Ok then :) [12:26:50] even though it might be that the answers will be "we do not know yet" [12:27:04] so all the answers are the current thinking about it :) [12:27:09] For sure :) [12:27:27] by the way, we are only days away from a public demo system :) [12:27:29] at some point you're going to be storing revision data that will have something other than wikitext in it right? or am I also completely confused about what the wikidata project is about? [12:27:41] Looking forward for that public demo :) [12:27:53] apergos, yes, this is correct [12:28:10] Daniel_WMDE has extended MediaWiki so that we can deal with other content types than mediawikitext [12:28:22] most notably javascript, css, or wikidata data [12:28:30] oooh [12:28:36] how will this be handled by .. the parser, various routines that normalize of otherwise munge revisions? what about improt and export of those? [12:28:41] internationalisation of gadgets. that sounds good [12:28:54] apergos: https://gerrit.wikimedia.org/r/#change,6101 [12:29:18] there's quite a discussion to be followed on the mailing lists and in gerrit [12:29:47] I saw the discussion on wikitech-l but it didn't really give me any answers to the above [12:30:24] Daniel_WMDe, can you jump in for this question? i think you are better equipped to answer apergos than me [12:30:47] Daniel_WMDE: ^ [12:31:36] apergos: there will be aq contenthandler object for every type of content [12:32:03] apergos: the contenthandler (resp. the content wrapper created via the contenthandler) decides how the content is dealt with [12:32:29] apergos: e.g. the Content object has a getParserOutput function. it can implement that in any way. [12:32:46] apergos: the ContentHandler also has functions for serializing and deserializing Content objects. [12:32:53] in import and export, this would be used [12:33:11] the export xml will contain information about the content model an serialization format [12:33:18] apergos: does that answer your question? [12:33:59] it's a start (and I don't want to monopolize this session); I'll have some questions on wikitech-l or someplace about what this means for things like rev_len in xml output, not to mention the sha1 column. [12:34:00] thanks. [12:34:23] good :) [12:34:28] there are a bunch of places in the code that need to be changed so they don't access page text directly. but it's all doable (though EditPage is nasty) [12:34:34] jem-: you wanted to ask some more questions? [12:35:14] apergos: quick reply to that: rev_len will be a bogus number for non-text. sha1 however will be based on the serialized form. [12:35:35] wait for my email (you'll hate it, I alrady do and I haven't sent it yet) :-D [12:35:41] Lydia_WMDE: Here I go :) [12:35:58] After reading the data model, I don't know if there's a clear distinction between "two sources with two possible values for a property" and "two effective distinct values for a property"... [12:36:31] no, there is not [12:36:40] that is a very good question [12:36:55] :) [12:37:00] i hope that this distinction is not necessary [12:37:07] but i maybe proven wrong [12:37:22] but enabling the data model to express that distinction would make it quite more complicated [12:37:41] and i am not convinced that the number of use cases for this are too many [12:37:48] i.e. where this indeed makes a difference [12:37:54] Well... I think there can be [12:38:01] yes, i know there can be [12:38:04] For example, sons of a person [12:38:11] especially with things like issue of a person [12:38:12] yes [12:38:13] children* [12:38:41] we will have to see how well we fare with the current model [12:38:53] and if there is sufficient demand for changing it, we will have to change it [12:38:59] Or days in which a "moving" celebration is held for each year [12:39:31] in this case the semantics would be clear [12:39:43] i can hardly think of a moving celebration where sources disagree on the date [12:39:55] unlike for the children example [12:40:37] as said, the current data model does not allow for this distinction [12:40:39] Yes, but if there were sources for this and next year's date, it will all be mixed up [12:40:55] it could be added in the form of "list of values" instead of just having atomic values [12:41:15] Yes, that's an option [12:41:20] but we have to see if this is really needed [12:41:31] it makes things quite more complicated, especially in the UI [12:42:00] this is one of the reasons we are collecting use cases [12:42:12] to capture exactly this kind of cases [12:42:22] and to see how frequent they are [12:42:37] Ok :) I'll work on the matter and add cases [12:42:41] thank you! [12:42:54] Oh, that is related to another question [12:43:04] (Talking about children) [12:43:14] the current way to solve this would be an interplay between statements and how they are starred or not [12:43:54] i.e. in which rank they are [12:44:17] but we have to see how this works out in the dirty reality of wikipedia [12:44:34] Ok, for the moment I take note [12:45:08] 15 minutes left [12:45:22] this is one of the more interesting cases. we need to see how it works out. [12:45:49] the data model is not set in stone, obviously, but it is our current working draft [12:45:55] So, question: How will be handled the connections between items? If parent and child are both items, both have to have the property which connects them? [12:46:25] no [12:46:47] there is no such thing as an autofilled inverse property so far [12:46:59] this might be added later, though, and wouldnt be to hard i guess [12:47:04] So... if they both have, they would be independent... and they could be contradictory [12:47:05] right now it would be manually added on both sides [12:47:12] right [12:47:18] Hmmmm [12:47:22] but contradictory data is something we have to deal with anyway [12:47:36] i also imagine bots to go through the data and look explicitly for such cases [12:47:44] and create reports for editors to look through [12:48:07] Yes, that could be it if there's no better solution [12:48:11] i trust humans a lot in such things, and wikidata is always meant to be a socio-techincal system [12:48:25] and too much automatism can be harmful [12:48:43] I understand [12:48:51] especially as the world is a really fascinating place where almost every rule does have interesting exceptions [12:49:02] jem-: automatically forcing consistency is not really desirable in this system, i think. it would remove a lot of flexibility. [12:49:18] I understand too [12:49:22] damn you world for being so complex ;-) [12:49:26] Denny_WMDE: i have a question :) [12:49:29] next question? [12:49:32] Daniel_WMDE: ! [12:49:39] (i already put it on the issues page) [12:50:02] basically - if a user sees a number in the infobox, and goes to change it... what should be the user interface, what should the semantics be? [12:50:05] 10 minutes left [12:50:07] i think in time there will be an interesting ecosystem of bots, human editors, etc. that will be interesting to investigate to see how the data is maintained and used [12:50:16] is the user replacing the old value, but keeping the source? that's usually the wrong thing [12:50:24] would the user be adding a new value with source? [12:50:34] Daniel_WMDE: right, i agree that would be the wrong thing [12:50:35] should the full list of values be presented, always? [12:50:51] that is why in shortipedia we had statements being immutable [12:50:51] Denny_WMDE: i'm actually inclined to say that values can not be edited at all. [12:51:01] i agree [12:51:12] is that reflected in the api? [12:51:26] well, we currently do not have an api for phase 2 [12:52:03] it also does not answer your question about what the workflow will be [12:52:07] for editing values [12:52:33] it will be hard enough to see what happens when the values are changed in wikidata [12:52:35] not totally. but i would suggest to not have an api for editing values either [12:52:44] for remote editing from wikipedia, it will be even more interesting... [12:52:55] Daniel_WMDE, i am inclined to agree [12:53:14] modifying statements, however, may be useful. but only in special cases [12:53:28] hm.... well, adding qualifiers would be ok. [12:53:33] a little bit of it is probably needed [12:53:34] exactly [12:53:42] we have to see how it really works out [12:53:48] yea [12:53:57] maybe statements become immutable once they have sources or s.th. [12:53:58] but the editing ui is indeed not trivial [12:54:02] and before that they can be changed [12:54:04] uh, I wonder how nightmarish it will be to have mediawiki's cache cope with a change in a wikidata set [12:54:11] i'd suggest to show the full list of values to anyone editing a property [12:54:36] DarkoNeko, sufficiently nightmarish that we had a number of long discussions on this already, and keep it as an expensive item in the back log [12:54:37] DarkoNeko: exactly as nightmarish as having it cope with a change in a template [12:54:39] no difference [12:54:46] ah, then job queue too ? [12:55:00] Denny_WMDE: the long discussion is really about the cross site communication mechanism. [12:55:07] but yea, it'S an interesting topic :) [12:55:13] Daniel_WMDE, yep [12:55:14] DarkoNeko: for the re-rendering, yes [12:55:30] it is not within one system. so this makes it a bit more interesting [12:55:36] but it is not completely new problem [12:56:07] mmhmmm [12:56:32] we had a discussion started on the mailing list about this last week [12:56:37] and have some notes in the wiki [12:56:52] if you want, i will check out the links [12:57:09] http://meta.wikimedia.org/wiki/Wikidata/Notes/Caching_investigation [12:57:17] thanks Lydia_WMDE [12:57:42] note that the notes don't include ryan's input [12:57:54] we'll have to talk about this once more [12:58:00] they are on the public mailing list discussion? [12:58:05] yes [12:58:10] good [12:58:56] alright - we're almost at the end of this office hour [12:58:59] one last question? [12:59:52] I'll save more for the next one :) [12:59:58] hehe [12:59:59] ok [13:00:02] so the hours of the office hour are meant to osciallate [13:00:16] so that we have different time zones being served by different office hours [13:00:26] this time it was not exactly convenient for the americas [13:00:40] i would be curious, where are you from? [13:00:51] especially those who are here for the office hour [13:00:55] I'm from Spain :) [13:01:08] But I agree rotation is good [13:01:25] country or time zone is enough, i don't want you to expose more private data than needed ;) [13:01:56] alright folks [13:01:57] Spain has UTC+2 which I think it's the same of Germany [13:02:03] let's wrap this up then :) [13:02:05] yes, it is [13:02:07] (Summer time currently) [13:02:08] thanks for showing up everyone! [13:02:21] Ok, it was a pleasure :) [13:02:25] * DarkoNeko bows. thanks to You [13:02:27] thank you for the questions! [13:02:28] you're more than welcome to hang around and ask more questions or use the mailing list [13:02:40] there will be a german office hour in 3.5h [13:02:42] * DarkoNeko is half allergic to mailing lists [13:02:48] haha [13:02:52] :) [13:03:06] * DarkoNeko have a monstruous backlog on the teward and toolserver ones ._. [13:03:15] * jem- will keep hanging around, of course [13:03:37] DarkoNeko: "mark all as read" is one solution to that problem ;-) [13:03:43] * jem- also knows a lot about backlogs [13:04:08] Lydia_WMDE, that...is true, but... :D [13:05:21] thanks for the hour [13:05:47] :) you're welcome [13:05:53] i will publish logs in a bit [13:06:22] Talking about logs, Lydia_WMDE [13:06:31] jep? [13:06:52] I think you should change the dir for the logs in the topic [13:08:09] * Lydia_WMDE looks [13:08:19] wm-bot's logs are better suited for permanent logging, while I have to restart logging after each restart, and I don't log time [13:08:41] (because the idea of my logs is to copy them into a wiki page) [13:08:46] jem-: ok - what should be the new link? [13:09:02] http://bots.wmflabs.org/~wm-bot/logs/#wikimedia-wikidata/ [13:09:15] Ehmmmm [13:09:21] Better: http://bots.wmflabs.org/~wm-bot/logs/%23wikimedia-wikidata/ [13:09:28] thx [13:09:43] Thanks to the wm-bot operators :) [13:09:56] (you could probably have changed the topic too) [13:10:03] :) [13:10:20] Oh :) [13:10:32] Isn't it protected? [13:10:45] not at the moment it seems [13:10:58] Ok then, I take note [13:11:47] ... I could review my bot's code, but my already mentioned wiki-tasks backlog is really huge [13:12:20] So it's better to use the work that is already done :) [13:12:33] hehe yeah [13:12:36] i know the feeling... [13:12:56] We should start a club :) [13:13:17] :P [13:13:25] do we get membership cards? [13:13:28] * Lydia_WMDE wants one [13:14:03] * jem- too, if making the cards isn't going to his backlog [13:14:19] later g2g to work [13:14:25] jem-: haha [13:14:28] thanks for the information [13:14:29] aniket: see you :) [13:15:55] Ok, I'm leaving off to have lunch (Spanish schedule, yeah), so see you later, and for all the team: keep up the good work! [13:16:36] :) [13:16:39] laters jem- [13:59:53] Tomorrow is a normal working day no? Or is there some holiday I don't know of? [13:59:58] * JeroenDeDauw pokes Lydia_WMDE [14:04:24] JeroenDeDauw: I guess there is �Labour day� in Germany (and some other countries). But is that a wmf holiday as well? [14:04:49] I dno [14:04:54] Denny_WMDE: work day tomorrow? [14:05:58] JeroenDeDauw: yes tomorrow is a holiday [14:06:15] we should try not to be out too late because of the demonstrations [14:06:27] JeroenDeDauw: tomorrow is no working day [14:06:38] WHAT?!!! I haven't coded real stuff for a week, and then there is a holiday?! DAAAAAAAAAAAAH! [14:06:46] sorry [14:07:02] So the office is closed? :p [14:07:09] yes [14:07:12] you will especially enjoy the irony that is the "Day of work" tomorrow [14:07:22] o_O [14:14:54] Not "Day of work" - Labour (the unions of workers) Day. [14:16:15] Also, in the USA at least, it is widely expected to have many protests. Good day to stay home. [14:17:20] Amgine: in berlin as well [15:00:34] Lydia_WMDE: Quick translation question: Main article = Hauptartikel, Main articles = Hauptartikel ?? [15:01:15] Amgine: jep :) [15:01:33] Thanks Lydia_WMDE. [15:02:14] sure [15:48:12] does someone have the link to the ui study on how editing of interwiki links could look like in the future? [15:48:24]