[15:20:38] <halfak>	 o/
[15:21:36] <schana>	 morning halfak
[15:21:47] <halfak>	 Hey schana!  
[15:27:47] <halfak>	 o/ sabya
[15:28:00] <sabya>	 o/ halfak
[15:29:48] <guillom>	 \o_
[15:29:58] <halfak>	 :D
[15:31:33] <sabya>	 how was Dosa with Subbu ? :-D halfak 
[15:32:01] <halfak>	 Was great!  :)  My first time having dosa. 
[15:32:21] <halfak>	 sabya, do you know subbu?  
[15:33:10] <sabya>	 nope. I read you were writing to DarTar.
[15:33:29] <halfak>	 Oh!  Gotcha :)  
[15:34:39] * subbu looks up from his non-dosa breakfast and goes back to it ...
[15:35:04] <subbu>	 halfak, dosa is commonly eaten for breakfast in south india
[15:36:08] <sabya>	 halfak : I am getting familiar with Wiki editing tools since last couple of days. It will help me put things in perspective. Read the paper you sent me
[15:36:34] <halfak>	 Great!  That's a good paper.  It's one of my favorites. 
[15:37:15] <halfak>	 I just happened to be inspired to pick up XGBoost this morning. 
[15:37:31] <halfak>	 http://xgboost.readthedocs.org/en/latest/model.html
[15:38:07] <halfak>	 I'm experimenting using it to predict damaging edits to see how it compares to our RandomForest and GradientBoosting models. 
[15:38:21] <sabya>	 the paper is pretty exhaustive. I wanted to try Huggle on enwiki as well. but not having rollback permission. So tried it on some other wiki.
[15:43:21] <halfak>	 sabya, yeah.  Fun story -- I've never used Huggle on enwiki either :) 
[15:43:39] * halfak should get his rollbacker rights 
[15:43:53] <sabya>	 halfak, how long are you online here? I'll be back in 45 mins.
[15:43:58] <halfak>	 sabya, any luck experimenting with the labs cluster?
[15:44:01] <halfak>	 I'll be here :) 
[15:44:23] <sabya>	 great. I'll try to setup the cluster today.
[15:44:29] <halfak>	 sabya started collaborating with me when I was traveling.  Little does he know that I rarely leave this channel otherwise ;) 
[15:46:22] * halfak maxes his CPU on some experiments with XGBoosting models. 
[15:46:38] <halfak>	 Ooooh.  And it looks like the preliminary tests are pretty good. 
[15:47:19] <halfak>	 Even simple trees are getting nice scores on enwiki's "damaging" predictions. 
[15:50:11] <guillom>	 BTW Tilman pointed me to http://tuprints.ulb.tu-darmstadt.de/5225/ earlier today; it might be interesting to others here.
[15:50:38] <guillom>	 "The Writing Process in Online Mass Collaboration: NLP-Supported Approaches to Analyzing Collaborative Revision and User Interaction"
[15:51:10] <guillom>	 "econd, we investigate activity-based roles of users in Wikipedia and how they relate to the collaborative writing process. We automatically classify all revisions in a representative sample of Wikipedia articles and cluster users in this sample into seven intuitive roles."
[15:51:19] <guillom>	 halfak: ^
[15:54:20] <halfak>	 guillom, yeah.  Been working from Daxenberger's notes.  Some big methodological holes there though. 
[15:54:46] <halfak>	 E.g. training and testing the edit type classifier on edits to a biased sample of articles. 
[15:55:00] <guillom>	 I haven't looked at it yet; Tilman sent it to me because he knew I was working on contributor roles.
[15:55:13] <halfak>	 But I think the overall methodological vision is solid and will work great with iteration. 
[15:56:27] * guillom looks at his quarterly goals for the next quarter, tries not to fall victim to the [[w:planning fallacy]] this time.
[16:10:35] * halfak quotes from [[:en:planning fallacy]]
[16:10:37] <halfak>	 "The planning fallacy, first proposed by Daniel Kahneman and Amos Tversky in 1979, is a phenomenon in which predictions about how much time will be needed to complete a future task display an optimism bias (underestimate the time needed)"
[16:11:54] <apergos>	 always
[16:12:25] <halfak>	 o/ apergos 
[16:12:32] <apergos>	 heya
[16:12:42] <halfak>	 Good morning!  (or other timezone apt. greeting)
[16:12:43] <apergos>	 so I have a law for (a part of) that
[16:12:55] <apergos>	 "Nothing ever takes 5 minutes."
[16:13:06] <halfak>	 Except for editing a Wikipedia article. 
[16:13:12] <halfak>	 And it's often 7 minutes. 
[16:13:17] <halfak>	 But 5 is pretty darn common ;) 
[16:13:34] <apergos>	 well:  "I just want to push this patch out. It will only take a minute"
[16:13:38] <apergos>	 = famouslast words
[16:13:45] <halfak>	 You must already be in an "activity phase" though
[16:14:04] <schana>	 "If you wait till the last minute, it only takes a minute."
[16:14:18] <halfak>	 Nonesense I'm basing my assertions on: http://arxiv.org/abs/1411.2878
[16:14:22] <apergos>	 "if you wait til the last minute, you'll be doing it again tomorrow"
[16:15:24] <apergos>	 you can paraphrase this as" fast or done, choose one" :-P
[16:17:01] <apergos>	 (I"m looking at the paper now)
[16:21:00] <halfak>	 apergos, that's one work that I'm hoping to re-write for a few different audiences.  It seems like we stumbled onto a strange regularity in human behavior.  I've been talking to neuroscientists about physiological cycles.  I expect that there some neuromechanical explanation for these attention patterns. 
[16:21:36] <halfak>	 E.g. the symbols that form an action get ~7 minutes of memory to work with and we'll optimize our /activities/ to be constructed of 1-7 minute actions. 
[16:22:00] <halfak>	 Regretfully, most nueroscientists I have run into are looking at cycles on the nano-second scale. 
[16:22:34] <halfak>	 When I say I'm looking at patterns that are *minutes* long, I usually get laughed at :D
[16:22:44] <wiggins>	 hey halfak, I'll be forcing students to use Quarry this afternoon, approx. 4 Eastern; will also show IRC channel but might be the only one actually online to mediate questions
[16:23:23] <halfak>	 o/  wiggins 
[16:23:30] <halfak>	 Cool.  I'll be around. 
[16:23:34] <halfak>	 yuvipanda, ^ 
[16:24:10] <wiggins>	 awesome, thanks--hopefully it'll go down easy since at least 3 students know SQL, but schema opacity could trip us all up
[16:24:59] <halfak>	 wiggins, I think that is likely.  In a way, discovering the MediaWiki database (and probably any sufficiently complicated other DB) is a lot like discovering the norms of some community. 
[16:25:20] <apergos>	 so what about people who are spending time drafting the article, reading sources...?  that never gets counted but it is a part of "time editors spend contributing" i.e. "wiki-work"
[16:25:26] <halfak>	 Some of it is documented, but you don't really know how to operate within them until you either (1) observe someone else or (2) practice them yourself. 
[16:25:49] <wiggins>	 well, that'll make my key point for the day: metadata matters. A lot.
[16:25:53] <halfak>	 apergos, that's right.  I guess I suspect it to be somewhat uncommon, but it is excluded from any labor hours measurement. 
[16:26:31] <halfak>	 wiggins, would love to see some notes on what metadata would have been most useful and how your students would *like* to be able to browse it. 
[16:26:41] <apergos>	 I wonder what a survey would reveal about that
[16:26:42] <halfak>	 E.g. search queries and click on table names to get schema
[16:28:10] <wiggins>	 halfak, for sure, if I can record/summarize, will do. Tasks are: 1) find "interesting" recent query & attempt to interpret; 2) find series of queries by same user & attempt to guess what they're up to; 3) pick a query, try to modify & run it
[16:29:25] <halfak>	 wiggins, sounds like that will be fun and super informative :)  
[16:29:28] <wiggins>	 added to activity slide: feedback for Wikimedia, what would have made this easier?
[16:29:34] * halfak wishes he had prof. wiggins as instructor
[16:29:42] <halfak>	 Awesome!
[16:31:06] <wiggins>	 it might be part of their portfolio assignment for the day, then they can elaborate a little more
[16:32:01] <wiggins>	 really curious to see how far they can get with it, given the vastly divergent backgrounds in the class
[16:41:22] <sabya>	 gonna try setting up my lab cluster
[16:42:56] <halfak>	 sabya, great.  Note that there are puppet roles for "ores".  These set up the redis node, load balancer, web node, and worker nodes. 
[16:43:13] <halfak>	 These may come in handy if you want to replicate prod, but for testing you might not want to use them. 
[16:43:55] <sabya>	 ok, any getting started doc?
[16:44:18] <halfak>	 sabya, regretfully not. 
[16:44:30] <halfak>	 But this seems like a good time to make one. 
[16:44:38] <halfak>	 Maybe we can do that as I answer questions. 
[16:45:15] <sabya>	 sure. I'll create one.
[16:46:32] <halfak>	 sabya, maybe you could make it an etherpad or wiki page so that I can copy-paste bits in there for you. 
[16:47:35] <sabya>	 ok. let me try etherpad
[17:02:08] <yuvipanda>	 hey wiggins!
[17:02:10] <yuvipanda>	 nice to see you here :D
[17:06:17] <yuvipanda>	 halfak: I should add 'search' to Quarry
[17:06:38] <halfak>	 Yeah.  Would be really cool to filter queries by table used as well.  
[17:06:55] <yuvipanda>	 we can easily do that, yeah
[17:06:56] <halfak>	 yuvipanda, is it possible to get an AST for SQL queries easily?
[17:07:05] <yuvipanda>	 there are enough parsers for it yeah
[17:07:19] <halfak>	 Even a messy one would probably be OK. 
[17:07:35] * yuvipanda nods
[17:07:44] <yuvipanda>	 and it'll also give me experience with elasticsearch
[17:07:49] * halfak sets up an experimental ORES node for sabya to play with 
[17:10:34] <Amir1>	 halfak: do you have some time to check the PRs for i18n? 
[17:10:55] <halfak>	 Yeah.  Will look as soon as I have sabya started. 
[17:11:04] <halfak>	 Amir1, check out https://etherpad.wikimedia.org/p/setting_up_ORES_lab_cluster in the meantime. 
[17:11:35] <Amir1>	 yuvipanda: hey, please check this patch: https://gerrit.wikimedia.org/r/274912
[17:11:41] <yuvipanda>	 Amir1: halfak btw, ensure_packages vs require_packages is a red herring
[17:11:42] <Amir1>	 awesome
[17:11:46] <Amir1>	 thanks halfak 
[17:12:05] <halfak>	 yuvipanda, bummer.  Any thoughts on the underlying issue?
[17:12:09] <halfak>	 sabya, http://ores-experimental.wmflabs.org/
[17:12:16] <halfak>	 That's the public-facing address. 
[17:12:30] <yuvipanda>	 yes
[17:12:35] <sabya>	 ok. thanks halfak 
[17:12:40] <halfak>	 sabya, you should be able to SSH to ores-experimental-01.revscoring.eqiad.wmflabs
[17:12:47] <halfak>	 If you have your ssh set up right for labs. 
[17:13:06] <halfak>	 Turns out you can drop the "revscoring" and just SSH to ores-experimental-01.eqiad.wmflabs
[17:13:07] <sabya>	 had uploaded the keys. let me check.
[17:13:13] <Amir1>	 :(
[17:13:19] <halfak>	 sabya, will need to set up proxy through the bastion too. 
[17:13:20] <leila>	 hi schana. 
[17:13:22] * halfak gets link. 
[17:13:28] <schana>	 hey leila
[17:14:02] <halfak>	 sabya, see https://wikitech.wikimedia.org/wiki/SSH_access
[17:14:04] <Amir1>	 yuvipanda: okay. let me check the whole thing again
[17:14:09] <Amir1>	 probably tomorrow
[17:14:12] <leila>	 I'm reviewing the tasks, question: in T113384 can you add an example of a working url as a comment in the task?
[17:14:16] <leila>	 schana, ^
[17:14:21] <Amir1>	 but your insight would be valuable 
[17:14:24] <schana>	 sure, leila
[17:14:34] <leila>	 thanks, schana.
[17:14:43] <Amir1>	 o/ leila :)
[17:14:46] * halfak needs to run away for lunch
[17:14:48] <halfak>	 back in a bit
[17:14:50] <leila>	 hey Amir1. :-)
[17:15:39] <Amir1>	 sup leila?
[17:16:03] <leila>	 not much, Amir1. How is it going on your end?
[17:16:27] <Amir1>	 not bad, waiting for ops to move ores to prod
[17:16:45] <Amir1>	 then we deploy the extension in fa.wp and then wikidata
[17:17:08] <leila>	 yeah, that's what I'm hearing. :-)
[17:18:15] <Amir1>	 I hope this happens soon
[17:18:36] <Amir1>	 I saw article recommander rocks 
[17:19:22] <leila>	 schana is working on it these days, Amir1. We're aiming for some campaigns and tests to start soonish.
[17:19:48] <Amir1>	 awesome
[17:19:50] <Amir1>	 :)
[17:20:22] <Amir1>	 If I can help, I would be thrilled 
[17:21:49] <leila>	 I'll let you know, Amir1. You already have a lot on your plate, but if you'd like to help, I'm sure we will bug you soon. D:
[17:21:50] <leila>	 :D
[17:22:18] <Amir1>	 Noruz is coming so I would lots of free time
[17:22:28] <Amir1>	 and I'm looking into possible way to fill it
[17:23:05] <Amir1>	 *I would've 
[17:32:57] <sabya>	 SSH worked!
[17:33:02] <sabya>	 sabya is thrilled
[17:39:33] <yuvipanda>	 Amir1|afk: halAFK https://gerrit.wikimedia.org/r/#/c/275560/ should probably fix one 
[18:07:38] <halfak>	 sabya, \o/
[18:07:43] <halfak>	 yuvipanda, looking
[18:08:25] <sabya>	 any other steps? halfak 
[18:09:19] <halfak>	 sabya, yeah.  See the fabfile in ores-wikimedia-config. 
[18:09:20] * halfak gets
[18:09:31] <halfak>	 https://github.com/wiki-ai/ores-wikimedia-config/blob/master/fabfile.py
[18:09:47] <halfak>	 fabric essentially lets us execute commands remotely on a set of machines.  
[18:09:57] <halfak>	 We use it to set up a new node and to deploy changes to nodes. 
[18:10:16] <halfak>	 See "initialize_staging_server": https://github.com/wiki-ai/ores-wikimedia-config/blob/master/fabfile.py#L65
[18:11:22] <halfak>	 I'm not sure if there is a good way to purposefully run the fabfile against this machine, but there should be. 
[18:11:36] <halfak>	 Alternatively, you could just review the fabfile and run the commands manually. 
[18:12:20] <sabya>	 ok.
[18:16:24] <yuvipanda>	 halfak: just a fyi, I'm going to merge that in about 15minutes and see how that goes
[18:16:45] <halfak>	 Sounds good.  How do you want to test it?
[18:23:10] <yuvipanda>	 halfak: I don't :D
[18:23:13] <yuvipanda>	 (j/k)
[18:23:24] <yuvipanda>	 I'm just going to merge it and make sure nothing breaks, and see how that goes
[18:23:25] * halfak rages for 0.01 seconds. 
[18:23:52] <halfak>	 We could stand up 3rd web node for a short period of time and see what happens.
[18:24:28] <yuvipanda>	 nah
[18:24:36] <yuvipanda>	 this affects all of uwsgi
[18:24:39] <yuvipanda>	 including prod services
[18:24:45] <yuvipanda>	 so we'll find out wherever
[18:24:49] <yuvipanda>	 so in cases like this
[18:24:51] <yuvipanda>	 I'll disable puppet
[18:24:54] <yuvipanda>	 in all these places
[18:24:56] <yuvipanda>	 and enable it in one
[18:24:56] <halfak>	 heh.  Let the others test it. 
[18:30:47] * sabya is sleepy. going to bed. will get back on setting up the labs instance tomorrow.
[18:31:06] <halfak>	 o/ sabya 
[18:31:12] <halfak>	 Thanks for hacking today :) 
[18:31:42] <sabya>	  thanks for helping me out. :-)
[18:31:52] <halfak>	 o/ schana 
[18:32:01] <schana>	 hey halfak
[18:32:15] <halfak>	 Interested in digging into ORES stuff today?
[18:33:11] <schana>	 what did you have in mind?
[18:34:52] <yuvipanda>	 halfak: can you make the requirements.txt in ores-wikimedia-config to be the output of a pip freeze?
[18:35:25] <halfak>	 yuvipanda, yes
[18:35:31] <yuvipanda>	 halfak: ok, thanks :D
[18:35:35] <halfak>	 schana, still working on that packaging stuff. 
[18:35:39] <yuvipanda>	 halfak: I'll create the repo and stuff
[18:35:42] <halfak>	 We only did the first bit before. 
[18:35:57] <halfak>	 schana, would love to have you get an overview of the system so that you can help get it into prod. 
[18:36:19] <schana>	 halfak: I could probably do an hour today, but could do a lot more tomorrow
[18:36:33] <halfak>	 kk.  So, there were the two tasks I directed you to last time. 
[18:36:43] <halfak>	 Seems like yuvipanda's ask is one we should add. 
[18:36:44] <schana>	 yeah, I'm looking at them now
[18:36:45] * halfak gets links. 
[18:39:39] <schana>	 halfak: I'm heading out to lunch, but should be back in about an hour to 1.5 hours from now
[18:39:49] <schana>	 if you wanted to work on it then
[18:40:22] <halfak>	 schana, OK if I start assigning a few things to you? 
[18:40:30] <halfak>	 We can always re-shuffle later
[18:40:54] <schana>	 halfak: sure
[19:11:18] <halfak>	 schana_lunch, when you get back, check out https://phabricator.wikimedia.org/T129109
[19:11:31] <halfak>	 I spent some time adding links and clarifying what should be done. 
[19:11:44] <halfak>	 Time for me to go to aspen. 
[19:21:20] <leila>	 schana_lunch: for whenever you're back: do you think we should have 2 shorter backlog grooming meetings every week, or 1 longer one every week. I think at the moment, a lot of changes are happening, so maybe 2 shorter ones every week are more useful. What do you think?
[19:25:05] <halfak|Aspen>	 leila, ^ is this the whole team backlog grooming?
[19:25:13] <halfak|Aspen>	 Or another, separate one?
[19:25:49] <halfak|Aspen>	 Woops.  I'm not supposed to be here
[19:25:52] * halfak|Aspen runs back to aspen
[20:09:25] <leila>	 halfak|Aspen: no. it's just the content creation work.
[20:09:29] <leila>	 hi halfak|Aspen.
[20:43:54] <halfak>	 o/ leila 
[20:43:57] <halfak>	 Just got back from aspen
[20:44:07] <leila>	 hey halfak.
[20:44:16] <halfak>	 "content curation work"?
[20:44:28] <halfak>	 *creation
[20:44:31] <apergos>	 aspen the channel or aspen the place IRL?
[20:44:34] <leila>	 schana: do you think we should have 2 shorter backlog grooming meetings every week, or 1 longer one every week. I think at the moment, a lot of changes are happening, so maybe 2 shorter ones every week are more useful. What do you think?
[20:44:38] <halfak>	 aspen the channel :) 
[20:44:41] <apergos>	 :-)
[20:44:44] <leila>	 halfak: article rec. :-)
[20:44:47] * schana thinks
[20:44:52] * halfak doesn't use his teleporter during work. 
[20:45:10] <halfak>	 leila, gotcha.  
[20:45:11] <leila>	 increasing content coverage to be exact, halfak. :D
[20:45:11] <apergos>	 can I borrow it then?
[20:45:34] <halfak>	 apergos, only one downside, you need to die and let your clone take over every time you use it. 
[20:45:48] <apergos>	 you got one of the cheap off-brand ones
[20:45:54] <halfak>	 :D
[20:47:31] * halfak didn't know it was cool to invite schana to backlog grooming events. 
[20:48:04] <halfak>	 Revscoring team does our backlog grooming on Saturday mornings to accommodate volunteers
[20:48:12] <schana>	 leila: I think one meeting / week should be enough to cover the changes happening. If we find ourselves needing more we can always increase the frequency *shrugs shoulders*
[20:48:15] <halfak>	 But we can move it
[20:48:55] <schana>	 halfak: I'm a software engineer; I love making decisions and arguing about them :D
[20:49:32] <leila>	 schana: sounds good. any preference for the day of the week?
[20:49:44] <leila>	 schana: I'm leaning towards early in the week so we can plan that week
[20:50:03] <schana>	 that would work, leila
[20:50:16] <leila>	 ooki, sending out the invites, schana.
[20:50:22] <leila>	 just one last thing: 30-min, schana?
[20:50:34] <schana>	 should be enough
[20:50:42] <leila>	 agreed.
[21:02:15] <halfak>	 Hey folks.  I just got an email asking about past work in Wikipedia. 
[21:02:26] <halfak>	 Anyone know if there has been a survey about mental health issues of Wikipedia editors?
[21:03:26] <halfak>	 HaeB, guillom, EGalvez ^ 
[21:13:34] <HaeB>	 don't remember anything offhand
[21:13:50] <wiggins>	 halfak, going to start off w/ your thanks/month on enwiki. any date range limits on that?
[21:14:01] <halfak>	 wiggins, wat?
[21:14:33] <wiggins>	 on quarry, one of your saved queries is thanks/month on enwiki. nice & simple; preview results only show through 2014
[21:14:47] <wiggins>	 but no metadata on query so can't tell if that's b/c you last ran it in 2014
[21:20:31] <halfak>	 Oh!  
[21:20:38] <halfak>	 Yeah.  Probably because I ran it in 2014. 
[21:20:45] <halfak>	 It would help a lot if you pasted a link ;) 
[21:20:47] <halfak>	 wiggins, ^
[21:20:59] <wiggins>	 right, that was my guess: http://quarry.wmflabs.org/query/216
[21:22:30] <halfak>	 Ahh yes.  Nice and simple.  Looks like I must have run that in Aug. 2014. 
[21:22:40] <halfak>	 I'd guess it to be about the middle of the month :) 
[21:22:42] <wiggins>	 Yes, I just ran a fork & it's current
[21:26:26] <wiggins>	 halfak, what's the namespace encoding? 0 = article?
[21:26:54] <halfak>	 wiggins, yeah.  0 = article.  Will get a more complete list. 
[21:27:06] <halfak>	 https://en.wikipedia.org/wiki/Wikipedia:Namespace
[21:27:28] <wiggins>	 thx!
[21:31:21] <guillom>	 halfak: nothing offhand either
[21:31:44] <halfak>	 Thanks HaeB & guillom 
[21:32:28] <wiggins>	 halfak, as examples we modified thanks/month to thanks/day and namespace starts w/ OMG! to WTF
[21:38:03] <HaeB>	 wiggins: perhaps also of interest: https://meta.wikimedia.org/wiki/User:Faebot/thanks
[21:40:26] <halfak>	 wiggins, if you ever want to see a bunch of curses, sort page titles alphabetically. ;) 
[21:40:53] <halfak>	 ^ also don't do that if you don't want to have a bunch of curses on your screen ;) 
[21:46:21] <wiggins>	 halfak: success! user gshahane modified query successfully. other group picked on an lvwiki query & couldn't tweak it usefully but pored through what was there
[21:47:17] <halfak>	 Cool :) 
[21:48:08] <wiggins>	 will be asking them to give feedback on querying system as portfolio entry for the week
[21:48:30] <halfak>	 wiggins, OK to share these notes with us?
[21:49:50] <wiggins>	 halfak, absolutely! aggregated anonymous notes. I put in my syllabus that I'll do that so they know the feedback goes to you.
[21:50:08] <halfak>	 great :) 
[21:52:55] <yuvipanda>	 wiggins: halfak \o/ awesome
[21:53:29] <wiggins>	 yuvipanda halfak thanks for the support! feedback to come. now prepping for guest lecture tonight... O_o
[21:53:56] <halfak>	 godspeed wiggins
[21:55:55] <yuvipanda>	 :D
[23:34:41] <yuvipanda>	 J-Mo: ping on email :D
[23:36:28] <J-Mo>	 I don't see it.
[23:36:37] <yuvipanda>	 J-Mo: oh
[23:47:58] <halfak>	 yuvipanda: what if I wanted to run a PAWS instance overnight. 
[23:48:06] <halfak>	 E.g. I want to parse an XML dump and that's going to take a while. 
[23:48:12] <yuvipanda>	 halfak: right now you can just let it run
[23:48:15] <yuvipanda>	 and it'll continue running
[23:48:17] <halfak>	 Cool :) 
[23:48:33] * halfak rages at "Untitled folder"
[23:48:36] <yuvipanda>	 heh
[23:48:53] <halfak>	 :D
[23:49:55] <halfak>	 yuvipanda, how do I access XML dumps from PAWS?
[23:51:49] <yuvipanda>	 halfak: should be in /public/dumps?
[23:52:17] <halfak>	 Awesome!  
[23:52:24] <yuvipanda>	 halfak: same as tools basically
[23:52:26] * halfak tried /mnt and /srv before asking
[23:52:42] * halfak needs to work on his knowledge of tools
[23:55:11] <yuvipanda>	 :D
[23:55:51] <yuvipanda>	 halfak: I just showed madhuvishy how to run a webservice with interactive development on notebooks
[23:55:55] <yuvipanda>	 it's pretty cool
[23:56:05] <yuvipanda>	 we don't have it in PAWS yet, need to stabilize other things first
[23:56:11] <yuvipanda>	 but I've demonstrated it's totally doable :D
[23:56:25] <halfak>	 yuvipanda, wooot!  That's going to be pretty cool. 
[23:56:32] <yuvipanda>	 yeah
[23:56:35] <halfak>	 The stuff that's in our .wsgi files. 
[23:56:36] <yuvipanda>	 it's just flask
[23:56:45] <halfak>	 Just put it in the notebook :) 
[23:56:48] * yuvipanda nods
[23:56:53] <yuvipanda>	 there's no reason
[23:56:56] <yuvipanda>	 .ipynb files can't contain code
[23:56:59] <yuvipanda>	 same way .py files do
[23:57:05] <yuvipanda>	 and be used for all the same things
[23:57:06] <halfak>	 Then, when we can import from notebooks, you put the routes there too. 
[23:57:08] <halfak>	 yeah. 
[23:57:21] <yuvipanda>	 notebooks: The Revenge of SmallTalk
[23:57:25] <halfak>	 lol
[23:57:56] <yuvipanda>	 I also need to add PHP support to this at some point