IRC log of #schooltool for Monday, 2012-06-25

*** ignas has quit IRC01:06
*** yvl has quit IRC09:10
*** yvl has joined #schooltool09:18
*** ignas has joined #schooltool12:39
*** menesis has quit IRC13:05
*** povbot_ has joined #schooltool13:31
*** povbot has quit IRC13:31
*** ignas has quit IRC13:32
*** ignas has joined #schooltool13:38
*** yvl has joined #schooltool13:45
*** menesis has joined #schooltool14:26
*** replaceafill has joined #schooltool16:21
*** th1a has joined #schooltool16:29
th1ahi yvl, menesis, aelkner, replaceafill.16:31
yvlgood morning16:31
aelknermorning16:31
replaceafillgood morning/afternoon16:31
th1aA thunder storm is right overhead, so we'll see if the power goes out.16:31
th1ayvl: Would you like to start?16:32
yvlok16:32
yvlI've played around with some stuff:16:32
yvlenabled zeo16:32
yvladded configs of celery and redis to paste templates16:33
yvland put everything under suprevisord16:33
yvleverything runs as daemonized services now (and schooltool can run as well)16:34
yvlso starting up will be somewhat faster16:34
yvl(for developers)16:34
th1aCan we set it to restart schooltool if it happens to go down?16:34
yvlbecause usually devs only need to restart Zope16:34
yvlactually, yes16:34
yvlbut only if it goes down16:35
yvlit may get clogged for the lack of better word16:35
th1akk16:35
th1aI assume menesis is ok with this.16:35
yvlI don't think he looked over the changes16:35
yvlyet16:35
menesishaven't looked16:36
yvlI'm now working on actual task queuing api16:36
th1aBut in principle menesis is ok with supervisor.16:36
yvland then I'll work on whats needed for tests16:36
yvlyes th1a, AFAIK16:36
yvlthen there's the web part16:37
yvlI plan to work on these this whole week16:38
yvland maybe few days of the next week16:38
yvldepends on how many rakes there are along the way16:38
th1aCan we use ZEO for hot backups?16:39
* yvl shrugs16:39
yvlZEO raid, probably yes16:39
yvlbut that is another story16:40
yvlsee http://pypi.python.org/pypi/gocept.zeoraid16:40
yvlzeo just allows more than one zope to connect to the DB16:41
yvlfor example, ST *and* remote tasks :)16:41
yvlalso CL may be interested in those16:42
yvlor other actual clients, when you think about it16:42
th1aYes... my general understanding was that it was the safe way to read the db for a live backup as well.16:42
yvlwell, a safe way for a live backup16:42
yvlis to copy the DB file16:43
yvlfrom command line16:43
yvlZODB is append-only and it's smart enough to reject unfinished transactions16:43
th1aDo we need that now for the celery stuff?16:44
yvlyes and no16:45
*** pgulley has joined #schooltool16:45
yvlI could have done stuff without it16:45
yvlbut this way is easier16:45
yvland sane16:46
yvlI think16:46
yvlfor example, supervisord16:46
yvlwe could just run the services manually16:46
yvlfor all I car16:46
yvle16:46
yvlzeo16:46
yvlwe could just post to relevant URLs16:46
yvlI mean, it's software.16:47
yvleverything's possible16:48
yvlit seemed like a less of a hassle to me16:48
th1aOK.16:48
*** dreich has joined #schooltool16:48
th1aAll right sounds good yvl.16:49
th1aKeep menesis in the loop.16:49
yvlI am, th1a16:49
yvl(we already talked briefly before)16:50
th1aThanks yvl.16:50
th1amenesis?16:50
menesisno progress on bugfix release, sorry16:52
menesisI added a few bugfixes to trunk16:52
menesisother than that, worked on philippines16:53
th1aOK, that's fine.16:54
th1aI'm not feeling like the bugfix release is urgent.16:54
th1aThanks menesis.16:54
th1areplaceafill?16:54
replaceafillok16:55
replaceafilli moved the Done buttons to the bottom of the page the year views16:55
replaceafill(/sections, /courses, /resources, etc)16:55
replaceafilli've styled them as secondary action buttons too16:56
replaceafillth1a, are we going to style "all" the Done links?16:56
replaceafilllike the one in person index?16:56
th1aYes... they certainly should all be the same.16:56
replaceafillshould all the be at the bottom of the page too?16:56
replaceafillthey*16:57
th1aThat's partly why my feeling is that they should be a bit different than form buttons.16:57
th1aI'll need to look at them, I've been reluctant to put them under the search results, since they can be long.16:57
replaceafilltrue16:57
th1aLet's just review all the changes later.16:58
replaceafillsure16:58
th1aMe just telling you things doesn't help.16:58
replaceafillagree16:58
th1aThat is, me pulling guesses out of my butt doesn't help.16:58
replaceafill:D16:58
replaceafillok16:58
replaceafilli also worked on the new assign skills view for courses16:58
replaceafilli'll have it done today i think16:58
replaceafillwill show it tomorrow to david16:59
replaceafillquestion16:59
replaceafilldavid said he'll be only available early in the morning, correct?16:59
replaceafilland late at night16:59
aelknertoday, he said he can meet in the late afternoon16:59
replaceafillah ok16:59
th1aOr evening.17:00
aelknerbut as a rule i think you're right17:00
th1aI think we should try to get as much done as we can asynchronously at this point.17:00
replaceafillah ok, i'll show him my changes tomorrow17:00
replaceafillmenesis, sorry about my misunderstanding of the ommit changes17:01
replaceafill(the list conversion part)17:01
replaceafilli just thought you just missed the /resources failing test17:01
replaceafilli didn't know ignas requested the change17:01
menesisno problem, thanks for noticing problems I left17:01
replaceafill:)17:02
replaceafilli think that's it from me17:02
th1aSo the table format issue is resolved?17:02
replaceafillyes17:02
replaceafilloops sorry17:03
replaceafilli thought you meant the style issue17:03
replaceafillwhich is fixed17:03
replaceafillmenesis fixed it17:03
replaceafillrelationship views look good now17:03
replaceafillah17:04
th1aOK.  What are you working on now?17:04
replaceafillyvl, thanks for the chromedriver advice17:05
replaceafillth1a, the assign skills to courses view17:05
th1aOK.17:05
replaceafillth1a, what time can we look at the buttons style changes?17:06
replaceafillin the afternoon?17:06
th1aYeah.17:06
replaceafillok, i'll be here17:07
* replaceafill done17:07
th1aThanks replaceafill.17:07
th1aaelkner?17:07
aelknerok, feeling much better now17:07
aelkneri made the changes to the virginia importer that i discussed with welsh on fri and loaded the data on my demo server17:08
aelkneri'll be showing it to him this afternoon, he said he'd contact me to look at it17:08
th1aOK.17:08
aelkneri pivoted to finishing up the core sections sheet importers17:09
aelknersomething occurred to me when i looked at setting up the sample data17:09
aelknerremember how we decided to use comma delimitted lists for courses and resources17:10
aelknerand then the start end terms, i convinced you that just have a single Terms column would be nicer than start and end cols17:10
*** paulproteus has quit IRC17:10
aelkneri still think that's the case, but i think that using a '-' to delimit them is not a good idea17:11
aelknerour names chooser for ids is to take the title and use only the alphanumeric parts17:11
aelknerand convert spaces to '-'17:11
aelknerso '-' is a valid part of an id, meaning it doesn't make a good delimitter17:12
aelknercould we go with start/end17:12
th1aOK.17:12
th1aJust require a list?17:12
aelkner2010-fall/2010-spring?17:12
aelkneri don't like the list idea because it implies flexibility that we don't have17:13
aelkner2010-fall, 2010-spring?17:13
aelkneri don't like that, and to have a list of all four terms (if that be the case) would be sloppy17:13
th1aWhat's wrong with commas now?17:13
th1aIt isn't sloppy.17:13
aelknerit implies that they could just pick which terms, but we require them to be consecutive17:14
aelknerso MP1, MP3, MP4 would not be valid17:14
th1aJust do start and end then.17:14
th1aTwo columns.17:14
*** paulproteus has joined #schooltool17:14
aelknerok, but we decided that all cols are required, no empty cells17:14
th1aAnd if there isn't an end we assume it is just one.17:14
th1aIt is ok with me if there is an empty cell.17:15
aelknerok, just for the end term, otherwise, the other cells are required17:15
aelknerand they can specify the end term as the same as the start term if they like17:15
th1aWe just decided that we weren't going to use the "blank cell takes the value from the row above" thing.17:16
aelknerright, so those cells will be required because we need their vales, but the end term can be optional17:16
th1asure.17:17
aelknerok, that's what i'l do17:17
aelknerthat's it for me17:17
th1aLet's try to wrap this up now.17:17
th1a(the imports...)17:17
th1aI have a meeting in Providence with someone from http://www.fabacademy.org/17:18
th1aHopefully they can use the new CanDo.17:18
th1aI'll need to use one of the demo instances, we can get that sorted Wednesday morning.17:18
th1aOK.  Thanks guys.  I think jelkner has to start using this next week so we need to be ready.17:20
th1aSee you Wednesday.17:20
* th1a drops the bag of gravel.17:20
yvlGood luck guys!17:20
aelknerthanks guys, cya wed17:20
replaceafillthanks everybody17:20
*** replaceafill_ has joined #schooltool17:30
*** replaceafill has quit IRC17:30
*** replaceafill_ is now known as replaceafill17:30
pgulleyreplaceafill, dreich and I have our instances running and were starting to dig into some of the problems17:37
pgulleycan we re-organize the views?17:37
pgulleyin pyquiz, that is.17:37
replaceafillre-organize?17:37
dreichmove them from one big file to a lot of small files in one directory17:37
replaceafillah sure17:38
dreichcool, we'll do that17:38
replaceafillfeel free to do whatever you think it's best17:38
pgulleyawesome, will do.17:38
replaceafillpgulley, did you install python2.6?17:38
pgulleyyes, we did.17:39
dreichwe had to use a ppa for it that we found on the internet somewhere, but it works beautifully17:39
replaceafillglad to know17:39
*** highvoltage has quit IRC19:31
*** highvoltage has joined #schooltool19:56
*** menesis has quit IRC20:03
*** ignas has quit IRC20:21
*** menesis has joined #schooltool20:54
pgulleyreplaceafill, we're trying to tackle the issue of editing deployed quizzes, and need advice20:55
replaceafillpgulley, yes?20:56
pgulleyor clarification. How would you have us handle students who have already taken the deployed quiz who may be affected by the edit?20:56
replaceafillthat's a good question for jelkner20:57
replaceafillhe and i have different pov on the issue20:57
replaceafillimho we have to detect if the quiz has already been taken20:57
replaceafillif it hasn't, we can change it20:57
replaceafillunless jelkner wants some other kind of logic20:57
replaceafilllike making the student taking it again or whatever20:58
replaceafillfor me the design was: put the quiz on stone (by publishing it)20:58
replaceafillbut jelkner says he's an "average" user20:58
replaceafilland average users make mistakes20:58
pgulleyso maybe there is a step in between deploying the quiz and having it available to take? just to force users to double-check20:59
replaceafillthat could work21:00
replaceafillto me it seems like to many steps, but maybe jelkner is fine with it21:00
replaceafilltoo*21:00
dreichI like your first idea, it being editable up until the first student takes the quiz21:00
replaceafillyes21:00
replaceafilland we should have that kind of "detection" anyway21:01
replaceafillbtw21:01
replaceafillanother thing i don't like about the current design21:01
replaceafilli don't remember if i changed it21:01
replaceafillPersons in pyquiz are just containers for answers, right?21:01
pgulleybasically21:01
replaceafillbut the answers are indexed by strings21:01
replaceafilli'd like the answers to be indexed by question21:02
replaceafillso you don't have to look into all the answers to get the answer to a question21:02
replaceafillyou know what i mean?21:02
replaceafillit's not an usability issue, so there's no bug for it21:02
replaceafillbut if you could tackle that change, that would be nice21:02
pgulleyright, so persons become containers for whole answers grouped by quiz?21:03
replaceafilli mean, now, person objects look like21:03
replaceafillpgulley = {'1': <answer>, '2': <answer>, ...}21:04
replaceafilli'd like them to be21:04
replaceafillpgulley = {<question>: <answer>, <question>: <answer>, ...}21:04
replaceafillthe question becomes the key21:04
replaceafilli didnt make that change because i thougth jelkner could ask for21:05
replaceafillseveral answers to the same question21:05
replaceafilllike history21:05
replaceafillbut in any case it would require a change in the code21:05
replaceafillit felt like YAGNI at that time21:05
replaceafilland maybe still does21:06
replaceafillbut i think it would get rid of some unnecessary traversing code21:06
pgulleyIt seems like a simple fix, either way.21:06
replaceafillcorrect21:06
pgulleyso we have it, no problem21:07
replaceafillas long as you don't have to evolve the data ;)21:07
replaceafilli think you're in a very nice time to modify everything21:07
replaceafillsince jelkner will start fresh21:07
replaceafillit becomes harder once he already has data21:07
replaceafillplease fix it if you have time for it21:07
replaceafilland if you think it will make the code easier to handle21:08
replaceafill(that's the way we do it in schooltool gradebook btw)21:08
pgulleyokay, so since we don't know exactly how we want to handle editing deployed quizzes, we will go ahead and fix this now21:08
replaceafillsure, ask jelkner about the "setting in stone" part :)21:09
*** highvoltage has quit IRC21:11
* replaceafill goes to get lunch, bb in ~ 1 hour21:13
*** th1a has quit IRC21:50
*** ignas has joined #schooltool22:22
*** ignas has quit IRC22:30
pgulleyreplaceafill, if you're here: it looks like this is already the case22:33
pgulleyusers are dictionaries of Question ID's to Answer Objects22:33
pgulleywe /think/22:33
*** ignas has joined #schooltool22:35
*** ignas has quit IRC22:43
*** ignas has joined #schooltool22:44
dreichnope, turns out that's wrong22:46
dreichpeople are {answer id : answer object}, as you thought22:46
dreichand it looks like that's pretty core zope stuff, too...22:46
*** ignas has quit IRC22:51
*** menesis has quit IRC22:56
*** th1a has joined #schooltool23:07
replaceafillpgulley, dreich i'm back23:22
pgulleyalright. we're just wrapping up here.23:22
replaceafillthey were indexed by string, correct?23:23
pgulleyit looks like the fix you requested is a little more complex than we thought.23:23
replaceafill(i mean, the answers)23:23
replaceafillpgulley, ah ok23:23
pgulleyyes- strings of floats?23:23
replaceafillnp23:23
replaceafillwe can continue with that approach meanwhile23:23
replaceafillgood work on the bugs fixed gents23:24
pgulleyalright.23:24
pgulleywe've made progress, at the very least, and it should move quicker herein23:24
replaceafill:)23:24
pgulleyawesome, we're heading out23:27
replaceafillsee u tomorrow guys23:27
*** pgulley has quit IRC23:27
dreichyep, see you!23:27
*** dreich has quit IRC23:27

Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!