[00:00:20] Ooh. Thanks Ironholds [00:00:58] np! [00:52:41] Ironholds, I just saw your tweet about stemming. You probably know, but, I added a link to elasticsearch's docs about stemming in various languages at https://www.mediawiki.org/wiki/Help:CirrusSearch#Stemming [00:53:02] (specifically: http://www.elastic.co/guide/en/elasticsearch/reference/current/analysis-lang-analyzer.html [02:28:06] quiddity: yay! [13:37:57] goood morning science people [13:37:59] And happy Friday! [15:36:10] hey halfak :) [15:48:04] g'morning Ironholds :) [15:48:23] Been listening to Hail Mary mallon [15:48:30] Hard to do while I'm writing though :S [18:06:06] halfak, whatcha writing? [18:06:34] Writeup from the CSCW workshop on next steps for the field :) [18:06:40] oh, cool! [18:06:56] Next bits of theory we should develop and ways that orgs like WMF can increase access for researchers. [18:07:26] dang. I wish I could think up that kinda stuff. Mostly I just do papers that look interesting and sometimes they are. [18:07:33] and then, ooh, other unrelated shiny interesting thing [18:08:43] did I miss standup? I think I missed standup [18:09:52] You did. But we just did it really fast and then went to breakfast with Lila. [18:10:56] ah well [18:16:31] * Ironholds is spending his morning studying. Is fun! [18:16:39] and then I have to spend my afternoon webscraping the Harvard website [18:17:59] halfak, can I ask an incredibly stupid question? [19:29:18] halfak: is felipe working on a reskin of the developer docs? [19:41:14] halfak, around when you are :) [19:43:36] Ironholds: Thank you for setting up the meeting. I've got some clarity from Erik about the output he's expecting for this project. I'll send you a note shortly. [19:44:08] np! [19:49:33] Which Felipe [19:49:46] Ortega [19:49:55] I think [19:50:24] Oh. That would be wonderful ;) [21:42:10] halfak: batcave? [21:42:22] Oh! I forgot. [21:42:37] I'm out at a coffeeshop now, but I'll join up and stick a headphone in one ear :) [21:42:59] ...or I forgot my headphones :( [21:43:10] Wait... no i didn't [21:55:28] leila, got some time to catch up today? [22:02:58] * Ironholds pokes halfak [22:03:00] a question! [22:25:04] J-Mo, so, I opened that document [22:25:15] "mean joins per idea" [22:25:21] arithmetic or geometric? ;) [22:26:58] which do you think is more appropriate to the data, Ironholds? [22:27:51] J-Mo, I dunno, I'll look at the probability distribution! [22:28:12] but "arithmetic mean" is the trigger phrase for halfak's KGB sleeper agent conditioning [22:28:29] I wanted to see if referencing it would set him off shouting in russian about the Fifth Internationalists [22:31:04] the distribution is probably highly skewed, though not as much as the distribution of endorsements. So arithmetic mean is probably not very useful. That's why I usually provide median as well, when dealing with these kinds of data. Happy to let you investigate the the numbers and find the best fitting statistic :) [22:31:50] *thumbs up* [22:31:52] I'm down with mean [22:32:06] the entire reason to not use arithmetic mean is it ends up pointing to users who don't exist. Median solves for that for me :) [22:32:13] bleh. down with median! [22:32:15] I brain gud [22:49:47] * halfak loves receiving regression outputs with questions ^_^ [22:51:02] hahah [22:52:46] çrsj [22:52:51] ugh [22:52:55] Wrong keyboard, sorry [22:53:25] #eslproblems [22:53:42] heh [22:54:03] * guillom sings the week-end song. [22:54:25] bbl [23:16:06] Ironholds: sorry. I was in a series of meetings. Will you be around later tonight or in the weekend? [23:16:21] I need to finish something in couple of hours. [23:18:21] leila, I'll be around Saturday early and Sunday all [23:18:31] ooki. I'll ping then Ironholds. :-) [23:19:22] kk