[15:58:33] tnegrin, can you attach a hangout to the meeting? [15:58:42] yah [16:02:29] danke! [16:11:10] All the echo pings --> DarTar [16:11:32] * halfak is fleshing out metrics pages and starting conversations about names and stuff. [16:26:02] halfak, can you give my apologies to the group? Won't be in the meeting. Mobile stuffs. [16:26:16] * YuviPanda pats Ironholds [16:26:19] S'ok [16:26:20] will do [16:26:38] ta :) [16:29:42] Ironholds: no worries [16:29:50] halfak: coming [18:23:53] o/ YuviPanda. What do you think about scheduling an office hours for quarry? [18:24:01] oh [18:24:05] think it'll be useful? [18:24:24] we haven't really publicized it on ml and stuff [18:24:29] Yup. The features have changed a bunch since you first deployed. [18:24:31] ml? [18:24:58] Oh yeah! It still hasn't gone to wiki-research-l has it? [18:25:00] mailing lists [18:25:00] no [18:25:05] or wikimedia-l or wikitech-l [18:25:40] OK. That sounds like a good idea. [18:25:47] Do you still have the link to that etherpad? [18:25:56] Also, we might want to do a documentation push [18:27:15] halfak: yeah, we should [18:27:20] also some more messaging on the home page [18:27:39] * halfak is updating our one screenshot [18:28:00] heh :D [18:29:13] Is there a list of published queries somewhere? [18:29:21] halfak: nope [18:29:30] I should add filtering capabilities, etc to that page [18:29:33] the the 'all queries' page [18:29:52] Indeed. Seems like default filter should be "published only" [18:29:56] hmmm [18:29:59] maaaaybeeee [18:30:33] yeah [18:30:36] Anyway. let me not interrupt you. I'll do a little more with the docs today and start thinking about what they should look like. [18:30:48] halfak: no, am just in a meeting [18:30:56] this is one of those 'I am just going to sit and listen' meetings [18:31:00] so no interruptions at all, really [18:31:08] halfak: do file bugs for things we still need to do? [18:31:15] Will do [18:31:18] need to add element [18:32:15] * halfak just got bit by revision_userindex again [18:46:05] <halfak> YuviPanda, can I document the download links for a query? I assume they are stable (e.g. http://quarry.wmflabs.org/run/1535/output/0/json?download=true) [18:46:40] <YuviPanda> halfak: yes, but the second '0' isn't always zero [18:46:47] <YuviPanda> I'm going to implement a 'latest' for that [18:46:59] <halfak> Gotcha. It's a query revision number? [18:47:00] <YuviPanda> you don't need the download=true for CORS [18:47:04] <YuviPanda> query run number [18:47:06] <halfak> CORS? [18:49:27] <halfak> Can you do a JSONP request? [18:49:53] <halfak> Hey HenriqueCrang! [18:50:03] <HenriqueCrang> hi halfak ! [18:50:09] <halfak> How you doin'? [18:50:17] <HenriqueCrang> crazy :D [18:50:22] <YuviPanda> halfak: no, you can just do a regular $.get [18:50:25] <YuviPanda> and it will 'just work' [18:50:34] <YuviPanda> because I'm setting appropriate CORS headers [18:50:41] <YuviPanda> it won't work on ancient browsers, but that's ok :) [18:50:41] <HenriqueCrang> finishing lots of stuff for the brazil program in our last month [18:50:50] <halfak> YuviPanda, But cross-domain? [18:50:56] <YuviPanda> halfak: indeed [18:51:08] <YuviPanda> http://caniuse.com/#search=cors [18:51:23] <halfak> I didn't know this existed. Cool! [18:51:40] <halfak> HenriqueCrang, I can imagine. [18:51:51] <halfak> I've been making progress on the revscoring project. [18:52:09] <YuviPanda> halfak: :) [18:52:11] <halfak> I was wondering whether your silence on the email thread was due to overwork or lack of interest. [18:52:15] <halfak> HenriqueCrang, ^ [18:54:04] <HenriqueCrang> ow, sorry! [18:54:37] <HenriqueCrang> people in brazil are sending a lot of stuff to my personal email account, I've missed the other inbox in the last day [18:54:49] <HenriqueCrang> Still interesed in our stuff! [18:54:55] <HenriqueCrang> I'll check that now! [18:54:57] <HenriqueCrang> sorry! [18:55:26] <halfak> No worries. You get to be busy. :) [18:55:44] <halfak> I just want to know if I should keep sending you on the thread or not. :P [18:58:03] <halfak> I have complete unit tests for the feature garden. It seems to work very nicely. [19:00:49] <subbu> hi folks .. is there a youtube streaming link for ellery's talk? or is just the hangout? [19:00:55] <DarTar> subbu, nope [19:01:02] <DarTar> just hangout I’m afraid [19:01:07] <subbu> k [19:12:44] <halfak> I wonder what biases are introduced by disregarding earmarks that do not appear in tables. [19:14:40] <halfak> Wow. That's some high AUC [19:41:01] <halfak> Wooo Python! [19:41:14] * halfak is literally playing with sklearn right now. [19:45:13] <ottomata> Whoa, I didn't know this channel existed [19:45:14] <ottomata> HI! [19:45:43] <halfak> WAT [19:45:45] <halfak> o/ [19:45:48] <halfak> Hi ottomata [19:45:49] <halfak> :) [19:45:56] <ottomata> so many channels! [19:46:12] <halfak> indeed [19:46:43] * Ironholds has spent all day doing recursion [19:46:52] <Ironholds> friends don't let friends play with session generation [19:47:00] <halfak> :P [19:47:05] <halfak> Doing sessions stuff? [19:47:08] <Ironholds> yep [19:47:14] <halfak> I wanna look at that expanded dataset [19:47:15] <Ironholds> it has generated, this entire afternoon, 20SLOCs [19:47:24] <Ironholds> but goddamn are they efficient SLOCs. [19:47:28] <tnegrin> any questions on IRC? [19:47:30] <Ironholds> and they cover aaall the edge cases. [19:47:34] <tnegrin> for Ellery [19:51:16] <Ironholds> ...aaand that hyper-efficient code is apparently completely irrelevant [19:53:21] <tnegrin> historical reasons [19:53:50] <Ironholds> ? [19:54:16] <tnegrin> his school wanted to do historical work on earmarks (bush v. obama eras) [19:55:25] <Ironholds> ahh [20:05:03] <subbu> halfak, i was spacing out .. but what was gabriel's qn? [20:05:15] <halfak> Why aren't we building classifiers right now? [20:05:15] <halfak> A: We are :) [20:05:21] <subbu> ah, ok. [20:06:57] <subbu> halfak, yes, we should meet up. [20:07:31] <halfak> subbu, it's been too long. [20:07:35] <subbu> coincidentally i was looking LDA topic classifiers before the talk ... it is a bit too much dense math for me. [20:07:45] <subbu> my math has rusted quite a bit since undergrad days. [20:08:00] <Ironholds> too dense math for The Parsoid Guy = we're boned, everyone go home and say goodbye to your kids. [20:08:04] <halfak> Oooh Dirchlets. [20:08:08] <subbu> http://www.cs.princeton.edu/~blei/kdd-tutorial.pdf [20:08:25] <halfak> Dirchlets are the generalization of my favorite probability distribution, the Beta. [20:08:43] <subbu> maybe you can help me understand it some time. [20:09:12] <halfak> Not sure I'll have some insights, but I can be your study buddy. [20:09:51] <subbu> Ironholds, not that much math in parsoid. ;) [20:10:29] <subbu> halfak, where do you work out of these days? home / coffee shop? [20:11:01] <halfak> Usually from home. I bought a house and then traveled a bunch so I've barely seen it. [20:11:07] <halfak> I go to UMN once/week though. [20:11:11] <subbu> oh, congrats :) [20:11:17] <halfak> :D [20:11:25] <halfak> Speaking of which I need to run to UMN in a couple minutes. [20:11:28] <Ironholds> it's SUCH AN AWESOME HOUSE [20:11:31] <subbu> k. [20:11:38] <subbu> oh, the one you visited? [20:11:39] <Ironholds> I went into it and looked around and went; I want to be Aaron when I grow up. [20:11:50] <halfak> lool [20:11:55] <subbu> :) [20:12:09] <Nettrom> Ironholds: I'm sure you're only wanting that for his citation count [20:12:17] <Ironholds> halfak, hey, you have a nice house full of nice people and an excuse to beat on things with a sledgehammer [20:12:25] <Ironholds> I have a guitar and gray hairs [20:12:36] <subbu> how is boston working for you? [20:12:39] <Ironholds> Nettrom, hey, I'm publishing one paper and one note at CHI this year tyvm. [20:12:41] <Ironholds> I'm getting there! [20:12:53] <Ironholds> subbu, pretty well sof ar! I'm looking forward to actually having money and so being able to buy, e.g., bookcases. [20:12:54] <halfak> Nice Ironholds [20:13:05] <Ironholds> halfak, I am sorry mobile sucked in all the session time :( [20:13:09] <Nettrom> Ironholds: awesome, gogogo! [20:13:30] <Ironholds> I was wondering if we should pitch that to bossland as a useful real-work project. [20:13:41] <Ironholds> it's an opportunity to test the robustness of our session analysis methodology [20:13:56] <Ironholds> and it's one that allows us to be transparent and encourage other people developing comparable to datasets to actually make those datasets comparable. [20:14:21] <Ironholds> also it lets me snark at Carsten Schaepers and his terrible paper of doom (tm) [20:15:36] <subbu> alright, time to head to a coffee shop now. [20:16:16] <halfak> Ironholds, i'm game. [20:16:21] * halfak runs out the door. [20:16:24] <halfak> --> UMN [20:17:13] <Ironholds> take care! [20:51:48] <DarTar> hey Nettrom: yt? [20:52:28] <ottomata> hiya ewulczyn___ [20:52:58] <ottomata> we want to get you your shell access, but first, we need you to make a wikitech account [20:53:00] <ottomata> you can do so here: [20:53:00] <ottomata> https://wikitech.wikimedia.org/w/index.php?title=Special:UserLogin&returnto=Main+Page&type=signup [20:53:23] <ottomata> this will get you in to the LDAP system, and then we can match your LDAP userid with the shell one we are going to give you. [20:53:45] <DarTar> ottomata: he’s AFK, but I’ll grab him as soon as he pops by [20:53:49] <ottomata> k cool [21:11:25] <ewulczyn___> ottomata: my wikitech Username is Ewulczyn [21:12:10] <ottomata> hmm [21:12:13] <ottomata> ah, that's probably why I couldnt' find it [21:12:14] <ottomata> ok [21:13:19] <ewulczyn___> the capital? Yeah I wanted it to be Ewulczyn (WMF), but thats not allowed [21:15:42] <DarTar> ewulczyn___: so you didn’t go for Ellery in the end? [21:15:46] <Nettrom> DarTar: here [21:15:53] <DarTar> woah that,s a lot of underscores :) [21:15:57] <DarTar> hey Nettrom [21:16:04] <DarTar> Nettrom: check your inbox [21:16:27] <Nettrom> DarTar: I saw the email from Toby, if that's what you're referring to [21:16:46] <Nettrom> oh wait, I see it now [21:16:47] <DarTar> the ops request ticket with my note [21:17:04] <DarTar> so, it’s happening – sorry that it took so long [21:17:24] <DarTar> I was pinging you to ask your preferred username but ops will get in touch to ask [21:18:00] <Nettrom> DarTar: okay, that sounds good... I've been busy with my CHI paper, so delay is not a problem, but maybe for you it is :) [21:19:07] <DarTar> Nettrom: that’s fine, we’ve been onboarding a new team member (say hi to ewulczyn___) and prepping for the upcoming quarterly meeting [21:19:24] * Nettrom says hi to ewulczyn___ [21:19:26] <Nettrom> :) [22:29:09] <ewulczyn___> atglenn: Hey Ariel, this is what I get when I try to ssh to stat1003: [22:30:07] <Ironholds> ewulczyn___, ariel ain't in this channel, but what's the error? [22:31:04] <ewulczyn___> ssh -vvv stat1003.eqiad.wmnet [22:31:04] <ewulczyn___> OpenSSH_6.2p2, OSSLShim 0.9.8r 8 Dec 2011 [22:31:04] <ewulczyn___> debug1: Reading configuration data /Users/ellerywulczyn/.ssh/config [22:31:04] <ewulczyn___> debug1: /Users/ellerywulczyn/.ssh/config line 17: Applying options for *eqiad.wmnet [22:31:04] <ewulczyn___> debug1: Reading configuration data /etc/ssh_config [22:31:04] <ewulczyn___> debug1: /etc/ssh_config line 20: Applying options for * [22:31:04] <ewulczyn___> debug2: ssh_connect: needpriv 0 [22:31:05] <ewulczyn___> debug1: Executing proxy command: exec ssh -a -W stat1003.eqiad.wmnet:22 bastion-wikimedia [22:31:05] <ewulczyn___> debug3: Incorrect RSA1 identifier [22:31:06] <ewulczyn___> debug3: Could not load "/Users/ellerywulczyn/.ssh/private_key" as a RSA1 public key [22:31:06] <ewulczyn___> debug1: permanently_drop_suid: 502 [22:31:07] <ewulczyn___> debug1: identity file /Users/ellerywulczyn/.ssh/private_key type -1 [22:33:33] * Ironholds blinks [22:33:33] <Ironholds> huh [22:34:26] <ewulczyn___> Ironholds: I'm not sure. How can I drill deeper than running ssh with verbose output? [22:34:27] <Ironholds> cygwin? [22:34:31] <Ironholds> ahh [22:35:11] <Ironholds> hrm. Poke ariel, then :/. apergos in wikimedia-staff [22:50:24] <halfak> Hey gwicke. I won't have time to write an email. on my way out of town. I'm going to throw some links at you. https://meta.wikimedia.org/wiki/Research:Revision_scoring_as_a_service [22:57:00] <halfak> https://meta.wikimedia.org/wiki/Research:Ideas/WikiCredit:_Measuring_value_added_to_Wikipedia [22:57:03] <halfak> That one too. [23:32:02] * gwicke is too late to say thanks