[00:50:24] there's a slight problem with the pagemove functionality as it applies to media files [00:50:37] specifically, audio files that have TimedText accompaniments [00:51:18] there doesn't appear to be any way to tell whether a file *has* a TimedText accompaniment, which can result in TT files being orphaned when the audio is renamed [00:51:46] this has happened at least once to my knowledge, and probably more [00:53:38] it's probably possible to find such orphans [01:11:45] Base-w - would it be possible to prevent such orphans from being created? [01:16:03] * Base-w has no idea, just passes by [01:17:27] Dragonfly6-7: file a task in phab [01:18:03] p858snake - you're assuming a skillset not in evidence [01:19:58] robla: I've done my duty :) -- https://phabricator.wikimedia.org/T119032#1896800 [01:20:26] * robla reads [01:21:45] p858snake - I could probably climb my way up the learning curve, but I'm doing a lot of things at once [01:22:53] Dragonfly6-7: https://phabricator.wikimedia.org/maniphest/task/create/ [01:23:09] if you need to login, use the "Oauth" button at the buttom of the login screen [01:24:23] bd808: awesome, thanks! I've got a half composed comment on T119029 that comports with what you wrote [01:24:41] sweet [01:25:34] 'oauth' ? [01:25:44] I see nothing of the sort [01:26:01] Dragonfly6-7, https://phabricator.wikimedia.org/T122038 [01:26:45] thank you, quiddity [01:27:01] Dragonfly6-7, That ^ should cover the bug you're reporting. Re: login - just click the button with the daisy icon. [01:27:26] it will send you to mediawiki, and ask to allow oauth. [01:27:31] Dragonfly6-7: of its "Login or Register" with Mediawiki on the next row [01:31:59] p858snake - thank you, but I don't really need to log in myself now. [01:32:00] .... do I? [01:32:11] no [01:32:25] but you can subsribe to the task to get alerts when people comment on it etc [01:32:33] look at the other features of phab etc [01:32:36] if you desire [01:32:42] Can you add more hours to my day? [01:36:40] lolololol [01:53:13] bd808: gwicke: Tim-away : see https://phabricator.wikimedia.org/T119029#1896858 [01:54:19] looks like I may have missed my opportunity to catch Tim before he arrives here in SF [01:55:03] !hiera | mutante [01:55:23] ah, wm-bot triggers per channel [02:02:06] Hi, did the database dumps stop delivering page text? [02:04:14] Specifically the stub-articles, at least, I haven't checked the others [02:35:52] * Hazard-SJ guesses that everyone here is sleeping XD [02:36:17] !ask [02:36:20] :p [02:36:44] !and_I_shall_receive ? :P [02:37:07] it works on #mediawiki :) [02:37:16] I think it used to deliver the text, but now it doesn't, at least not the dumps I'm checking [02:37:31] This would probably be Wikimedia-related [02:37:48] do you know when it changed? [02:38:03] the dumps person is out right now, timezone [02:38:12] but a ticket would be great [02:38:25] Since we can't access page text from the databases generally... no idea, but I was of the impression that it used to give the page text :/ [02:40:12] A complete copy of all Wikimedia wikis, in the form of wikitext source and metadata embedded in XML [02:40:16] that's what you are trying, right [02:40:26] https://dumps.wikimedia.org/backup-index.html [02:42:36] mutante: I found my issue >< [02:43:04] Hazard-SJ: :) [02:43:16] I don't know how I ended up using stub-articles, it should have been the pages-articles dump :P [02:43:26] ok, cool [02:43:54] And it indeed has the text :) [02:44:06] Thanks, anyway :D [02:44:12] sure, no problem [02:48:13] I messed that up 7 days ago, apparently, lol [02:48:43] That was when I last updated the dump I used [10:01:31] 1500 characters are not enough for this task, students get confused because I was forced to remove some details :( https://codein.withgoogle.com/dashboard/task-instances/5038541679624192/?sp-page=1 [10:09:51] -> #gsoc :) [10:15:03] "Many MediaWiki documentation pages are only available in English. By adding a few annotations to the page source, the page can be imported into MediaWiki's translation tool, which helps translators find out about the page and provides a workflow for translating it. You will be adding annotations to an English-only documentation page to make it translatable." ? [18:52:47] Reminder: Lightning Talks start in 10 minutes. https://www.youtube.com/watch?v=cIqhGxtXFbg [18:54:02] Hi Megan! [18:54:16] thanks for putting out the reminder... I was just about to do that [18:54:43] Hi! no problem [19:12:09] question: RE: survey results: isn't this heavily biased towards people who wind up on the site? i.e. i would guess that _many_ people interact w/ wikipedia through off-site lookup experiences: e.g. info boxes or extracts shown in Google or Siri queries [19:12:25] i.e. we can't run a survey on other platforms' users [19:13:55] bgerstle: shucks, didn't see your question in time... [19:14:05] let me invite Leila into this channel [19:14:08] kevinator: np, i just learned your IRC handle :-) [19:14:19] i didn't know which autocomplete from "kev" it was ;-) [19:16:05] bgerstle: kevinator just showed me your question. [19:16:21] leila: cool, thx kevinator [19:16:52] bgerstle: We only understand the behavior of people who come to Wikipedia through these surveys. [19:16:58] right [19:17:22] bgerstle: we had some discussions about learning what happens between when the user leaves Wikipedia and when they come back to it. [19:17:46] bgerstle: do you have ideas about how we can look outside of Wikipedia, because those are very important and interesting aspects of the user behavior, too [19:17:51] leila: right, i guess is there anything we can glean from API usage of specific partners? [19:17:52] are any of these mobile features available on the web? [19:18:03] i know we recently talked to apple about how they use our data for Siri [19:18:10] getting more data from them could be really valuable [19:18:22] dtm which ones are you thinking of [19:18:22] i have a huge wishlist including a wish for a personal wikipedia portal and feed [19:18:27] bgerstle, any [19:18:30] * bgerstle is the iOS tech lead [19:18:48] i'm just now hearing of Wikipedia lightning talks and of this app [19:19:03] dtm there's no "feed" on the web (or anywhere else, AFAIK) [19:19:16] bgerstle, first off, anything that makes people log in is good [19:19:22] whatsoever [19:20:14] bgerstle, is mobile the only platform with "near you"? [19:20:19] dtm: unfortunately logging in doesn't get you much in the app aside from editing [19:20:29] well good [19:20:31] dtm: not sure about that, i think mobile web has a "nearby" feature [19:20:37] i mean i can simply give my location. [19:20:44] dtm: our hope is to explore more features related to login, like article collections [19:20:45] it doesn't need some GPS integration lol [19:21:12] dtm yeah, most user's don't know their lat/long, and we'd have to do some kind of reverse lookup to get the lat/long of a locale (e.g. "San Francisco, CA") [19:21:18] bgerstle: I think we should stay on site, understand Wikipedia site and the behavior on it first some more, and then go to external usage and consumption of Wikipedia content. [19:21:30] bgerstle, i definitely would like more accessible mobile editing. but i want full wikicode features usually. so i'd like to have wikicode, but not crammed into a microscopic display. is this possible? [19:21:50] leila: sure, they aren't mutually exclusive, i'm just curious if anyone has given though to qualifying the "weight" of the lookup case—given that most lookups are probably satisfied off site [19:22:10] dtm: we're focusing on reading use cases atm, but we have rudimentary wikitext editing [19:22:19] you do get a preview before submitting your edit [19:22:47] bgerstle: my real concern is that by going out at this point, we will make our work too thin. We have a lot of data to learn from internally. Once we go to other companies to ask for data, we need to do a lot of social/human work, to get some data that it's usually not as rich. So, at the moment, my recommendation is for us to stay, in some months from now, we try to learn what happens to the user when they leave the website. [19:22:55] One question for Joshua, if we have time: Why so much differences between the Android and iOS apps? [19:23:04] bgerstle, i have a tiny iphone 5 and still, i instantly disable mobile view in Safari. i just want a usable desktop interface, on mobile. [19:23:05] awesome, would love for maps to be there too ))) [19:23:13] leila: great presentation :) I'll send you a more detailed email to follow up. [19:23:30] guillom: there are long and short answers :-). short: we try to not build the same thing at the same time on multiple platforms [19:23:46] bgerstle: why? :) [19:23:56] bgerstle: yup. I understand your point. I don't know of any work on that front, though Partnership may be interested/doing something I don't know about. [19:24:07] bgerstle, *applause* [19:24:09] please clap 4 me [19:24:12] so we let android gather usage data on a new feature before spending time implementing on another platform [19:24:15] clapping proxy GO [19:24:17] dtm i'm remote as well :-P [19:24:24] okay we'll all clap harder then [19:24:25] guillom: great. looking forward to it. [19:24:28] CLAPPING NOW [19:24:56] leila: yeah, guess all that can be said about it is "soon" [19:24:57] bgerstle, how about augmenting iOS's terrible text search function? [19:25:17] we only have so many analytics people, so we probably good to focus on understanding "core" use cases [19:25:44] dtm: we have both Universal Links and Spotlight indexing on the backlog :-) [19:25:47] bgerstle, i guess what i'm asking for is a mobile app that can do everythign that desktop editing can do, but un-retarded from mobile. [19:25:50] I wouldn't even say "soon" bgerstle. ;-) [19:25:53] bgerstle: Good point. So are the two apps team coordinating so that the next features are split between the team to be built, tested, and if successful, also implemented by the other team? [19:26:02] guillom: basically [19:26:09] guillom: maps is the best example of this [19:26:27] bgerstle, i want "mobile" to mean "i walk around with this and might even stick it in my pocket", not "this is crippled" [19:26:29] guillom: we're also holding off on doing add'l' work to do "link preview" on "older" iOS devices (older than 6S) [19:26:31] bgerstle: great. This has never been made explicit to me. Thank you :) [19:27:20] guillom: it's not really documented anywhere, but yeah that's kind of been the approach which organically grew out of the re-org and our coincidentally divergent development strategies (which Josh mentioned—Android = incremental; iOS = overhaul) [19:27:33] (not documented anywhere AFAIK) [19:27:38] got it; Thanks :) [19:28:00] welcome :-) [19:28:58] dtm: right there with you :-) [19:29:29] by switching to github instead of requiring people to learn gerrit or phabricator [19:29:31] dtm we really want to understand what the *mobile* use cases are. IOW make the apps optimized for how people interact w/ Wiki projects while mobile—instead of just watering down desktop experience [19:30:20] yurik: just checking, was that directed at me or the current lighting speaker? [19:30:29] bgerstle, current speaker )) [19:30:31] bgerstle, yeah i'm saying "adaptive", not "watering down" [19:30:39] dtm yep [19:30:43] yurik: I'll relay your comment :-) [19:30:50] bgerstle, well done [19:30:56] yes please )) [19:31:19] dtm: thanks :-) [19:31:26] bgerstle, BRAVO! [19:31:28] dtm do you have an iOS device, and if so, are you registered as a beta tester yet? [19:31:46] bgerstle, I have iOS 9 on iphone 5 and ipad mini 1 (yuck) and no. [19:31:53] and i have android but they're just idle toys [19:32:21] my ios is not jailbroken but my android is [19:32:25] if that matters. [19:32:30] dtm let's continue discussion on #wikimedia-mobile (and/or #wikimedia-ios) [19:32:37] want to free up the channel for lightning talk convos [19:32:39] i want the big room :) [19:33:00] bgerstle, which one? lol [19:33:11] dtm mobile is more generic, ios is specific to iOS [19:33:14] up to you [19:34:23] my point exactly ))) [19:37:31] the most important thing to me is regarding wikipedia as a culture... what is anyone doing to increase quality assurance and reduce [[WP:BATTLEGROUND]]? such as the introduction of actual reputation (achievements and statuses between "plebian" and "all-powerful admin plutocracy cabal")? such as requirement of login accounts? updating of policies to rely less on arbitrary subjective survival of the most powerful? anything? [19:38:44] i think that the increased standardization of policies and code down to the manual of style and citation formats etc, would serve to equalize the culture and reduce WP:BATTLEGROUND. [19:41:04] there needs to be standards for qualifying editors. current wikipedia culture treats quality assurance as if it's a form of eugenics. so everything is arbitrary and relative; new editors don't have a clear path to improve, and incompetent editors are encouraged to endlessly "try try again". [19:41:14] and just tear down the whole [19:41:33] yurik: do we need to use templates? is there a way to get the data for graph in a structured format w/o parsing the page? [19:41:53] (thinking about how we would do this in apps/mobile) [19:42:39] in addition to technological improvements, we need a wikipedia university sorta thing to closely and tightly mentor and RATE and give qualifications and REPUTATION to editors, like you would in a real university. people need to be able to earn and maintain respect and minor amounts of tenure. [19:42:48] is this in any way relevant to this channel or lightning? [19:43:05] bgerstle, graphs can use any URL (on our domains). The problem is how to store data in general. [19:43:18] dtm no idea, maybe a mailing list? i have a sense that this feedback (which is great) might get lost on IRC [19:43:39] bgerstle, as for mobile - don't worry - you simply pull it as an image - rendered by graphoid service [19:43:51] yurik: but then it's not interactive :-( [19:44:14] yurik: "any URL" which points to...? [19:44:26] bgerstle, well that sux. these topics (and me) are better suited to interactive chat until i get oriented with who might be interested in this stuff [19:44:51] bgerstle, correct. For interactive, when you click on the graph, it loads vega js lib, and calls the api.php to download graph spec itself. If graph spec does not embed the data, it will reference the URL [19:44:56] dtm yeah, i'm trying to think of who from our community & community tech teams you should talk to [19:45:08] quiddity: ^ where should dtm send feedback about WMF technical direction? [19:45:15] bgerstle, the data can be on any WMF domain [19:45:22] bgerstle, quiddity: here's my wishlist https://en.wikipedia.org/wiki/User:Smuckola#Wikipedia_wishlist [19:45:27] which does include mobile [19:45:35] yurik: i guess showing an image, and then tapping through to render the graph interactively in a webview would work, assuming it's functional (and performant) on mobile web [19:45:38] well the portals overlap with mobile kinda [19:46:05] yeah [19:46:11] bgerstle, it works fine on my android - try it https://www.mediawiki.org/wiki/Extension:Graph/Demo [19:47:14] bgerstle, quiddity i'd call it sociotechnical, mostly ultimately social and cultural [19:48:38] dtm (bgerstle), I know you've mentioned it to the commtech team (channel) which is perfect. I've also added it to https://meta.wikimedia.org/wiki/Wishlists And there'll be some more requests for related feedback in the new year. [19:48:46] yurik: seems to work [19:49:04] quiddity: thanks 👍🏻 [19:54:27] quiddity, hellos. [19:54:43] quiddity, thank you [19:57:24] :) [20:27:15] quiddity, bgerstle: i just wrote a ratinoale atop my wishlist, which is summarized by "we need toys. lots of toys." [20:27:33] "Because we don't (or can't?) have a single point of final crosscultural leadership or even a benevolent dictatorship like a Steve Jobs or a Linus Torvalds, we need technology and policy to cull and prevent chaos and WP:BATTLEGROUND. And we need toys. Lots of toys." [20:28:19] dtm, +1. I mentioned our existing toys at https://www.mediawiki.org/wiki/Talk:Requests_for_comment/Redesign_user_preferences#Gadgets [20:29:27] quiddity, thank you [20:29:42] and feel free to qualify whether it's DON'T or CAN'T ;) [20:29:53] i see Jimbo as a power vacuum [20:29:58] personified [20:30:35] he's the type of person with the makings of personality traits to be a Jobs or Torvalds. but Torvalds refuses to live in a bubble or an ivory tower. [20:31:45] it's almost like Jimbo has a spine but it's like an anti-spine. he resolutely refuses to be involved? or, he thinks he is, from within an ivory tower? i dont know. i dont get it. [20:32:29] is he reacting to massive broad powerful volunteer consensus, which tells him "hands off"? [20:33:35] or is he the ultimate trickster? an agent of chaos, setting up a deliberately chaotic system to watch the annual statistical charts while the masses claw out a civilization? [20:33:55] wut [20:35:04] he's like a conscientious macro-observer who refuses to lead, even though all roads go to him. an anti-hero. [20:36:13] lmao [20:41:48] :-D [21:20:33] MusikAnimal: 13 xtools issues closed, xtools-articleinfo works again [21:21:00] really!?? [21:21:24] well I'll be! [21:21:31] who is responsible for this [21:22:45] you are! I see it on GitHUb [21:23:17] ori: tytytyty [21:23:34] :) [21:24:05] very occasionally i do things other than breaking testwiki [21:24:14] ha [21:24:15] how did you navigate that code [21:24:19] that is the real question [21:24:40] i took breaks to cry a little [21:25:47] I'm sure! [21:26:31] what else got fixed? [21:26:46] wow, gadget works [21:28:18] ori: I've updated https://en.wikipedia.org/wiki/MediaWiki:Histlegend [21:28:33] did I fix the gadget? if so it was not deliberate [21:28:45] it works now! [21:28:54] are you User:Ori on enwiki? [21:29:06] no that's not you [21:29:11] no https://en.wikipedia.org/wiki/User:Ori_Livneh [21:32:16] wanted to make sure you got some WikiLove [21:34:03] awww! thanks! [21:34:30] absolutely. You may not realize how badly people have missed that tool [21:35:10] and on top of that, fixing it was deemed near impossible, as we had many experienced PHP devs giving it their all, to no avail [21:35:24] Wait, Xtools is back?!?!?!?! :0 [21:35:29] Well done!!!!7 [21:37:06] indeed [21:37:48] it wasn't totally broken, was it? I hope I am not getting credit for more than I actually did [21:38:00] articleinfo was totally broken, not sure about the rest [21:38:12] ori, it is quite broken... :P [21:38:27] still, you mean? [21:38:29] most of the xtools suite has been working [21:38:46] it's just that article-info was probably the most widely used tool [21:38:56] My bad, I did mean articleinfo [21:39:08] also people really loved that gadget, which I'm guessing whatever you did fixed it (it hits the xtools API) [21:39:09] ah cool [21:39:32] the article blamer still isn't functioning properly, if you're up for that task... https://en.wikipedia.org/wiki/Wikipedia_talk:Copyright_problems#Article_Blamer_seems_to_be_broken [21:40:44] how to reopen an issue on github? [21:41:02] it looks like https://github.com/x-tools/xtools/issues/74 isn't actually fixed yet [21:42:09] Base-w, find a repo dev :) [21:42:26] I re-opened it [21:42:51] IIRC there is a lot of code that assumes the first part of the domain name is a language code [21:44:41] can some stuff be added so it does not think like this in case of *.wikimedia.org? [21:47:42] Base-w: works now [21:48:22] most .wikimedia.org sites have hard-coded entries in the source. i just added one for outreachwiki. not beautiful, but in a pinch.. [21:50:25] thanks. yeah works for outreachwiki, but does not for e.g. ua.wikimedia.org [21:51:36] yeah, it should really use the specified wiki's api to get things like language / db name [21:53:47] API or I believe this stuff is also present in some tables on toollabs's replicas [22:29:58] Base-w: ok, made it query the wiki's api https://github.com/x-tools/xtools/commit/78ea49a7fd274840d6212de5cc92d28e1729d1b2 [22:30:08] so now this works: https://tools.wmflabs.org/xtools-articleinfo/index.php?article=%D0%92%D1%96%D0%BA%D1%96%D0%BC%D0%B5%D0%B4%D1%96%D0%B0+%D0%A3%D0%BA%D1%80%D0%B0%D1%97%D0%BD%D0%B0&project=ua.wikimedia.org [22:32:37] ori: cool, thanks!