[10:27:47] Hey YuviPanda. I was just swinging by to see if you were around. [10:27:48] :) [10:28:12] There's a thread on analytics-l about fixing the batch inserts EL bug. Looks like they won't pick it up this quarter. [10:28:18] * halfak gets link [10:28:32] halfak: heya [10:28:41] http://lists.wikimedia.org/pipermail/analytics/2014-August/002391.html [10:28:41] halfak: oh :( [10:29:11] Something that we could still look into if we find someone with the bandwidth. Not sure we should prioritize that though. [10:29:50] Dan seems to think that the batch inserts strategy wouldn't be that hard to push on. [10:31:38] halfak: true, but my instinct is that it's a bit harder to do 'right' i.e. ensure throughput while also not dropping events [10:31:53] That's my impression too. [10:31:55] but I haven't actually looked at the code [10:32:50] See dan's message here: http://lists.wikimedia.org/pipermail/analytics/2014-August/002399.html [10:33:06] halfak: yeah, I read the entire thread [10:33:51] halfak: store_sql_event doesn't take a list, just individual ones. [10:33:51] But seriously though, don't let me rope you into more work. I think we should prioritize instead. Right now, rather than pushing more work on you I should really be getting Quarry set up so that I and start pushing changesets. [10:34:55] * halfak appreciates the flow diagram in comment form. [10:35:14] halfak: :D I will do this but yeah, not now. quite a lot of quarry stuff left [10:35:22] +1 [10:35:32] halfak: you saw the user page? [10:35:38] User page? [10:35:47] halfak: http://quarry.wmflabs.org/Yuvipanda [10:35:54] Ooh! [10:36:22] Nooo! You've run more queries than I have. [10:36:25] halfak: they're linked to from the query lists page [10:36:29] I have to get my internet points up. [10:36:53] Wait a second.... are "Thanks" public? [10:36:58] http://quarry.wmflabs.org/query/210 [10:36:58] halfak: yeah [10:37:02] WHAT! [10:37:09] I was told otherwise by many people. [10:37:14] not who is thanked [10:37:25] Quarry is solving problems for me. :) [10:38:02] Oh. OK. That might still be useful for my purposes. (Actually external researcher's purposes.) [10:38:11] Discoverability FTW! [11:12:35] halfak: heh :) [11:12:50] halfak: we talked about tagging the queries as well [11:13:10] +1. Queries to the logging table would be a cool one to look at. [11:13:23] It would help me develop the MW Events spec. [11:15:54] halfak: yeah. we could auto tag based on the tables it touches [11:16:11] halfak: and then have human done tags too [11:16:42] I'm a little more skeptical of the utility of human tags. [11:16:56] Seems like we'd need to do a lot of work to build management into that. [11:17:20] right [11:17:23] E.g. we'll need oversight so that I can't tag something with "Yuvi_lives_at_" [11:17:26] table tags sound useful to start with [11:17:51] I suppose we can put that in queries now :S [11:17:58] indeed [11:18:05] I should have a 'report' button that goes to OTRS [11:18:30] And people on the other side that can have delete/hide/etc power. [11:18:36] That's a mop I'd pick up. [11:18:50] yeah [11:18:57] nothing abusive yet, of course :) [11:19:08] halfak: adding stars now. [11:19:23] halfak: I've also linked to the user's meta page and their talk page in profile [11:27:03] YuviPanda, good call [11:27:41] halfak: I should get flow on the Quarry page [11:28:42] Now this would be interesting. I want flow on my talk page. [11:28:58] I've got a request in to make flow available in Research: on meta [11:30:21] halfak: yeah, you can poke quiddity until it happens ;) [11:32:47] * halfak pokes quiddity :) [11:32:56] Mwahaha pings for when you wake up. [11:33:03] heheh [11:34:09] halfak: what can I do to make this more useful to the Oxford folks, btw? [11:34:51] Hmm... getting something related to views in it. [11:34:54] A public hadoop. [11:35:05] All of the data requests I got were about geo-location. [11:35:16] I've been writing emails at Ironholds trying to think of ways we can deliver on them. [11:35:41] yerp :/ [11:36:23] if you guys can produce something that's not too big and can be public, I can deliver it into labsdb [11:37:47] I'm actually providing Han-Teng with a dataset because I proposed the project and so get to control for the variables [11:37:57] although sanitising geoIP data in R is a Hard Problem [11:38:06] R :P [11:38:17] because it requires iterating on non-primitives if you want to store it in a reasonable way [11:38:19] oh shutup. [11:38:36] I'd document all the ways in which I dislike python except, oh wait, that would be the first time anything about python was ever documented. [11:38:38] :P [11:39:12] 'oooh, I have blinders on, can not see the light, light is undocumented anyway' :P [11:39:45] the trick is to import light. Then you deal with the fact that half of light is methods in a new class that requires object.method and half is functions [11:39:59] and that python makes no real distinction between them, and won't let you read the body of the function. [11:40:35] and then you try and find the documentation only to give up in despair 4 hours in, and resort to treating them as functions or methods until the nonsensical errors go away. [11:40:42] *throws hands up* PYTHON [11:40:44] n00b :P [11:40:47] (I like Python. But you insulted R.) [11:41:03] (only I get to insult R. I write it.) [11:41:08] heh [11:41:36] R passes large things around by copying them. boo :P [11:41:45] * Ironholds blinks [11:41:59] so does C. Do you not know what an array is? :P [11:42:39] Ironholds: err, passing an array passes a pointer to the array, not the array itself [11:42:45] eh, fair point [11:42:52] arrays are pointers, pointers are arrays [11:43:12] I could get into a linked-list-versus-array discussion here that has interesting implications for R versus Python best practises but eh, effort. [11:43:14] it passes structs by copying them, which is why you have pointers to structs everywhere [11:44:05] Ironholds, you can view python code of a function. [11:44:14] halfak, ooh, how? [11:44:24] http://pastebin.com/UUBE2VRp [11:44:39] aha. Cool! Thanks :) [11:44:45] although it being in a module is a BIT weird. [11:44:51] Note that I cut it off at 200 chars because it is a big function [11:45:02] I guess python code gets compiled by default(?) and so direct body access is not as simple [11:45:18] whereas R is uncompiled unless you explicitly byte-compile it. [11:45:18] not really, no. it gets bytecode compiled, but nobody distributes only those [11:45:36] .pyc files are an optimization that is pretty much fully transparent [11:45:42] sure, but when you import foo, foo.* is compiled when it's in your environment, yes? [11:46:03] when you library(foo), foo.* comes in as R code. Other than stripping out newlines and comments the body is fully accessible. [11:46:14] not really, no. even .pyc files have the same source tree, it's just in binary form, not text form [11:46:26] huh. Okay. [11:46:30] only with things like PyPI or Jython do things get 'compiled' [11:46:43] .pyc is more similar to PHP Opcode cache than anything else. [11:46:51] it just prevents a text-based parse, nothing more [11:47:07] okay. So change "compiled" for "binary" above. [11:47:08] Done ;p [11:47:22] (also: comparing [thing I don't know] to [thing I don't know]. Heh ;p) [11:47:25] :P [11:47:36] don't those make for the most productive conversations? :) [11:49:16] this is the last wikimania I spend giving you free therapy, dude ;p [11:50:03] We had a therapist available for delegates... [11:50:21] EdSaperia: was that Ironholds? :) [11:50:23] Ironholds: :P [11:50:39] EdSaperia, yes, I know. [11:50:40] | (•□•) | (❍ᴥ❍ʋ) [11:50:45] I broke them the first day of the hackathon. [11:51:03] (I kid. I didn't know we had a therapist. Weirdly specific thing to get.) [11:53:23] running off to lunch [11:53:25] o/ [11:59:38] * Ironholds tries to think of an elegant way to do this. Eurgh. [12:19:31] phuedx: halfak staaaarssss! :) http://quarry.wmflabs.org/Yuvipanda [12:21:04] "self starred queries"? [12:21:18] weird syntax ;p [12:21:31] Why not just one "Starred queries"? [12:23:46] Ironholds: because J-Mo wanted those. all the veribage is weird. [12:24:46] heh [12:24:55] suggest just going for "bookmarked", not starred [12:25:08] because, you know. People who aren't githubbers know what bookmarked means. [12:25:10] Ironholds: idea being you can use it to show off your queries that you like [12:25:13] heh [12:25:16] true, true [12:25:18] * YuviPanda hugs github [12:57:39] Re. "self-starred queries" I think that we should have "unlisted" vs. "listed" queries. [12:58:03] I don't like that we have a hack on top of stars to express worthiness. [12:58:09] YuviPanda & Ironholds ^ [12:59:03] well, I think all queries should be public by default, and we should have another thing on top to say 'I think these are nice' [12:59:08] we don't have to use stars, yeah. [12:59:32] we can call them something else. 'Featured'? [12:59:37] but that seems a bit weird. [12:59:55] internally they can use the star table - if it is your own query, we just use another string [13:00:26] by public by default I mean they should be public, period. no private queries. [13:01:03] Well, unlisted queries would still be public, just not in "the list". [13:01:31] E.g. a github repo for a python package that isn't in PyPI [13:02:10] Could also do "published" / "unpublished" [13:02:21] Maybe "published" vs. "draft" [13:02:28] That would fit a more wikipedian mindset [13:02:36] halfak: published sounds nice, but also has the connotation that it isn't visible until then [13:02:49] halfak: but yeah, having default state be 'draft' and then adding a 'published' to it sounds doable [13:02:56] :) [13:04:52] halfak: I'll redo it later today [13:04:58] <3 [13:05:04] I got some +1s for hourly pageviews from some OII folks. [13:05:19] halfak: hourly pageviews exist in labs, just not in a db form [13:05:34] If we had a table that was <views> that would solve a lot of problems that people are hammering stats.grok.se for. [13:05:47] <YuviPanda> indeed, and it'll need a fuckin huuuugeee machine :) [13:05:52] <YuviPanda> and joins across that would also be hard [13:05:56] <halfak> Indeed. :/ [13:06:06] <halfak> No joins necessary. Could be hadoop. [13:06:10] <YuviPanda> right [13:06:22] <YuviPanda> we have the data, just needs a lot of work :( [13:06:24] <halfak> Filter & aggregation would be very useful;. [13:06:24] <YuviPanda> and machines [13:06:48] <halfak> OK. wishful thinking. Add to backlog & depriotize. [13:06:58] <YuviPanda> heh :) [13:07:42] <halfak> Do we have anything like a public hadoop? [13:08:02] <YuviPanda> halfak: we as in WMF? nope [13:08:10] <halfak> kk. Just checking [13:09:14] <Ironholds> hmn [13:09:21] * Ironholds cannot tell if plyr call broken or what [13:59:28] <phuedx> *sigh* [13:59:33] <phuedx> YuviPanda|brb: … [13:59:38] <phuedx> stop moving so dern quickly [13:59:42] * phuedx can't keep up [14:30:29] <Ironholds> hey halfak, I just discovered a really interesting hydrodynamics phenomenon [14:30:43] <Ironholds> if you pour a cup of tea down a slope it forms a chai distribution [14:30:46] * Ironholds lets himself out [14:33:44] <tnegrin> argh [14:37:07] <Ironholds> tnegrin, the joke, or? [14:38:03] <tnegrin> the joke [14:38:59] * Ironholds nods. VICTORY. [14:47:34] <halfak> Nice work Ironholds. That was tortuous. [14:47:54] <halfak> Hmm... wrong word there. [14:47:58] <halfak> It tortured me. [14:48:25] <halfak> torturous [14:49:15] <Ironholds> I would also accept "tortuous" if you mean "so bad I'm going to sue you" [14:49:29] <Ironholds> or "torteius" for "so bad I've got no option but to make you into a pie as vengeance" [14:49:37] <Ironholds> *torteuous [14:50:59] <halfak> Now it [14:51:04] <halfak> 's lactuous [14:51:09] <halfak> 'cause you're milkin' it [14:54:24] * Ironholds groans [14:54:26] <Ironholds> nicely DONE! [14:54:38] <Ironholds> although I'd argue it's mortuous. Because I'm killing it. [15:00:38] <halfak> We're gonna get fired. [15:02:37] <halfak> Reason for termination: Torturing the English language without remorse. [15:03:32] <Ironholds> Compensation: we have made them learn French so they can ruin other people as well. [15:03:40] <Ironholds> Legal recourse: MAIS NON. [15:04:47] <tnegrin> DarTar: we're running the same giant query! [15:05:04] <DarTar> oops – are we? [15:05:06] <Ironholds> see, at least I'm running an INTERESTING query [15:05:19] <tnegrin> just counting hits right? [15:05:27] <Ironholds> had conversation with fundraising that went "can I use your data in my circadian patterns project" "eh, that's a pain" "alright! *gets it my own damn self from Hadoop*" [15:05:33] <tnegrin> I'll kill mine -- it's just for experimentation [15:05:49] <DarTar> tnegrin: on the phone now, bbl [15:05:52] <tnegrin> yours might run a little faster [15:05:53] <tnegrin> np [15:11:03] <quiddity> halfak, re: "make flow available in Research: on meta" - do you mean all 390 existing talkpages in the research: namespace, or just a few particular ones? [15:12:45] <quiddity> EdSaperia, I recommend "The Art of Looking Sideways". tis a large beast. excellent coffee table book of epiphanies. I've gifted 5 copies so far. [15:13:04] <EdSaperia> yessir [15:13:18] <quiddity> Ironholds, yes/no for colorbrewer2.org ? [15:13:31] <Ironholds> quiddity, I like it, leila found it for me [15:13:38] <Ironholds> It's RColorBrewer-compatible, so win. [15:15:21] * DarTar is a huge fan of colorbrewer [15:15:40] <DarTar> hey EdSaperia, nice to have you in the channel [15:16:21] <EdSaperia> In the basket [15:16:32] <EdSaperia> Hello DarTar! Did I meet you at Wikimania? [15:16:43] <DarTar> you did :) [15:17:17] <Ironholds> he was the most italian-looking italian [15:17:23] <Ironholds> if it helps distinguish ;p [15:17:27] <DarTar> @readermeter (and @wikiresearch) in twitterland [15:17:36] <DarTar> Ironholds: definitely not [15:17:45] <Ironholds> I dunno, you look pretty italian. [15:17:46] <DarTar> the prize goes to aubrey [15:17:52] <Ironholds> I don't know if I met aubrey. [15:17:58] <DarTar> wat [15:18:00] <DarTar> seriously? [15:18:11] <Ironholds> dude, I didn't see a single presentation the entire trip except the one I presented [15:18:16] <Ironholds> I spent 3 days cussing out polygons. [15:18:32] <Ironholds> basically my memory of the conference is a blur of code and, well, cussing. [15:18:54] <DarTar> -> https://wikimediafoundation.org/wiki/Thank_You/Andrea_Zanni [15:20:01] <Ironholds> ah! Possibly. [15:20:05] <Ironholds> He does look very Italian, it's true. [15:20:18] <DarTar> alright, shower and off to wiki [15:20:23] <DarTar> see you in a bit [15:22:38] <EdSaperia> :) [15:22:48] <EdSaperia> People are trying to get me to organise another conference [15:22:59] <Ironholds> a fun one? [15:23:12] <EdSaperia> I didn't run Wikimania because I wanted to be a conference organiser, though. [15:23:25] <EdSaperia> But if I had to run another conference, a commercial one, I'd run a VR conference. [15:23:41] <Ironholds> but would you really run it or would it ALL BE IN YOUR HEAD [15:23:44] * Ironholds looks mysterious [15:24:23] <halfak> EdSaperia, tried out occulus? [15:24:29] <EdSaperia> yyyyyyup [15:24:53] <EdSaperia> I can't wait till there's a load of competitors in this space, and HD units are consumer priced [15:25:13] <EdSaperia> I think I'll never take it off. I'll just leave this plane of existence entirely. [15:25:27] <EdSaperia> Only a matter of time [15:25:46] <halfak> +1. Gonna be fun. Think we'll look at the years of "monitors" as lamezors [15:25:47] <halfak> ? [15:28:03] <EdSaperia> Yes - I'm not interested in virtual environments at all, more unlimited screen real estate. [15:28:22] <EdSaperia> And when VR headsets have phones built in, which is basically obvious [15:28:29] <EdSaperia> And cameras on the front [15:29:16] <halfak> Heh. Good for augmented reality [15:30:12] <halfak> We'd need better input devices if we're going to be walking around. [15:30:53] <quiddity> what's that smartphone game people were playing at the conference, with the map-territory basis, using realworld landmarks, battling for vectors? [15:31:40] <quiddity> Ironholds, ^ someone was showing me when we had dinner that one ngiht. [15:31:40] <halfak> map-territory basis, using realworld landmarks, battling for vectors? [15:31:42] <halfak> woops [15:31:43] <halfak> lol [15:31:52] <halfak> Meant to paste this: http://imgur.com/xZDWSzZ [15:32:02] <Ironholds> quiddity, Ingress [15:32:26] <quiddity> Yes! Ingress will be fun with googleglass and thereafter. [15:33:18] <quiddity> oh, it's even made by Google. That's kinda creepy. [15:33:51] <halfak> Google bought 'em up. [15:35:44] <quiddity> https://en.wikipedia.org/wiki/Niantic_Labs looks like they were google from the start [15:37:20] <quiddity> (also, https://en.wikipedia.org/wiki/File:NES-Power-Glove.jpg ) [15:40:04] <tnegrin_> Ironholds: have you checked out the hadoop management tools? [15:40:28] <tnegrin> they will make your life better [15:40:34] <Ironholds> tnegrin_, yep, albeit ages ago on account of they were inaccessible for a couple months [15:41:41] <tnegrin> ok -- I'm watch Dario's megaquery and it's very useful to see the jobs running [15:42:11] <DarTar> what’s the hive equivalent of SHOW PROCESSLIST? [15:42:36] <DarTar> Ironholds: ditto [15:42:47] <tnegrin> it doesn't work like that -- you need to use the hadoop tools [15:42:57] <DarTar> boo [15:43:09] <tnegrin> it's a new world of distributed systems [15:43:16] * DarTar wears sunglasses [15:43:19] <tnegrin> I think there is some command line stuff but the web tools are easier [15:43:59] <tnegrin> ssh -N bast1001.wikimedia.org -L 8088:analytics1010.eqiad.wmnet:8088 [15:44:04] <tnegrin> http://localhost:8088/proxy/application_1406229821917_23653/mapreduce/job/job_1406229821917_23653 [15:44:11] <tnegrin> then you can watch your MR job run [15:44:59] <DarTar> nice, thx [15:46:16] <DarTar> best reviewer comment I’ve seen in a while [15:46:39] <DarTar> “this work is about the English Wikipedia only, so it’s product research not science” [15:46:57] * DarTar goes off to bulk delete 13 years or so of enwiki-only WP research [15:47:12] <EdSaperia> I'm not really a fan of AR. [15:47:19] <halfak> heh. It's a good thing that we don't accept studies of one organism into bio journals. [15:47:33] * halfak goes off to delete all genetic research done on fruit flies [15:47:59] <DarTar> :) [16:29:24] <Ironholds> so do we have our group meeting today or tomorrow? [16:29:26] <Ironholds> I'm still confused. [16:29:35] <leila> today [16:29:41] <leila> tomorrow is standup [16:29:42] <Ironholds> yay! Human contact! [16:29:45] <leila> :D [16:29:49] <leila> I'll move to the room [16:46:45] <YuviPanda|brb> halfak: quiddity yeah, Ingress. is pretty cool. They even released an iOS version [16:46:59] <YuviPanda|brb> J-Mo: quarry.wmflabs.org/Yuvipanda :) [16:47:47] <quiddity> YuviPanda|brb, I'd like to find, article titles that end with: ]}>:.,'"‒–—―…!‐-?;/\&*^%$#@™ [16:48:15] <quiddity> (no rush. I just wanted to get it out of my notepad :) [16:48:30] <YuviPanda|brb> heh, should be easy, I think [16:50:11] <YuviPanda|brb> halfak: Ironholds hmm, any idea why http://quarry.wmflabs.org/query/223 is an SQL error? [16:50:30] <YuviPanda|brb> ah, nevermind [16:54:20] <YuviPanda|brb> quiddity: hmm, either there are no such pages or my query is wrong http://quarry.wmflabs.org/query/223 [16:55:14] <Ironholds> YuviPanda|brb, those characters shouldn't be in the title [16:55:21] <Ironholds> actually, are titles URL encoded in the database? [16:55:26] <Ironholds> that'd do it too [16:55:26] <YuviPanda|brb> they aren'\t [16:55:27] <YuviPanda|brb> aren't [16:55:34] <Ironholds> hmn [16:55:51] <YuviPanda|brb> quiddity: maybe no such pages exist? do you know of an example? [16:55:59] <Ironholds> YuviPanda|brb, totally [16:56:03] <Ironholds> https://en.wikipedia.org/wiki/Talk:Star_Trek:_The_Next_Generation [16:56:11] <YuviPanda|brb> Ironholds: that doesn't end with a : [16:56:15] <YuviPanda|brb> I'm just looking for endswith [16:56:18] <YuviPanda|brb> there's a '$' at the end [16:56:25] <Ironholds> ohhh [16:56:28] * Ironholds thinks [16:56:45] <Ironholds> there shouldn't be [16:59:23] <quiddity> try "+" to get C++ [17:05:25] <YuviPanda|brb> quiddity: right, so that works [17:09:20] <YuviPanda|brb> quiddity: I might have it :) is running now [17:09:40] <YuviPanda|brb> quiddity: is probably a bit more inclusive than you'd like, but it's everything ending with something other than a-z, A-Z, 0-9, ) and + [17:10:14] <quiddity> That's fine! I would have suggested, but thought it might be too intensive/long-running. [17:10:26] <YuviPanda|brb> let's see if this runs in 10m! [17:11:40] <YuviPanda|brb> quiddity: ah, it's quite fast. lots of accents now :| [17:14:24] <quiddity> hehe [17:19:44] <YuviPanda|brb> quiddity: w00t. http://quarry.wmflabs.org/query/223 for a subset of characters. might freeze your browser for a bit [17:20:01] <YuviPanda|brb> I can limit it to mainspace as well [17:22:40] <YuviPanda|brb> quiddity: it's probably slightly more useful now. LOTS of quotes [17:25:51] <quiddity> it's killing my browser, that's for sure! [17:26:15] <YuviPanda> quiddity: heh, I should implement CSV download s oon [17:29:01] <Ironholds> TSV [17:29:03] <Ironholds> TSV! [17:30:19] <YuviPanda> yes yes TSV too [17:30:49] <Ironholds> no, instead ;p [17:31:20] <YuviPanda> :) [17:31:26] <YuviPanda> for any of that I need to rewrite the results backend [17:31:30] <YuviPanda> have been putting that off a bit [17:31:56] <Ironholds> too bad [17:32:17] <YuviPanda> quiddity: did you finally see results or did it just kill your browser [17:33:14] <quiddity> I had to kill my browser [17:34:06] <YuviPanda> quiddity: :( let me limit it to 1000 [17:34:29] <YuviPanda> quiddity: try now? [17:35:35] <quiddity> I'm in standup meeting. will try after ;) [17:36:13] <YuviPanda> quiddity: cool [17:53:08] <Ironholds> holy hellballs. [17:53:19] <Ironholds> halfak, so in response to a really thorny sanitising problem I finally checked out data.tables. [17:53:27] <Ironholds> I would like an explanation, please, as to why this code is not in core. [17:53:37] <Ironholds> it's like if data frames were primitives! [17:53:53] <Ironholds> or more accurately: it's like if the data type people used was actually the data type the language was oriented around. [17:54:14] <halfak> :) [17:54:17] <halfak> <3 data.tables [17:59:02] <EdSaperia> I just went and ordered an Oculus Rift V2 devkit XD [18:00:07] <quiddity> wooo! [18:04:35] <Ironholds> halfak, seriously, this is mad. [18:05:21] <halfak> EdSaperia, that's pretty reasonably priced. [18:05:23] * halfak considers [18:05:35] <halfak> Ironholds, data.tables. Remember what I said? [18:05:56] <Ironholds> yep. [18:06:00] <EdSaperia> Right? And 1080p for each eye! [18:06:03] <Ironholds> you were right, I should've listened. [18:06:39] <halfak> Took me a couple months too. [18:17:39] <YuviPanda> EdSaperia: halfak do you know each display on the DK2 is just a Galaxy Note2? :) [18:17:56] <YuviPanda> http://9to5google.com/2014/07/31/oculus-rift-dev-kit-2-teardown-reveals-samsung-galaxy-note-3-front-panel/ [18:17:58] <EdSaperia> Then it's even more reasonably priced! [18:18:07] <YuviPanda> how much was it, EdSaperia? [18:18:09] <YuviPanda> and what's the lead time? [18:18:11] * YuviPanda considers as well [18:18:20] <halfak> lol. two Galaxy note 2's? [18:18:21] <EdSaperia> $350, and ?t? [18:18:45] <YuviPanda> bah, it won't ship to India and I don't think it'll ship by the time I leave the UK :( [18:18:55] <YuviPanda> I'd totally buy it if I could [18:21:34] <EdSaperia> You still around? In London? [18:21:58] <YuviPanda> EdSaperia: I'm in Glasgow, in London in September for a m onth [18:22:13] <EdSaperia> Awesome, look me up, let's hang out! [18:22:26] <EdSaperia> | (•□•) | (❍ᴥ❍ʋ) [18:22:59] <halfak> Oh i see. just one [18:22:59] <halfak> YuviPanda, I could ship one to you. [18:22:59] <halfak> Then again, I won't be back for a bit. Better to have someone in SF ship it. [18:23:02] <YuviPanda> EdSaperia: totally :D [18:23:39] <EdSaperia> I wonder if I could be the first person to wear an oculus rift on top of a google glass [18:24:19] <YuviPanda> EdSaperia: sorry, already done that :P [18:24:32] <EdSaperia> ಠ_ಠ [18:24:34] * YuviPanda has a Glass, and had a friend's DK1 a few months ago [18:24:54] <YuviPanda> it was absolutely useless, of course [18:24:55] <YuviPanda> but still [18:26:46] <YuviPanda> EdSaperia: am ordering one now :) [18:26:56] <EdSaperia> woohoo [18:27:47] <YuviPanda> EdSaperia: very impulse-y :) Reconsidering since consumer version might be out by May 2015 [18:28:17] <EdSaperia> Gotta get ahead of the curve :) [18:28:36] <YuviPanda> hehe true :D [18:28:48] <YuviPanda> only problem is if I'll have time to play with it. I didn't for the Google Glass :( [18:29:34] <EdSaperia> Just start using it as a monitor [18:29:39] <YuviPanda> hahah :D [18:31:21] <YuviPanda> EdSaperia: I've given myself 3h, if I still want it 3h from now I'll get it [18:31:38] <EdSaperia> cool system! [18:33:53] <YuviPanda> :) [19:01:55] <YuviPanda> DarTar: not sure if you saw, but we have stars :) http://quarry.wmflabs.org/Yuvipanda [19:02:42] <leila> DarTar, whenever you have have few minutes, can you tell me about onboarding docs? maybe in person. :-) [20:07:11] <leila> Ironholds: yt? [20:23:41] <YuviPanda> J-Mo: new 'published' state instead of self starring, clearer I think http://quarry.wmflabs.org/Yuvipanda [20:23:48] <YuviPanda> thanks to Ironholds and halfak for solutions! \o/ [20:24:31] <J-Mo> that seems okay to me [20:24:51] <YuviPanda> J-Mo: yeah, and in the UI as well it's a distinct state from starred [20:25:01] <J-Mo> cool! [20:25:23] <YuviPanda> J-Mo: right, so I just need to get description for you [20:26:05] <J-Mo> sure, that'll happen :) what do you need most, and when do you need it? [20:27:11] <YuviPanda> J-Mo: most? would be UX suggestions :) I need to make it clearer that you should enter a title. [20:28:01] <J-Mo> cool. I'll work on that. [20:28:35] <YuviPanda> hey HenriqueCrang [20:28:53] <YuviPanda> HenriqueCrang: http://quarry.wmflabs.org/Yuvipanda has both 'stars' (for other people's queries) and 'publish' for your own queries [20:28:57] <HenriqueCrang> hi YuviPanda ! [20:29:40] <YuviPanda> HenriqueCrang: and the query runs screen has also been re-organized :) Should prevent issues from people not running queries because they think it'll fail / be killed [20:31:28] <HenriqueCrang> it's awesome! [20:31:36] <HenriqueCrang> just trying now! [20:31:40] <YuviPanda> HenriqueCrang: \o/ [20:32:02] <YuviPanda> HenriqueCrang: I've to head off now, but you can leave comments to me on IRC (I'll read them later) or file bugs (link at the bottom of quarry) [20:32:17] <YuviPanda> HenriqueCrang: would love to support you in any way possible for any possible use cases in Brazil :) [20:32:22] <HenriqueCrang> great! [20:32:32] <YuviPanda> HenriqueCrang: \o/ ciao for now :) file bugs / leave comments! [20:32:45] <HenriqueCrang> good to know, stared dude :) [21:03:38] <halfak> Hey phuedx. I'm in the githubs mergin your pull requests. :) [22:00:47] <Ironholds> halfak, I just reduced like, a slow-as-hell 8-line ddply call to aggs <- data[,length(unique(user)),by=c("page_title","project")] [22:00:48] <Ironholds> this is mad. [22:00:52] <Ironholds> so very mad. [22:01:01] <Ironholds> It's like...the clarity of Python with the complex data structures of R. [22:01:13] <Ironholds> I think I might be in love. If I have a secretary please get them to hold my calls for the next week. [22:01:45] <halfak> Oooh and it gets slightly better. [22:01:58] <halfak> Change that last bit to by=list(page_title, project)] [22:02:12] <halfak> This way you can rename columns and perform computations while grouping. [22:02:42] <halfak> e.g. ... by=list(log_bucket=exp(floor(log(value))))] [22:03:09] <halfak> <3 data.tables. [22:03:14] <halfak> So many nice design decisions. [22:03:29] <halfak> Gotta head to bed. Have a good one! [22:07:01] <Ironholds> ooh, clever [22:07:03] <Ironholds> take care! [23:17:45] <Ironholds> hey DarTar [23:17:51] <Ironholds> so, I assume our 1:1 earlier was not on? ;p [23:17:53] <DarTar> hey [23:18:26] <DarTar> dude, I’m sorry – I was in a goal review meeting with toby and kevin that lasted 2 hours [23:19:09] <DarTar> do you mind if we move our 1:1 permanently to Friday morning? [23:19:21] <DarTar> it looks like we’ve done this for the last couple of weeks [23:20:56] <DarTar> Ironholds: would that work? Like tomorrow 11 PT [23:21:05] <Ironholds> totally [23:26:17] <leila> DarTar, re the permanently part, there is a meeting at that time next week [23:26:21] <leila> which we should change? [23:26:27] <DarTar> ah [23:26:30] <DarTar> hang on [23:27:01] <leila> I know you have something else and have to cut that one short, but in case I can't move it. [23:27:04] <DarTar> hmm ok, Ironholds: early afternoon ok? [23:27:07] <Ironholds> ehh [23:27:12] <Ironholds> morning is better. UTC ;p [23:27:14] <Ironholds> and then EST ;p [23:27:22] <DarTar> oh right [23:27:47] <leila> Friday night doesn't start /that/ early Ironholds. [23:27:48] <DarTar> so let’s do tomorrow 11PT for now [23:27:51] <leila> ;p [23:28:02] <DarTar> ? [23:28:10] <DarTar> and we’ll figure out the next ones [23:28:19] <DarTar> when are you going to be in Boston? [23:28:30] <leila> 18th [23:28:41] <DarTar> leila: you scare me [23:28:44] <leila> :D [23:28:47] <leila> He put it in the calendar [23:28:52] <DarTar> ok [23:29:39] <leila> Ironholds, just to double-check: you will have internet access for the showcase, right? [23:29:47] <DarTar> Ironholds: invite sent for tomorrow, sorry again for not letting you know earlier [23:30:51] <Ironholds> leila, I will, yes [23:30:58] <Ironholds> if I have to claw it from molly's housemate's cold dead hands [23:31:00] <Ironholds> DarTar, np! [23:31:00] <leila> great! [23:31:15] <Ironholds> *housemates' [23:46:20] <J-Mo|away> leila, bad news: I have a meeting that wants to be scheduled at the time you identified for our reading group next Friday. Looks like I won't be able to make it this time :( [23:46:50] <leila> I'll reschedule then. the only reason I didn't was that you said you can't read it by then [23:47:02] <leila> DarTar has a minor conflict with it, too, [23:47:08] <leila> this will make it earlier, but what to do [23:47:14] <leila> thanks for letting me know. [23:47:57] <J-Mo|away> yeah, thanks for trying to accommodate me. I appreciate that. I won't be able to attend this time, but I would like to attend a future session! [23:48:09] <leila> yup. no worries.