[00:43:49] hey folks, anyone have any experience/thoughts on multilingual semantic wikis? [00:44:24] i'm merging two wikis, one in french and one in english, nad we're trying to decide whether to use subdomains (eg. fr.appropedia.org) or whether to try something using SMW [00:44:29] we're already using SMW btw [00:44:57] the audience are far more multilingual/internationally minded than the average person [00:45:09] Skud: hello! [00:45:37] Oh, I thought you were asking due to the potential Google Summer of Code project on that very issue, but I guess not. [00:45:44] Ah, Appropedia. [00:45:46] http://www.appropedia.org/Special:BrowseData/Medical_Devices?_single= is the sort of stuff we're dealing with. it's used by eg. Engineers Without Borders, WHO, etc. [00:46:09] oh, i didn't know there was a GSoC project doing something similar! [00:46:19] Awesome! I didn't know all those orgs used Appropeida, and Semantic Drilldown specifically. [00:46:29] it's just ramping up [00:46:40] medical devices was the pilot project but i'm about to start on a project with EWB [00:46:44] Well... maybe. GSoC is starting up in a month or two, and it's one of the potential projects. [00:46:48] Oh, I see. [00:47:04] but we're importing another wiki (ekopedia.org) which has lots of non-english pages [00:47:08] Cool. [00:47:43] The big question when dealing with multilingual SMW wikis is whether the semantic data will be kept in one language, or whether each language will have its own version of the data. [00:47:54] hmm [00:48:22] i'm guessing they'll want it integrated tbh [00:48:33] i can't imagine them wanting to fragment eg. their database of medical devices [00:49:23] they currently have some mixed language stuff in appropedia and their main concern is reader surprise on suddenly finding themself on a spanish page or something [00:49:33] So, data stored in English, and then ideally displayed in other languages? [00:50:37] hmmm. [00:51:26] i'm not sure whether the entire wiki will be heavily semantic [00:51:38] it's something that's just ramping up in certain parts of the wiki [00:52:23] Well, I'm just asking about the SMW-using parts. [00:52:43] yeah [00:52:57] but if we do this import, then all the multilingual pages will be SMW-using in a small way [00:53:04] (i.e. just language metadata) [00:53:24] You don't need to store the language semantically... [00:53:25] there's not really any display aspect to that [00:53:29] oh? [00:54:50] Why would you need to store it? [00:55:08] so that you can use it semantic drilldown [00:55:17] and so that we can use it to generate some template stuff [00:55:40] including maybe a hover popup or something on links to let people know they're about to suddenly wind up in another language [00:55:44] I think there's some misunderstanding here. [00:55:54] quite likely :) [00:56:20] You said that only a small fraction of the pages will use SMW - i.e., only a small fraction will be browsable in Semantic Drilldown. [00:56:20] this is just an idea that came up in brainstorming around approaches [00:56:39] ah yeah i wasn't communicating well there [00:56:51] But now you're saying that every page in a non-English language should be browsable in SD? [00:56:51] *at present* there is only a tiny amount that's semantic [00:57:02] the medical devices stuff uses it heavily [00:57:11] and we're about to do a second, similar area related to EWB projects [00:57:19] but there are huge swathes that are not currently semantic at all [00:57:46] if we do this import from ekopedia, one of the options is to treat language as a semantic thing, and to add semantic tags for the language of each page [00:57:59] Right, I got all of that. [00:58:01] But - why? [00:58:02] if we do that, the language tag will probably be the only semantic info on most of those pages [00:58:15] so at that point, a large proportion of the wiki will be semantic to some degree [00:58:24] but not heavily/complexly semantic [00:58:48] so as to "why" [00:59:02] we were just debating alternatives to taking the fr.appropedia.org approach [00:59:06] there are some downsides to that [00:59:22] No, not why do the import - why store the language semantically. [00:59:39] so the other options seem to be: [00:59:50] 1) use a subdomain approach (what wikipedia does, and what ekopedia currently does) [00:59:57] I don't see how Semantic Drilldown would really help with browsing different language pages. [01:00:02] 2) wiki namespaces (lol no, just including for completeness) [01:00:08] ...compared to the standard tools, like the Language extension. [01:00:08] 3) do nothing, just mix them all up [01:00:25] If you're not familiar with that one, you should definitely check it out. [01:00:31] looking now [01:01:05] It's used heavily on mediawiki.org - it makes the multingual stuff pretty easy to put together. [01:01:13] i probably shouldn't be surprised that there are already extensions for this :) [01:01:37] ok i'm finding a heap of different language extensions but not clear on which one you're referring to. link? [01:01:48] http://www.mediawiki.org/wiki/Extension:MultiLanguageManager [01:02:04] and https://www.mediawiki.org/wiki/MLEB which sumanah pointed me at [01:02:15] Oh, did I say Language? I meant Translate. :( [01:02:18] https://www.mediawiki.org/wiki/Extension:Translate [01:02:31] this is the first multilingual wiki work i've doneso i'm not familiar with the solution space [01:03:03] ah ok [01:03:25] ok what i'm seeing with the translate extension is that you use one wiki, one namespace, and that this adds UI stuff to assist in writing translations of pages and interlinking between them? [01:03:37] Yes. [01:04:03] that looks pretty sweet! [01:04:13] is there an assumption about the base language of the wiki? [01:04:45] Yes... every wiki has a base/default langauge. [01:04:47] yeah, looks like there is [01:04:53] this looks like a great extension [01:05:09] how does it interact with SMW? [01:05:29] like if i were in the semantic drilldown page for medical devices would i see Device and Device/fr? [01:05:37] Well... there are some rough spots there - that's what this GSoC project would be about. [01:05:45] ah nice! [01:05:50] is there a writeup of the GSoC project? [01:06:10] For actual querying/SD stuff, though, I assume everything would get stored in the main language. [01:06:31] and also, if you get a student on it, i'd love to hang around and help/test/generally be part of the process [01:06:33] This project would be mostly about data entry, i.e. Semantic Forms. [01:06:48] That would be great - they were looking for test wikis. [01:07:08] Although there's a good chance it won't happen, so I don't want to get your hopes up too much. [01:07:21] https://www.mediawiki.org/wiki/Google_Summer_of_Code_2014#Multilingual_Semantic_MediaWiki [01:07:22] sure [01:08:06] fantastic [01:08:24] what's the current status of GSoC applications? are they currently open? [01:08:33] appropedia could help spread the word and say that they're keen for someone to work on this [01:08:53] Students can still apply, until the end of the week... it's getting pretty late. [01:09:07] eep! [01:09:12] ok i will let them know today. [01:09:31] Again, this project would be intented mostly to help with data entry - the assumption, I believe, is that all data would be stored in one language. [01:09:42] yup [01:09:51] the EWB project involves data entry [01:10:06] and the timing is basically "summer, ish" (northern hemisphere time) [01:10:26] Ah, great. Yes. [11:38:10] (PS1) Raimond Spekking: Consistency tweaks in preparation for adding extension to translatewiki.net [extensions/SemanticSifter] - https://gerrit.wikimedia.org/r/119464 [13:12:17] (CR) Siebrand: [C: 2 V: 2] Consistency tweaks in preparation for adding extension to translatewiki.net [extensions/SemanticSifter] - https://gerrit.wikimedia.org/r/119464 (owner: Raimond Spekking) [15:31:04] (CR) Chad: [C: 2 V: 2] Remove current implementation of getInfo() [extensions/SolrStore] - https://gerrit.wikimedia.org/r/119237 (owner: Chad) [17:01:39] Yaron: MIT license? [17:01:50] Yaron: y u not in #wikidata as well? ;p [17:02:44] I... don't understand that first question. Shouldn't it have a verb or something? [17:03:12] Yaron: It's re "[MediaWiki-l] New extension DataVisualizer" [17:03:17] ls there any reason for me to be in #wikidata? [17:03:26] Oh, I see. Well, it's Nischay's extension... [22:39:41] question for all: is there a parser function that will magically insert the current user's username when evaluated? [22:40:00] I'm wanting to create a link to a Form that will create the User's own userpage. [22:42:23] looks like the getusername extension will do this, with potential security issues [22:42:53] doesn't look like a great option