[17:10:16] <Ironholds>	 good morning happy researchers
[17:13:28] <halfak>	 Good morning Ironholds :D
[17:14:40] <Ironholds>	 how goes it?
[17:15:52] <halfak>	 Not bad.  Just doing some programming homework grading. 
[17:16:07] <halfak>	 I didn't realize that evaluating tasks would be so similar to grading software engineering homework. 
[17:16:10] <halfak>	 Same set of problems. 
[17:16:27] <Ironholds>	 heh
[17:16:37] <Ironholds>	 why are you grading someone's homework?
[17:16:41] <Ironholds>	 My morning can be summarised with https://twitter.com/quominus/status/660949012007141376
[17:18:32] <halfak>	 Ironholds, evaluating task submissions for our SWE opening. 
[17:18:36] <Ironholds>	 ohh, cool!
[17:18:41] <Ironholds>	 halfak, make the questions confusing
[17:18:44] <halfak>	 Woops. 
[17:19:01] <halfak>	 ^ re. accidentally forgot needed day to do thing. 
[17:19:10] <Ironholds>	 'you are in a locked room. Looking around you see a puddle of water, a single shoe and a key. What's the theoretical difference in search efficiency between a set and ordered vector?"
[17:19:27] <halfak>	 ha!
[17:20:08] <halfak>	 Mostly it was, "We have this one type of problem that comes up.  Solve it for us in an elegant way."  
[17:20:21] <halfak>	 I've gotten emails about how this problem would never happen in the real world. 
[17:20:23] <halfak>	 lol
[17:20:25] <Ironholds>	 ...bahaha
[17:20:26] <halfak>	 Just like students. 
[17:20:30] <Ironholds>	 halfak, I see that with analysts all the time
[17:20:52] <halfak>	 Also, a lot of fails for not reading the instructions. 
[17:20:59] <Ironholds>	 like I was giving someone advice about where he fell down in a task he submitted waybackwhen and went "I expected to see more exploratory data analysis to validate that the data wasn't somehow corrupted or in need of munging and filtering"
[17:21:00] <halfak>	 Might be great engineers, so that is frustrating
[17:21:11] <Ironholds>	 "wait, but surely the data is clean by the time it gets to us"
[17:21:20] <Ironholds>	 after I stopped laughing I explained
[17:21:24] <halfak>	 :D
[17:21:53] <halfak>	 ASSUME NOTHING.  THIS.  IS.  KNOWLEDGE.  *kicks applicant's messenger down a hole*
[17:22:09] <Ironholds>	 like, I'm dealing with a dataset that covers a 2 week A/B test that's taken 4 weeks because the schema broke in the middle and the second week we don't know if it's worth a damn because a different team decided to turn on a new feature on 12 hours notice like eejits
[17:22:30] <Ironholds>	 "clean". Sure, and I've got a dukedom in Scotland to sell ya.
[17:23:18] <YuviPanda>	 I'm going to do some of those too!
[17:23:19] <halfak>	 Ha. 
[17:23:28] <halfak>	 YuviPanda, <3 want to split some up?
[17:23:34] <YuviPanda>	 halfak: sure
[17:23:43] <YuviPanda>	 halfak: I can do maybe two now before I've to go to meetings
[17:24:10] * halfak PMs 
[17:38:59] <J-Mo>	 hey halfak, someone's asking for tools to detect reverted edits on the mediawiki-api list. What's the link for your repo, again?
[17:39:29] <J-Mo>	 also, happy Monday, researchers! 
[17:39:30] <halfak>	 J-Mo, http://pythonhosted.org/mwreverts/
[17:39:52] <J-Mo>	 sweet. I'll post, and direct them to research-l if they need future help. Thanks!
[17:40:32] <halfak>	 Note that, if they are doing small scale work (less than 100k revisions), they might want this: http://pythonhosted.org/mwreverts/api.html#mwreverts.api.check
[17:45:24] <J-Mo>	 pointers posted!