[17:54:42] Hello. [17:56:47] * James_F is IRC liaison again today. [17:56:49] Welcome, all. [17:57:11] * brion waves [17:57:17] James_F is awesom-o [17:57:39] spagewmf: Let's not go that far. :-) [17:58:40] This week our compère is Deskana. [17:58:52] We'll have a run of presentations, and then take questions at the end. [17:59:06] However, if you have any queries, comments or questions, please ask them as soon as they occur. [17:59:22] First, this means that you won't forget them. [17:59:29] Second, it'll feel less dead in here. [17:59:51] And third, it'll mean others here can help answer your question, or refine it to get to the core of your question. [18:00:02] And we're about to start. [18:00:09] M.C. James_F is beyond compère [18:00:27] Dan Garry is the hôte. [18:00:27] giving MC Escher a run for his money [18:00:41] ey viola, it starts [18:00:42] hi multichill :) [18:00:49] * James_F grins at brion. [18:01:04] * multichill waves [18:01:22] Can everyone see and hear correctly? [18:01:29] Welcome, everyone. [18:01:39] No, the Google Hangout is not presenting anything. [18:01:43] had to reload my youtube, now it's working [18:01:47] oh hey i have an anniversary YAY [18:01:54] :) [18:02:06] *\o/* [18:02:11] brion: 9 years is incredible, a third of your life at WMF [18:02:25] jaufrecht: Refresh? [18:02:31] I cannot see/hear. reloading. [18:02:39] yes, it works. :-) [18:02:42] hi matt_flaschen. [18:02:53] Hmm.... where's the link to the live stream? [18:02:57] spagewmf: brion cheated a bit when he went to StatusNet :) [18:03:04] Hey leila. [18:03:05] :) [18:03:05] https://www.youtube.com/watch?v=aMQ5XVF0zm0 [18:03:05] ah, got it [18:03:30] well if i hadn't left and come back i think i'd be up to 11 ;) [18:03:31] Youtube works, but definitely no presentation in the Google hangout [18:04:29] Went to the wedding. It was awesome. :) [18:04:32] brion: And you'd have reached that milestone where you get that last week of vacation! [18:04:37] Can something be done about the audio quality? [18:04:51] it's not that bad Pine. [18:04:54] Don't think that proves it's a social network, though. People also meet each other at jobs, but I don't think that means a job is a social network (but maybe they both are!). [18:04:54] what IS jimmy wearing.... [18:05:07] Leila: I feel like I'm underwater listening to it. [18:05:12] agreed matt_flaschen. [18:05:26] it is most definitely "a network of social interactions and personal relationships" [18:05:30] Indeed. [18:05:31] ow yes, Pine. it's a bit like that. [18:05:33] yes. [18:05:40] Pine: Why would you listen to the meeting under water? [18:05:45] My videos of the wedding: https://commons.wikimedia.org/wiki/File:Walking_down_the_aisle_at_the_Wiki_Wedding.webm and https://commons.wikimedia.org/wiki/File:Vows_at_Wiki_Wedding.webm [18:05:47] oh cool! [18:05:52] guillom: :D [18:05:53] guillom: Didn't you see that Maldives Cabinet meeting photo? [18:06:00] new language, awesome. [18:06:03] Cheeky blog plug. [18:06:08] James_F: I don't think so. [18:06:09] lol [18:06:09] guillom: http://news.bbc.co.uk/2/hi/8311838.stm [18:06:10] guillom: Meetings are more entertaining if we hold them in different venues. [18:06:13] (The blog is awesome by the way, nudge nudge) [18:06:20] James_F: :) [18:06:22] That's cool that the university helped out. [18:06:25] * James_F grins at foks. [18:06:45] More info: https://blog.wikimedia.org/2015/07/15/konkani-wikipedia-goes-live/ [18:06:46] https://blog.wikimedia.org/ what this blog? [18:06:47] Jamesofur: Prada or her Miuccia line, circa 2012, I believe [18:07:00] yeah that's a better link, foks was faster at actually finding it [18:07:09] spagewmf: perfect [18:07:20] Jamesofur: that is the Wikimedia Blog... [18:07:27] yes...... [18:07:36] I think you missed the rest of the discussion :) [18:07:39] What's wrong with it? [18:07:43] nothing [18:07:48] nvm :) [18:07:51] It's summer. [18:08:11] marktraceur: It's Summer or Winter depending on your hemisphere. [18:08:16] Seeing Comms in perspective there... [18:08:18] the sound is a bit quiet, I have my volume up all the way and can mostly hear it. [18:08:20] marktraceur: It's not 'summer' [sic] whatever that is. :-) [18:08:57] I think Katherine's performance rates much higher than 1 out of 5, personally. :) [18:09:04] If you're doing arithmetics, you may be a summer. [18:09:07] James_F, language evolves!!!! [18:09:18] lots of red doesn't mean not lots of work, they had huge jobs that got offset by other issues and lack of resources [18:09:24] yup [18:09:39] * brion tempts the wrath of the bandwidth gods by watching video and updating ios beta at same time [18:09:48] speaking of summer :/ [18:09:52] I can't hear Kevin properly. has the mic changed? [18:09:53] James_F: I mean, it's summer in North America and Europe, kids on break are editing more [18:10:09] marktraceur: generally we have the reverse [18:10:13] can barely hear now [18:10:13] it goes down during the summer iirc [18:10:22] because people aren't at school etc [18:10:23] is someone taking questions here? [18:10:33] leila: no, he just speaks more softly than Dan. [18:10:37] * Jamesofur appoints James_F [18:10:39] aha guillom. [18:10:40] leila: yes, James_F is taking questions. [18:10:55] James_F: which pageview definition are these plots based on? [18:10:55] I can't really hear either. [18:11:03] leila: The official one. [18:11:03] There are 3 pageview definitions. :-) [18:11:20] +1 for more volume needed [18:11:23] byron: Could you tell Kevin to speak more loudly and/or hold the microphone closer to his mouth? [18:11:39] I'm afraid to increase the volume and Dan grabs the mic again [18:11:39] :D [18:11:41] James_F: Weird...OK then [18:11:41] guillom: Done. [18:11:46] The tiny yellow (apps) bar was 200 million hits per month. :) [18:11:53] That can't be percentage, looks like a 0-1 scale [18:11:53] I turned the gain up for Kevin [18:11:55] matt_flaschen: Yeah, 'tiny' is relative. [18:12:05] byron: we need much more [18:12:08] byron: I still can't really hear [18:12:22] Unfortunately turning up the gain means turning up the water effect [18:12:29] you should tell Kevin to hold the mic at 30 degrees from the horizon [18:12:29] :D [18:12:37] James_F: We should have a mandatory karaoke session so that all Metrics meeting presenters know how to hold a microphone :) [18:12:50] lol [18:12:50] I second the motion. [18:12:55] guillom: Or just use headset boom mics. [18:12:55] guillom: +1 [18:13:02] I can still hardly hear him [18:13:04] to be fair, guillom, it's very hard to hold the mic the way I was told to keep it for the last research showcase. :D [18:13:08] Everyone sings bohemian rhapsody, nothing else [18:13:20] Journey is off limits though. [18:13:29] still can't hear [18:13:45] only 50M searches per day? [18:13:46] I can hear fine. [18:13:57] It's a little watery but tolerable. [18:14:00] if Journey is off limits, I'm out. #bleedcitylights [18:14:19] OK, made a few adjustments [18:14:22] Don't stop, believin'... [18:14:24] James_F: any idea on the giant dip [18:14:26] (in both) [18:14:37] perhaps an outage... actually probably just outage [18:14:37] nvm [18:14:38] Total Eclipse of the Heart all evening [18:14:55] James_F, question for discovery team: Is there a plan to also work on precision (percentage of results that are relevant)? [18:15:01] Now, research update from Abbey [18:15:09] Jamesofur: In the edit performance results? Yeah, a major outage which (we think) was part of the transition from bits. [18:15:10] humm, so each unique visitor uses the search only 1.3 times per month? this is very low! [18:15:15] GOt abbey to hold the mic close [18:15:21] leila: Most people use Google. [18:15:24] byron: Thank you :) [18:15:40] How's the sound on the hangout right now? [18:15:41] yeah, but you expect more than 1 per month from each user, James_F. [18:15:54] ok for me, a little watery but ok [18:15:58] A lot of people probably don't even know Wikipedia has a search. [18:16:04] there are also search in App which is fully from App (though app users are not as many as desktop/mobile) [18:16:21] leila, I've seen a lot of people who exclusively access Wikipedia through Google. It wouldn't even occur to them to go to wikipedia.org or use a browser search bar. [18:16:22] sound's quiet again [18:16:42] leila: I don't think I have that hypothesis. [18:16:45] yeah, that can be matt_flaschen. a lot of low-hanging fruit for Discovery then. :-) [18:16:53] Are people using lav mics or larger mics? [18:16:57] byron is the one who can make changes to the mic (I think) [18:17:00] Larger mics [18:17:01] But fair question. Will ask it. [18:17:02] very quite, right? [18:17:06] quiet [18:17:16] leila, remember the guy we surveyed on the street? He was like "wikipedia... is that like... Google?" [18:17:17] For me volume is fine [18:17:18] byron: work on the 30-degree practice with the speakers. ;-) [18:17:19] Pretty quiet [18:17:29] owww I loved that guy milimetric. :-) [18:17:36] I'm at roughly 50% volume on my headphones and it's fine for me [18:18:11] leila, yep, there's a lot of room for the team to do unique search/discovery things that are easier to do with direct access to the data (in addition to other benefits of having our own strong search engine). [18:18:47] yup, matt_flaschen. [18:18:56] I love this research, I would binge watch those interviews if they were posted [18:19:07] milimetric, some of them are public. [18:19:08] ok, how's this now [18:19:21] milimetric: ask for them to become public? [18:19:26] It's ok byron, just have to turn up a bit [18:19:31] ok [18:19:35] better, thanks. [18:19:35] thanks, byron. [18:19:52] I just type "Bismillah wiki" into whatever bar is at the top of the Interwubwub [18:20:05] milimetric, https://commons.wikimedia.org/wiki/Category:User_testing . Not sure how much of the VE is there (or elsewhere public), but if not worth asking. [18:20:29] sweet :) thx matt [18:20:48] last edit was done in January 2013 matt_flaschen [18:20:55] There are relatively few videos there compared to how many they have. There is at least one VE video, but I think there is more public elsewhere. [18:21:22] ARGH. [18:21:29] LOUD [18:21:31] Sorry for ears bleeding. [18:21:32] leila, of the category page? Yeah, but that doesn't get bumped when people put stuff in the category. There is a VE one from March 2015. But they may have a more recently used Commons cat. [18:21:37] See, Aaron uses the microphone correctly! [18:21:38] J-Mo: are the videos Abbey referring to public? [18:21:40] ahhhhh much better volume [18:21:45] This is what happens when people ask for it to be 'louder'. [18:21:47] What just happened [18:21:56] not sure marktraceur [18:21:58] finally I can actually hear anything [18:22:02] transition issues [18:22:04] THANK YOU! i can hear again [18:22:05] I think Aaron had started earlier [18:22:06] marktraceur: I just skipped 5 minutes [18:22:09] hey, I don't need to have my volume all the way up [18:22:11] lol, I used the time lag to turn down the volume, thanks for the arghs :) [18:22:11] marktraceur: Aaron's sound was approximately 60 dB higher than the rest of the audio. [18:22:14] yeah, Aaron should talk on everyone's behalf. [18:22:17] Aaron is remote [18:22:23] FYI not in SF [18:22:33] Aaron's mic is also better than the mic in the office [18:22:36] can we turn the vol on the local microphone up by like a million times? [18:22:39] Less water [18:22:39] everyone should be remote tnegrin. ;-) [18:22:41] I think the levels are not mixed evenly for the online feed and the 5th floor mics [18:22:44] awight: Yeah, that's my issue too [18:22:59] victorgrigas: exactly [18:23:14] I move that the 5th floor mics be replaced with whatever kind of mic Aaron is using. [18:23:17] leila: some of them are, and are linked from here: https://www.mediawiki.org/wiki/Wikimedia_Research/Discovery_and_usability_for_brand_new_and_casual_editors [18:23:18] ah, thanks matt_flaschen. didn't notice that. [18:23:35] thanks, J-Mo. [18:23:39] brion: Not without feedback. [18:23:43] Pine - i dont think its the mics, I think it’s the mix [18:23:56] victorgrigas: It's also the way people use the mics. [18:24:06] victorgrigas: I'm referring to the water effect as well as the volume. [18:24:20] the 5th floor room mics jsut need to have their levels boosted either on the mic signal or in the mixer [18:24:20] Aaron's mike has more of a buzz effect, but it's better than the water effect. [18:24:44] guillom - good point, they might not be holding the mic close to their mouth [18:25:01] victorgrigas: Most aren't. [18:25:02] victorgrigas: For all the speaks so far it's been within 15cm of their mouth. [18:25:05] +er [18:25:15] or pointed at their mouths (uni mics) [18:25:27] victorgrigas: hence the suggestion that all presenters first must perform at a 5th floor karaoke session to get the mic setup correct. [18:25:30] That's more likely the issue. [18:25:33] yay for data! [18:25:38] I thnk its a mixer issue [18:25:51] leila: significant in a techical sense? [18:25:52] Pine: I don't know what in world you think we have "5th floor karaoke session"s, but sadly it's not this one. :-) [18:26:03] in statistical sense tnegrin. [18:26:14] p-value under a threshold [18:26:27] he should make that clearer — signficant but a small effect is counter intuitive for non-stats people [18:26:34] J-Mo, would be good to get those videos on Commons as well. There are tools for it, e.g. https://commons.wikimedia.org/wiki/Commons:YouTube_files#Moving_to_Commons_with_youtube2mediawiki . [18:26:35] thanks [18:26:41] On to Q&A [18:26:44] in practical sense, something that is not statistically significant, can be important, tnegrin. [18:26:46] OMG so many pings. [18:26:50] Asked my search question earlier. [18:27:18] "question for discovery team: Is there a plan to also work on precision (percentage of results that are relevant)?" [18:27:23] leila: for sure [18:27:39] noooooo [18:27:52] matt_flaschen: I've got it queued up. [18:27:58] tnegrin: yes, we tend to use terms that are technically correct but accessible, the usual way of framing the difference is significance vs effect size (was the effect “sizeable” as in, substantial and meaningful) [18:27:59] Owww [18:28:02] But no-one gave me a mic. :-( [18:28:05] Thanks [18:28:16] It's coming. [18:28:30] DarTar: yes — I appreciate that. [18:28:42] I would have guessed that most vandals were anon, so is this study missing a key group? [18:28:46] Claim one James_F. You are the voice of 56 viewers. :-) [18:28:53] * James_F grins. [18:29:05] I get more nervous when people refer to a significant difference outside of an actual statistical test ;) but I’m biased [18:29:08] I think it'll take a long time to socialize the "hey wikipedia is easier to edit now!!" news halfak [18:29:27] Here's the presentation I was talking about: https://www.youtube.com/watch?v=vGyrVg_qKSM [18:29:33] http://searchdata.wmflabs.org [18:29:37] ^ I talk about entry barriers and the potential of VE [18:29:39] milimetric: Yeah, I vaguely think it'll take another 5-ish years. [18:29:49] so it makes sense that it's not immediately increasing the newcomer numbers [18:30:16] To what extent does the transition to mobile platforms provide an opportunity to start fresh with no reputation for difficult editing? [18:30:22] If you like research and talking about social dynamics, check out our monthly showcases: https://www.mediawiki.org/wiki/Wikimedia_Research/Showcase [18:30:28] milimetric: Gut feeling rather than any data on that call from me. :-) [18:30:46] jaufrecht: I don't know. Good question. [18:30:57] Thanks for the answer. [18:31:04] can you guys hear S? [18:31:07] matt_flaschen: Any follow-ups? [18:31:07] good job Deskana [18:31:09] Not really [18:31:10] barely [18:31:10] milimetric, right. and there are social issues, right? VE can remove technical barrier, if your first edit gets reverted with no explanation or with very technical explanation, you will not come back. [18:31:13] q: for Discovery, just to clarify – are we disambiguating between searches that retrieve 0 results because there’s literally no match vs searches that retrieve nothing because of issues with our search engine, that plot doesn’t seem to reflect the difference [18:31:21] James_F, nope, appreciate the answer. [18:31:22] Thanks DarTar. [18:31:22] Everyone turn it down [18:31:23] no dynosaur [18:31:32] matt_flaschen I agree. I'll make a phab ticket for that. [18:31:37] yeah, leila, I agree with the social issues point, Aaron aluded to that too [18:31:44] uhun, milimetric. [18:31:48] In my opinion, not every zero-result search is wrong, though (though many certainly are), which was part of the reason for my question. [18:31:49] Question for 5th floor (maybe for Katherine or Juliet): is there a comms plan coming to market VE to Wikipedia readers so that James' 5 year estimate is accelerated? [18:32:01] Also, a correct zero-result search can encourage article creation. [18:32:20] Thanks, J-Mo [18:32:47] FYI on audio, we have a professional audio tuner slated to come to fix these issues out [18:32:52] good q leila ! [18:32:59] Pine: No, my 5 year estimate is /with/ aggressive outreach. [18:33:05] Aren't there 500 million unique readers? [18:33:13] So if there are really 50 million searches, it's only 0.1 [18:33:16] thanks James_F. [18:33:37] that figure is suspect matt_flaschen, we're trying to estimate uniques better [18:33:39] 50m per day vs. 500m uniques per month -> 50*30/500 = 1.3. [18:34:05] For certain values of '='. :-) [18:34:06] byron: happy to hear this. it's not good for remotees. [18:34:09] Coming from Google, not the Wikipedia search box? [18:34:27] matt_flaschen: 50M was daily, I /think/ [18:34:40] yes [18:34:42] Ah, right, Special:Search not the search box. [18:36:27] leila, in that case it's 3. [18:36:33] is the Mancunian saying DUI or DOI :) [18:36:40] James_F: question for Kevin. What pageview definition was he using when showing the pageview by access method? Is this the same pageview definition that Analytics API will expose publicly? [18:36:46] (assuming 500 million, but would be interested in hearing Analytics' best guess on actual unique count). [18:36:59] leila: Wilco. [18:37:05] leila: yes, its the new PV def [18:37:07] leila: yes [18:37:12] matt_flaschen: 500M/(50Mx30)? [18:37:12] ah https://en.wikipedia.org/wiki/Digital_object_identifier [18:37:14] Oh, OK, won't ask. [18:37:30] since May 1 there should be no more ref to the legacy data [18:37:34] thanks DarTar, milimetric. [18:37:43] Thanks everyone. [18:37:47] stop watching if you are remote or you will get hungry! [18:37:48] You can ask if Leila's purpose was to plug the API and put pressure on us :) [18:37:56] Q for aripstra: Is there a way to do this design research earlier in the dev process? [18:38:01] noo, I didn't have an agenda milimetric. ;-) [18:38:08] leila, 500 million/(50 million * 30) is users/search, and is 1/3. [18:38:09] milimetric: :-) [18:38:14] search/users 3 [18:38:35] you are right matt_flaschen [18:38:41] Is that a record for shortest metric meeting? [18:38:43] FYI: the restbase endpoints of the API are nearly done, the project's looking good [18:38:55] great milimetric. :-) [18:39:25] Dan’s favorite lunch and shortest meeting ever. Coincidence? [18:39:43] 3 per unique is not very low, matt_flaschen, given that we know not that many people use internal search [18:39:50] thanks for pointing my mistake out. [18:40:09] mmmmm lunch [18:40:37] leila, I said it was 0.1, until James_F pointed out that was daily... :) Definitely worth thinking about, so thanks for bring it up. [18:41:08] leila, 3 per unique is indeed pretty good. [18:41:24] maybe even too high, matt_flaschen. ;-) [18:41:59] Heh [18:42:19] i guess i'll go eat lunch too now. i finally finished seeing the metrics with a 10 min .lag :) [18:43:13] ha, how come there was that much lag subbu? [18:44:00] leila, i spaced out and didn't start watching till 10 mins. after :) [18:44:13] ah! [18:45:01] By the way, random question: wasn't there a motion to burn Gerrit awhile back and replace it with something else? [18:45:05] Is that happening? [18:45:39] Pine: it's happening slowly [18:45:50] ok thanks [18:45:51] Phabricator can do all the things Gerrit does except ... not as cleanly [18:46:01] So we'll migrate to Phabricator probably... at some point [18:46:30] but there's no urgency, a lot of the value of Phabricator was standardizing how we project manage so the community can just learn one tool and understand all the different projects / teams [19:49:02] Pine: git browsing is moving from git.wikimedia.org to phab:diffusion. Git commits are already viewable and searchable in phabricator. But migrating code review (and gerritbot, jenkinsbot, zuul, jenkins, ...) from gerrit to phab:differential sounds like "a giant rathole", https://phabricator.wikimedia.org/T18