[18:47:19] _o/ [18:52:40] * lvillaWMF waves [18:52:49] I am pseudo-James today [18:56:52] Hi lvillaWMF (psuedo-James)! [18:57:03] * lvillaWMF waves at rdaiccherlb [18:57:20] * rdaiccherlb always has typos :( [18:59:12] rdaiccherlb: what, IRC wasn’t part of your workflow in a prior life? ;) [18:59:29] (for those just joining, I’m pseudo-James (aka moderator) for the day) [18:59:39] ohai lvillaWMF [19:00:02] rdaiccherlb: typos are part of irc [19:00:04] ohai [19:00:19] * brion waves at fake-james [19:00:21] part of its charm [19:00:25] lvillaWMF - I just type more quickly than my fingers can do so accurately. It happens in emails, too. You've also seen how fast I can talk. [19:00:25] * subbu is waiting for the youtube stream link [19:00:30] heya...I've got R32 set up as a remote viewing room on 3rd [19:00:31] * PKM lurking [19:00:33] rdaiccherlb: :) [19:00:46] youtube url, btw: https://www.youtube.com/watch?v=kIguVBIfnCY [19:01:05] Hello everyone [19:01:39] Hi everyone, I'm helping with AV today, any comments (even positive ones) would be very appreciated. Thanks! [19:01:49] hi, victorgrigas [19:01:54] cndiv: thanks [19:02:09] s/even positive ones/especially positive ones/ [19:02:23] cndiv, you are always a rockstar at this. Thank you! [19:02:45] started. comments appreciated. [19:03:04] A little feedback [19:03:15] works, just had to reload [19:03:16] \o/ youtube started [19:03:23] focus! [19:03:24] nice. seems like lag is minimal. [19:03:27] yay urandom! [19:03:30] :) [19:03:40] harej: Yeah, new cameras coming soon. [19:03:50] welcome urandom :) [19:03:50] sounds real nice from here [19:03:57] chrismcmahon: great, thank you [19:03:58] I hit pause on the window and it stopped [19:04:15] cndiv - what camera [19:04:15] well, even the slide deck is blurry, leading me to believe it's just low-bandwidth [19:04:26] sometimes mikes are hot, but today is just right so far [19:04:34] victorgrigas: slick vaddio mechanical ones [19:04:54] model #? [19:05:01] couldn't tell you off the top of my head. [19:05:06] ok :) [19:05:23] Hi new folks! [19:06:13] you'll all see though, it's gonna be great. [19:06:49] that is a nice ship [19:07:11] be a shame if something were to happen to it. [19:07:24] "by the way, this is my ship" :O [19:07:38] it’s good to be the captain :) [19:07:43] :) [19:07:51] harej: it’s cheaper than housing in SF ;) [19:07:57] :( [19:08:04] heh [19:08:19] (not kidding, I’ve semi-seriously looked at it myself) [19:08:31] lvillaWMF: that was an extra reason we bought an RV, so that I can have a place to crash cheaply next All Hands/Dev Summit [19:08:31] He did tell me he had a boat but I didn't realize it was that big [19:08:54] crash might be the wrong word to use with RVs... [19:08:56] RoanKattouw: big family [19:08:58] Damon's presentation doesn't seem to acknowledge that VE has already "shipped" on several wikis. On frwp, itwp, VE has been "released" already. [19:09:18] guillom: I believe he’s going to get to that, but not sure [19:09:25] slide just said VE is default on 160 wikis [19:09:45] chrismcmahon: Missed that; thank you [19:11:00] (Side note: Are there really three typos in that chart? [19:11:10] #ocd [19:11:20] guillom: that's my count. [19:11:35] Yup I see 3 too [19:12:00] ship early, ship often. [19:12:36] I think we're back down to 0 in the latest slide. [19:12:41] agile FTW! [19:12:48] greg-g: where are you going to park it?! [19:13:09] rfarrand: Golden Gate, Holly Park, lots of options :) [19:13:18] It's not 800 bugs but 800 points [19:13:21] trust me, I know all about sleeping in cars and it is easy to get kicked out of cities [19:14:32] Thank you for the definitions clarification, Damon! [19:15:01] The only time I slept in cars was at interstate rest stops. [19:15:19] James' audio is alright, I'm assuming? [19:15:30] cndiv: +1 good [19:15:35] hospital parking lots always work, and the buildings are generally open 24/7 harej [19:15:36] cndiv: yes, quite posh. [19:15:36] brion: sweeeeet [19:16:08] dumb question someone may have already answered: what does it mean to be “live” on those 160 wikis? On by default? Or live as in opt-in/beta? [19:16:10] Context menus improvements: going out to wikis with VE next week [19:16:16] yay table editing [19:16:20] lvillaWMF, I believe that's the default count. [19:16:22] lvillaWMF: yes, on by default for newcomers. [19:16:35] not just newcomers [19:16:36] Pretty awesome: I think he said a million/month. [19:16:48] rfarrand: hospital lots, good call [19:16:49] on frwp it's on for all users iirc [19:16:52] "Suggesting targets for links" – I *love* this feature [19:17:11] guillom: on for all without an opt-out? [19:17:11] yes, on by default on those 160 wikis - the "Edit" tab points to VE [19:17:14] for all. [19:17:23] opt-out available [19:17:31] harej: wait until it shows a picture of the possible targets! [19:17:47] cool new context menu! [19:17:53] Looks great :) [19:18:06] niiiiiice [19:18:16] the new menu makes it a lot easier to visit/check links, try it on mediawiki.org [19:19:15] video worked, right? [19:19:23] oh you haven't seen it yet but when you do... [19:19:40] See the mocks at https://phabricator.wikimedia.org/T71474 to see an example of context menus [19:19:53] cndiv: yeah, kinda. [19:20:14] not all that smooth, but it got the point across. [19:20:15] ragesoss: yeah, I caught it myself. :-) [19:20:27] ragesoss: not sure what it was originally supposed to look like. [19:20:40] citoid is sweet [19:20:43] I think brad fixed some of that lag ;) [19:20:51] Citoid is the bestest. [19:20:54] though it screams of feature creep at this point ;) [19:21:00] WANT [19:21:08] lvillaWMF: all of WP is feature creep [19:21:17] lvillaWMF - I don't know [19:21:24] I kinda think it's necessary [19:21:37] THAT IS AWESOME [19:21:50] greg-g: I mean, yes? But it’s odd to have it on a slide literally right before “we’re feature finalized” [19:22:09] lvillaWMF: /me nods [19:22:10] * rdaiccherlb says this knowing that she isn't very good at citations yet [19:22:20] (not just have it, but have it with an implicit “this is not even alpha yet” statement) [19:22:25] rdaiccherlb: then how will we sort the elect from the masses, if it's not hard to add refs? [19:22:27] lvillaWMF, it doesn't say "we're feature finalized" [19:22:35] It says we want to be feature finalized at end of march. [19:22:50] Citoid has been in the works for a long time. [19:22:56] * rdaiccherlb makes tiny fist shake motion at ragesoss [19:23:03] superm401: you don’t get to be feature-finalized at the same time you also claim to be at release quality [19:23:16] you have to feature finalize and *then* get to release quality :) [19:23:39] lvillaWMF, citoid is landing a bit beyond the original projected ship date, for a number of reasons -- but has always been in the must-have list for en.wp [19:23:52] *clap clap clap* [19:23:53] greg-g: https://gerrit.wikimedia.org/r/#/c/193404/ ; heh, still wasn't merged [19:23:57] * aspiecat needs to work on that [19:24:00] [19:24:13] It's the infrastructure rather than the code that's behind for Citoid [19:24:18] On the wikis where VE is on by default, have we found that the amount of edits being made has increased? [19:24:43] aspiecat: I... don't know about this change/bug? [19:24:45] harej: James_F|Away and milimetric have lots of interesting data in that area but I don't know the answer to that offhand [19:24:58] greg-g: it was slowing down the wikitext editor a *lot* [19:25:33] hence the video by James [19:26:06] * rdaiccherlb claps [19:26:23] ooOOoooh the app [19:26:25] aspiecat: there is no varnish layer in mw-vagrant either which led to the smoke and mirrors [19:28:16] One of the things that was, and still is, a little confusing to me is "Similar pages". For disambiguations, it may not actually be pages that are very similar at all. [19:28:33] ragesoss: “pages with similar names”? [19:28:43] Are there plans to make that a label a little more dynamic based on the types of links being hidden? [19:29:04] "Other meanings"? "Other definitions"? [19:29:06] ragesoss: not that i know of, but file a bug in phab if you feel strongly :D [19:29:08] brion: pretty clunky. [19:29:10] Isn't it always disambiguation links? [19:29:13] I've started adding wikidata descriptions if I find stuff without them.... [19:29:14] Question: Is the app using the "Wikidata picture" in any case, or always the picture in the Wikipedia article? [19:29:33] guillom: it uses what’s returned from pageimages extension [19:29:45] which, ideally, should turn into the wikidata image at some point. maybe? [19:29:46] guillom: neihter, I think it’s using the PageImage [19:29:47] manybubbles: TFA doesnt' have a wikidata description. [19:29:49] manybubbles++ [19:29:50] I... didn't know about that extension :) [19:29:53] It really should allow specifying the lead image. [19:29:53] Thanks [19:30:09] guillom: https://www.mediawiki.org/wiki/Extension:PageImages [19:30:09] The first image may e.g. make perfect sense for the first section (where it is), but not make sense as a lede. [19:30:15] it would be great if we could explicitly semantically mark the 'lead image', cf https://phabricator.wikimedia.org/T90914 [19:30:17] yep, should be a task in phab if there isn’t already [19:30:18] thanks kaldari [19:30:24] if not please add :D [19:30:26] So that suggestions list, with images and wikidata descriptions, that's also what link suggestions in VisualEditor will look like in the future [19:30:52] I'm on Wikipedia Beta on Android and the article of the day, HMS Bellerophon, doesn't have collapsed section headings as, IIRC, it used to. Hard to scroll to the end [19:30:54] superm401: please link T90914 if you find another bug for this. [19:30:54] RoanKattouw: cool [19:31:10] spagewmf, swipe in from the right [19:31:16] for ToC [19:31:39] cscott, doesn't seem that related, but okay. [19:32:24] oh, ugh [19:32:30] hard to moderate with dead battery [19:32:34] That share feature is pretty cool. [19:32:35] quiddity: yes, that or tap the "bulleted list in the mirror" dot-dash icon [19:32:46] lol! love this concept! [19:32:49] so cool! [19:32:50] I didn't realize the collapsed section stuff went away [19:32:55] This is an awesome idea. [19:32:58] thats gonna be great :) [19:33:30] I like that it has the wikidata description [19:33:43] I want that [19:33:46] Dark overlay really won't work for some images, though. Will there be some image analysis or dark/light options? [19:33:59] I need to RT that [19:34:02] HMS Belleerophon article of the day has "Similar pages: Billy Ruffian (band)". ?? [19:34:07] ragesoss: I think I filed a phab about that a few weeks back [19:34:40] Specifying lead image: https://phabricator.wikimedia.org/T91683?workflow=create [19:35:18] I'm not sure about iOS, but Android 5 has a lib that generates a compatible palette from an image [19:35:29] Random question: What font is used in these apparently standardized WMF slide decks? [19:35:37] s/lib/framework utility/ [19:35:49] harej: I suppose heatherw knows? [19:35:51] lvillaWMF: "filed a Phab", I like it :D [19:35:53] sourcesans pro [19:36:04] source san...that :) [19:36:06] is it foss? [19:36:09] kristenlans: bug seems somehow wrong ;) [19:36:16] I thought it was philing. [19:36:23] ooh, I like philing [19:36:30] harej: http://www.google.com/fonts/specimen/Source+Sans+Pro [19:36:43] Isn't it always disambiguation links? <--- nope, lots of others. See enwiki's list at https://en.wikipedia.org/wiki/Wikipedia:Hatnote#Hatnote_templates [19:36:48] e.g. [[abaci]] [19:37:04] lvillaWMF: then you can even elide Phab. "I philed it a month ago." [19:37:11] quiddity, I mean for the app feature, not for all hatnotes. [19:37:27] it compiles all types. e.g. [[abaci]] ;-) [19:38:34] Ah [19:39:10] Whoo, 1000x increase. :) [19:39:30] Nice [19:39:34] Do we know that Knowledge Graph and Wikidata have the same inclusion criteria? Knowledge Graph may have hundreds of millions of items, but Wikidata has notability rules that may prevent it from getting *that* big. [19:39:53] Isn't Knowledge Graph more of a search engine index than a database? [19:39:58] superm401: i'd like to build a more general semantic tagging feature for images, rather than handling lead images as a one-off. [19:40:20] cscott, yeah, when I thought about it as a markup option (rather than a magic word) it seemed more related. [19:40:38] I doubt Knowledge Graphs is as useful for data re-users as Wikidata. [19:41:00] Also, Google is going to index everything from Wikidata, so we'll never "catch up". [19:41:28] superm401: I recommend reading the Knowledge Vault paper to better understand where they are headed. [19:42:07] omg, I think I have ebola. [19:43:03] lvillaWMF, didn't read the article (yet at least), but the Wikipedia summary ("mix of both high-confidence results and low-confidence or ‘dirty’ ones and machine learning is used to rank them") supports what I'm saying. [19:43:04] https://en.wikipedia.org/wiki/Knowledge_Vault [19:43:38] Plus the fact that we actually have PD dumps, obviously. [19:43:48] superm401: it certainly supports the “never catch up” point; not sure it supports the “useful for data re-users” point at all [19:44:02] in Firefox Android, clicking the Ebola article images does not bring up a Media viewer "gallery" where I can swipe through the images. Is that on a roadmap? [19:44:14] turns out lots of people are perfectly happy to pay license fees for data with known reliability ;) [19:44:23] lvillaWMF, does Google even license it out for a fee? [19:44:29] spagewmf, the MV experiernce in the app is quite a bit nicer already [19:44:34] lvillaWMF, yeah, if the automatic ranking were accurate and the data were actually available, it might be useful to re-users. [19:45:15] superm401: I believe they’ve announced limited access to an API, though admittedly I can’t find a cite on that right now [19:45:44] for data reconciliation iirc [19:45:46] Eloquence_mm: sure, but it seems the web site on mobile is showing me a media viewer, without < > arrows to advance to the other images. Lemme try chrome [19:46:00] spagewmf, that's correct, that feature is not implemented yet [19:46:20] caching and offline access is likely important for pay-per-byte users. once you download an article, we should really never download that data again. [19:46:51] superm401: yeah, as part of the Freebase retirement announce, they said there would be a forthcoming Knowledge-Graph based API [19:47:17] any questions from IRC? (other than “these 160 wikis, what happened to editing on them”) [19:47:39] language converter is one thing that happened [19:47:42] I didn't realize they even bought Freebase, let alone were killing it. [19:47:59] API is interesting and I'm sure will be used, doesn't allow the same use cases as dumps though. [19:49:06] Cool that they're supporting transferring some of Freebase to Wikidata. [19:49:37] there's also a different level of specialized templates, bureaucracy, etc, with different wikis [19:50:20] enwiki uses *all* of the features of wikitext, and all of the corner cases, & etc. some smaller wikis use a smaller set of mediawiki features. [19:50:45] very true cscott [19:51:42] on the other hand, nlwiki uses a {{largethumb}} template extensively, because they want *two* standard thumbnail sizes on their wiki. VE doesn't support editing {{largethumb}} images. [19:52:12] and on the gripping hand, at wikimania I heard of one small wiki *turning off* language converter, so that they could use visual editor. [19:52:32] so there are lots of different decisions that wikis make that bear on VE utility. [19:55:26] * JeanFred waves at rfarrand [19:55:32] JeanFred: ! [19:55:35] Hi [19:56:15] anything else from IRC? [19:56:17] cscott: s/wikis/opinionated cabal members on a wiki/ :) It's funny-strange that WMF has to engage with the concerns of 100s of wiki communities, but the communities don't have to agree amongst themselves. [19:56:18] * guillom chatouille JeanFred. [19:56:45] * JeanFred waves at guillom [19:56:54] spagewmf: well, the {{largethumb}} decision, just to take one example, far predates VE. [19:57:15] sounds like we’re good :D [19:57:16] +1 to considering global south! [19:57:28] \o/ [19:57:32] over and out? [19:58:25] spagewmf, decentralization leads to experimentation, which can be a really good thing. [19:58:26] http://www-personal.umich.edu/~jbednar/WIP/rfs%283%29.pdf [20:06:59] is there an online version of that burndown chart that Damon showed? [20:08:54] ragesoss: yes [20:09:03] don’t remember where though [20:09:30] ragesoss: it’s in Phabricator somewhere [20:10:08] yeah, it’s generated by Phab [20:10:36] I see... Sprint List. [20:10:43] the chart icon at the left. [20:17:13] is there a WMF-wide definition of a Point? [20:18:44] ragesoss, kaldari https://phabricator.wikimedia.org/project/sprint/view/1015/ [20:28:13] James_F: what is the VE definition of a Point, and is it the same definition for all teams that use sprints and burndown charts in Phabricator? [20:28:37] ragesoss: "1–3-ish hours' work for a reasonable developer". [20:29:06] ragesoss: No, points are strictly incomparable between teams, that's a fundamental tenet of agile pointifying. [20:30:02] ragesoss: Strictly incomparable === by definition you must not try to compare them; it's not just a "sometimes they won't compare", it means "they will never ever work and trying to compare them will lead to misery". :-) [20:30:37] James_F: velocity being incomparable is slightly different from the definition of a point being the same or different. [20:33:05] ragesoss: I agree. However, even with the same basal definition, what one team at one point in time considers to be "1–3-ish hours' work" or even what a "reasonable developer" is can vary significantly. [20:34:16] yeah. i'm familiar-ish with the definition that is based on "ideal hours", so 1 point is one hour of perfect, distraction free work for the developer who made the estimate. [20:36:23] have you been doing points / estimation / burndown prior to the sprint tracking in Phabricator? Or will this Q3 sprint be the baseline for establishing a velocity for your team? [21:03:59] ragesoss: No, and no.