# IRC log for #webwork, 2012-05-25

All times shown according to UTC.

Time Nick Message
00:36 mgage_ joined #webwork
00:40 CIA-39 joined #webwork
00:44 nmichalo joined #webwork
00:44 ChanServ joined #webwork
00:58 mgage___ joined #webwork
00:59 mgage___ Hi
03:01 JohnTravis joined #webwork
05:09 _ilbot joined #webwork
05:09 Topic for #webwork is now WeBWorK (http://webwork.maa.org) is an open-source online homework system for math and sciences courses. WeBWorK is supported by the MAA and the NSF. | Release notes: http://goo.gl/Ry5HN | Channel logged at http://goo.gl/jELTn
15:21 aubreyja joined #webwork
16:25 lanc joined #webwork
16:42 mgage__ hi
16:52 aubreyja joined #webwork
16:52 aubreyja joined #webwork
17:13 mgage___ joined #webwork
17:15 mgage____ joined #webwork
18:25 lanc joined #webwork
18:30 lanc Hi there, I'm authoring some problems for the University of British Columbia, and I would like to confirm something about the 'homework' assignment type in WeBWorK: is it true that there is no way to have multiple problems (distinct .pg files) show together on a single page?
18:31 lanc My current solution is to combine the problems into a single .pg file as parts (a), (b) ... of a single problem. This is basically the same, but it doesn't quite match the instructor's ideal.
18:31 lanc Thanks very much, Lance
18:59 mgage__ Hi Lance,
18:59 mgage__ lanc: hi
19:00 mgage__ you can combine homework problems together using the "gateway" mode.  This is usually used for quiz type presentation where the number of tries and perhaps the time allowed is limited.  But there a number of parameters you can set for the gateway quiz that might make it work just the way that you want it to.
19:03 mgage__ I've tended to combine problems parts into one problem in what I have written.  There is also a multipart problem -- which is still not fully developed:  http://webwork.maa.org/wiki/SequentialProblems#.T7_XR59Ys8Q   -- you should search the wiki for multipart and also for sequential for additional information
19:04 mgage__ One difficulty with gateway quiz is that there is no easy way for the various problems to communicate data with each other.  For that you need to use multipart or the plan that you described above.
19:05 mgage__ Just out of curiosity what is the downside to having the problems in a single pg file?
19:05 mgage__ I'd like to hear the story of the use case -- so we can consider it for future developments.
19:06 mgage__ lanc: hi
19:11 lanc I've been hesitant to use the gateway mode, mostly because I'm not very familiar with it, but I will look into your suggestion.
19:12 mgage__ it takes a little longer to set up because there are several options but 20 minutes fiddling the first time should let you know whether it will work for you or not.
19:12 lanc I tried to code this as a sequential problem, but I couldn't figure out how to implement limited attempts on each part - is this possible?
19:14 mgage__ I don't know off the top of my head.  the sequential/multipart problem is still unfinished in my opinion.  we threw something together that kind of worked, but there are many options that one could reasonably want, e.g. can you redo an earlier part of the problem, can you even see an earlier part of the problem,  etc. etc. etc.  so a version 2.0 redesign of the current implimentation is in order.  You intererested?
19:16 lanc To be honest, there isn't much in the way of a downside to having them all in a single pg file. The instructor wants there to be a set number of problems on each assignment, and if they are included as parts of a problem, it 'looks' like there are fewer problems. We certainly won't be upset if it's not possible.
19:17 lanc By interested, do you mean am I interested in there being a redesign, or am I interested in helping with the redesign?
19:18 mgage__ ok -- the reason for putting all of the problems in separate pages (particularly in the early days of the internet and browser development) was so that if a connection went down or a browser crashed at most one problems worth of work was lost.  -- that's not as important as it once was.
19:19 mgage__ Either, but we are looking for a larger crew of people interested in help with development -- at what ever level they are comfortable with.  the enterprise has grown quite a bit and it's getting harder to keep up :-)
19:23 aubreyja Hi -
19:23 mgage__ hi jason
19:23 mgage__ just looking at Geoff's code.
19:23 lanc I see. I must admit that when I envision having multiple problems visible on one page, I'm not quite sure how student would submit answers, and get feedback - i.e. would it be a page at a time or a problem at a time?
19:24 aubreyja I think this will be very popular
19:24 lanc The issue is mostly one of presentation, it doesn't really matter if the problems are on one or more pages
19:24 mgage__ it can be adjusted -- for example if there are several answer blanks on a page and you allow grading of partial answers, then students usually fill out part of the problem, check it and then continue.   This is often what we want
19:25 mgage__ but you can also require that they get the whole thing right without being told which answer blank contains a mistake.
19:25 aubreyja quick question: I haven't been following the github code for a month or so - is the openwebwork repo what I should switch too?
19:26 mgage__ I'd say in your case you should pull from my github jason -- I'm a little ahead of openwebwork right at the moment.  If you don't find anything that completely bombs I'll update openwebwork to my repo
19:26 lanc As to the redesign of sequential/multipart problems, I don't think it's a big priority from our perspective.
19:26 aubreyja sounds good - I'm doing it now
19:27 mgage__ for your mac -- get source tree -- it will give you a nice graphical picture of where the different repos stand with respect to each other
19:27 aubreyja cool
19:27 lanc I would be happy to contribute, but I'm not sure that I'm well-qualified. The little bit of digging I've done into the WeBWorK source code is my first experience with perl..
19:28 aubreyja btw, we should add mathjax as a git submodule so it gets pulled down automatically like in svn
19:29 mgage__ ok -- there are a number of people interested in multipart so it will probably improve at least this summer.  -- the more one works with it the more complicated the feature set. :-)   I also have some reservations (not really tested) as to whether it is a big advantage educationally -- I've tended to present long problems in which they can look ahead to see where the problem is going.
19:30 aubreyja also, re multipart, this is kind of a hack but it works:
19:30 mgage__ lanc -- everyone's welcome, -- perl (at least the way we use it) is not that much different from all the other scripting languages.  you can shoot me (gagemath.rochester.edu) an email and I'll try to keep in the loop on projects.
19:32 mgage__ thanks jason, I'd forgotten about that page in the wiki.
19:32 mgage__ (keep you in the loop on projects) -- I'm typing too fast
19:33 lanc okay, i'll do that.
19:33 mgage__ jason -- do you know where geoff stores the persistent data for the achievements ?  -- he's created a new table ? or some kind of cookie storage?
19:33 lanc thanks for the help, I'll go try out the gateway idea
19:34 mgage__ have fun -- let us know how it works out   Gavin LaRose is the designer of gateway and the goto guy if you have suggestions or questions
19:34 aubreyja I was going to ask that - his achievements are in the file manager, maybe he is creating a log file there or in logs
19:35 aubreyja swapped out svn pg for your github pg and so far so good, now for ww2
19:35 lanc left #webwork
19:36 mgage__ if you could look into grabbing mathjax as a git submodule that would be helpful.
19:37 aubreyja yep, I was going to do that
19:37 aubreyja my sysadmin would have a heart attack if he knew what I was doing right now :)
19:37 aubreyja switching to a new vcs and getting the bleeding edge code
19:38 mgage__ I've found that using  svn co https://github.com/openwebwork/pg/trunk   grabs just what you want for a non-contributing repository -- so we could consider switching to that ---- the current version of svn is out of date with respect to github/openwebwork
19:38 mgage__ don't tell him :-)
19:38 mgage__ that's my policy
19:39 mgage__ I'm going to slap a 2.5.0 tag on the current version and then update the svn trunk.
19:39 aubreyja that's a good idea - on the wiki I set up a template for the svn url so we could just switch it in one place and it would propagate to every page that uses it
19:40 aubreyja (in theory, we haven't changed it since I set it up)
19:40 mgage__ ok --- another thing you could do -- perhaps this is what you meant -- is switch the POD documents so that they are updated from github.com/openwebwork
19:41 aubreyja ah, right, good idea - I'll do that
19:41 mgage__ that could be done anytime now -- github.com/openwebwork is pretty stable -- and it should be quite easy to keep it stable since there are alternative ways to share bleeding edge code.
19:42 mgage__ If you are just upgrading webwork2 you may need to spend a little time adjusting the new prelocal.conf and postlocal.conf files -- the idea is that this will be a one time thing and after that you won't have to mess with global.conf.dist anymore.
19:42 mgage__ brb -- it's the dog's meal time
19:42 mgage__ :-)
19:43 aubreyja just now looking at that
19:43 aubreyja any changes to webwork.apache2-config?
19:45 aubreyja interesting - I did a diff with our webwork.apache2-config and the webwork.apache2-config.dist in your github and the only difference is $webwork_dir 19:46 mgage__ yep - we're trying to get the changes down to the very bare minimum 19:48 aubreyja I guess we need that for the evals below it, but if we put the webwork libraries in @INC we might be able to get away without even needing that 19:49 mgage__ I think the main reason we need it is to find the webwork libraries. 19:49 aubreyja (that's what linux dists would want us to do also) 19:50 mgage__ hmm -- so what mechanism do they use to add it to @INC? is that easier than adjusting the$webwork_dir variable?
19:50 aubreyja right, the evals are "veal use lib…" but if they are already in @INC we
19:50 aubreyja wouldn't need to find them
19:51 mgage__ right -- and we wouldn't need to define the pg directory in global.conf.dist either
19:51 aubreyja well, modules from cpan go somewhere like /usr/lib/perl5/site-perl/
19:51 mgage__ but how do we reset the @INC ?
19:51 aubreyja whereas stock perl and the core library goes in /usr/lib/perl5/vendor-perl
19:52 aubreyja we would use the existing @INC - the /site-perl/ is already in it.
19:52 mgage__ I see, we could consider doing a redesign like that -- it might make it more compatible with standard installation methods.  Let's talk about that with Remy (and some of the UR CS guys if they come) and Steve Jacobs when you come.
19:54 mgage__ It's kind of a big change in philosophy but it wouldn't be that hard to do --- and if it brings us in step with current open source practice then that would be a plus
19:54 aubreyja sure - and that directory is set in the %Config hash and we have access to it whenever perl is installed
19:56 mgage__ definitely worth looking into -- Remy told me about an RIT Remix program (or some similar name) that helps automate creating spins (distros with specific programs) -- I was going to ask him more about that next week when I'm back in Rochester.  See if we can get some USB sticks set up for the code camp
19:56 mgage__ we're you able to get code from my github? I can send you a pull request if that will help.
19:56 aubreyja ah, cool - I think this camp will be great and very good for the project
19:57 aubreyja yep, just doing config stuff
19:57 mgage__ hope so. I'm working on listing bite sized chunks.
19:57 aubreyja (not much to do, but checking carefully before I replace webwork2)
19:58 mgage__ yeah -- most of the overrides go into postlocal.conf --- sometimes you have to override a couple of items if the first item was used to define the second.
19:59 mgage__ still taking suggestions about ways to organize or reorganize those three files.
20:29 aubreyja I think it looks good - the only think you might want to consider moving to prelocal or postlocal is maybe the choice of Authen module and the corresponding settings hashes (e.g. in my case I had to edit global.conf for ldap) and the display mode selection and defaults (although maybe we can assume users will want our recommended choices in general here)
20:35 mgage__ try overriding the Authen module setting in postlocal.conf and leaving the setting in global.conf.dist as is. (also use global.conf.dist  and remove global.conf unless you make massive changes to global.conf) -- that way you have an automatic default setting for Authen but you also have local overrides.
20:36 mgage__ if it works you can add a commented out template to postlocal.conf.dist so that people will have a better idea of what
20:36 mgage__ needs to be changed
20:36 aubreyja ok, I'll have to do that this evening (time to run some errands and pick up the kids).  everything seems to be working except I get _LOGIN_MESSAGE on the front page instead of the login message
20:36 aubreyja didn't we help someone with that problem?
20:37 mgage__ there is something missing from one of the Localize packages
20:37 aubreyja maybe
20:37 mgage__ either Localize.pm or the en.pm in the directory Localize doesn't have a translation for that phrase
20:37 aubreyja probably - I'll bet we never installed the prerequisite perl modules
20:37 mgage__ that would do it -- run check_module.pl
20:37 mgage__ check_modules
20:38 aubreyja yep, missing Locale::Maketext::Lexicon...
20:38 aubreyja alright, well that was pretty easy in any case
20:39 aubreyja I've got to run, but I'll be around this weekend if anything comes up
20:39 mgage__ updates should be even easier once we get this right because all of the overrides you care about will be in postlocal.conf
20:39 mgage__ ttyl
20:39 aubreyja yep, ttyl
20:39 mgage__ I may not be around Saturday -- but I will be Sunday and Monday
22:20 mgage___ joined #webwork