[14:52:05] neilk_: Thought you'd like to know about https://secure.wikimedia.org/wikipedia/mediawiki/wiki/Special:Code/MediaWiki/86054 , see also comments section [16:43:49] RoanKattouw_away: thanks [17:52:49] alolita: AFT scrum is on now, which ext? [17:58:13] alolita: I'm in 2003 now, I briefly heard you [18:12:31] fuuuuuuuck [18:12:49] who came up with RJ11 connectors with 3 pairs :/ [18:19:09] TrevorParscal: Actually, about scalability, one thing that bothers me a little bit is ClickTracking volume [18:19:19] That table has almost 10M rows now on enwiki [18:19:26] I think we can afford to toss some of the old data [18:19:29] we can back the tracking down to a percentage [18:19:36] I will bucket that too [18:20:14] OK [18:20:23] And, I just found out the clicktracking table has no indexes. None [18:20:46] <^demon> Someone called me from Kentucky asking for a Mike Holler. [18:20:48] *^demon sighs [18:20:51] That's slightly disturbing [18:21:09] Wrong Mike? [18:21:27] <^demon> I told him that :) [18:21:55] <^demon> He started asking about a machine shop or somesuch. Definitely knew it was the wrong number then. [18:28:43] Mike Hunt [18:28:45] TrevorParscal: Hmm actually the large volume seems to be the result of a fluke in March, probably me having turned on clicktracking for sidebar clicks and suck [18:30:00] Ooooh [18:30:03] SectionEditLinks stuff [18:30:30] *That's* why we were getting half a million click tracking events per day for a while in March [18:34:20] And sure enough, 8.4M/9.6M click tracking rows are SEL [18:34:26] So that allays most of my fears [18:53:10] mmerz: hi [18:53:21] is denis joining us [18:53:42] hi alolita [18:54:37] Yes he will be here in a minute [18:55:14] (or second) [18:55:23] hi denis [18:55:28] hi everybody [18:55:30] mmerz: ok cool [18:55:35] denis: hi! [18:55:35] hi mmerz alolita [18:56:07] denis, mmerz: apologies about not being able join in yesterday [18:56:26] denis: i have looked at the docs you sent - thanks [18:57:04] alolita: i hope they helped in understanding our goals [18:57:20] yes they did [18:57:29] and i hope even more we could ensure, that there isn't too much work to do [18:57:36] so here are my initial questions - [18:58:02] a. we've a couple of tools that we've used for surveys [18:58:13] mmerz had mentioned a different tool that you use [18:58:41] EFS survey? [18:58:47] am i correct in assuming that your team will handle building the survey in the survey tool you use [18:59:05] and we will help in hooking up the survey to central notice [18:59:16] exactly! [18:59:41] with specific requirements that you've already identified [18:59:48] yep [19:00:04] ok cool [19:00:52] i am assuming that you're not adding any editor/user data to the survey [19:01:26] as mentioned, this would be beneficial, but it is not essential [19:01:38] once the survey is clicked on by the user via central notice, the user switches over to the survey system [19:01:38] we could use the existing implementation [19:02:14] yes, the user data could be sent to the survey via url parameters in our implementation as well [19:02:57] mmerz: do you collect them from de.wp logins and then pass them via web parameters to the survey [19:03:51] yes we could do this, but we should do so only, if it means little changes to your existing code [19:04:20] well - it depends on what kind of data you're pulling in for each user [19:04:44] if your existing implementation seems to work well, we would like to use this data, but it is not essential as the essential questions are in the questionnaire itself [19:05:25] ok that's good to know [19:05:56] so let's talk about the cn display first [19:06:47] the survey is displayed only once per user (both anonymous and logged-in) [19:06:55] correct [19:07:08] there is no need to pull more data than used for the editor survey (less the data that might not be available for DE) [19:07:49] mmerz: cool [19:08:10] there is no concept of a sample lottery right now [19:08:31] it should be easy to implement: [19:08:47] randomizing 5 out of 100 [19:08:53] right? [19:08:53] exactly [19:08:57] ok cool [19:09:38] is there anything else in the launching of the survey that I am missing [19:09:38] the lottery value is saved to the users cookie, so ever user takes part in the lottery only once [19:10:04] ok cool [19:10:12] the most tricky part will be to survey anonymous users (not logged it) [19:10:15] yes [19:11:01] so let's talk about the survey results [19:11:06] the banner should be visible to editors only (but including anonymous editors) [19:11:16] ok cool [19:11:32] what do you mean by anonymous editors [19:11:34] this should be allright to implement as well, right? [19:11:46] how is that tracked? [19:11:50] users that are not logged-in [19:12:19] but how do you measure them to be editors [19:12:30] is it their first time click on edit? [19:13:03] yes, this would be a good way of doing it [19:13:23] as soon as they perform an action that only editors would perform [19:13:36] let me talk to my developers about it and see what is a good measurement [19:13:37] eg. viewing the editor interface [19:13:52] that would be great! [19:14:12] so tell me about how you collect the data from the survey [19:14:51] could you please specify the question [19:15:29] what is format of the data that is output from the survey; what is the format of the file where is data is saved [19:15:49] ah, ok, I see [19:16:14] the raw-data output can have a variety of formats [19:16:30] we will use SPSS for quantitative analysis [19:17:15] the formats should however be convertable to any format [19:17:26] what format du you use? [19:21:01] in any case, we can make the data available to you in any format you like [19:21:06] all surveys are output by the survey tool so we prefer csv (comma separated value - open formats) [19:21:25] but i guess we don't care because you would be processing the data for analysis [19:22:03] csv is fine [19:22:56] yes, we will be processing the data for analysis, but you have access to it as well, so we will make it available in csv [19:23:50] ok cool - that works well :-) [19:23:59] i think that's all the questions i have for now [19:24:16] cool [19:24:23] what i will do is walkthrough these requirements with my dev team [19:24:37] and then let's touch base next week to finalize a timeline [19:24:46] for deliverables [19:24:57] what would your first estimate be? [19:25:14] what is your target timeline for the survey to be run live [19:25:27] alolita: can we assume, that our proposal concerning a milestone (survey in june) is available? [19:26:06] denis: can you clarify what you mean :-) [19:26:28] when in june [19:26:36] mid june [19:26:46] the original plan was first week of june [19:26:51] i see [19:26:57] yes [19:27:03] but this will depend on the WMF editor survey as well [19:27:08] so our dev team will be in berlin second week of may [19:27:28] do you have a timeline for the WMF editor study? [19:27:33] so let me take a look at our deliverables for other projects and available resources [19:27:53] we just completed the editor survey which should be running next week [19:28:23] there should be enough time between both surveys [19:28:31] for the next few weeks and propose a timeline then [19:28:47] its completely driven by availability of our engineers [19:29:19] cool - let's plan to talk same time next friday then - does that work for both of you [19:29:33] alolita: if needed, we might be able to help with additional ressources [19:29:55] denis: cool - so let's talk earlier next week and figure this out [19:30:01] since we really don't have a lot of time [19:30:18] if we target june first week as a launch date [19:30:44] yes, lets target this launch date [19:30:45] we only have about 4 weeks of actual time [19:31:05] true [19:31:05] let's target june 6 [19:31:15] or june 15 [19:31:37] all development, testing, localization done before that [19:31:55] alolita: sounds great [19:32:06] yes that sounds great [19:32:22] and if we don't have all the dev time we need; we borrow from you :-) [19:32:37] and of course need you to test and verify as we make progress [19:32:49] so let's sync up next week [19:32:56] what day works for you [19:33:04] mondays and tuesdays are not good for me [19:33:10] alolita: mmerz1 isn't available next week [19:33:16] ouch [19:33:21] I am on the move most of next week, so how about monday the 25th [19:33:24] ok - so let's stay on email [19:33:29] wanted to propose monday then [19:33:36] yes, or email! [19:33:37] 25. April [19:33:37] thru next week and sync up on the 25th [19:33:41] at the same time [19:33:45] coll - fix it [19:33:46] 12 noon PDT [19:33:50] cool [19:33:56] ok awesome [19:33:59] great [19:34:13] perfect [19:34:17] denis, mmerz: thanks much and i will be sending you email :-) [19:34:30] ttyl! [19:34:33] bye [19:34:34] thank you! [19:34:36] thanks for speeding up things, alolita [19:34:44] great evening now [19:34:46] ;) [19:34:49] denis: glad to help; bye! [19:34:52] bye! [19:34:57] bye