Perl 6 - the future is here, just unevenly distributed

IRC log for #webwork, 2012-07-16

| Channels | #webwork index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
01:26 djun joined #webwork
05:10 mgage__ joined #webwork
08:04 djun joined #webwork
11:44 CIA-105 joined #webwork
16:00 djun joined #webwork
17:00 mgage__ djun: hi
17:01 djun hi mike
17:02 mgage__ I fixed the issue with BEGIN_SOLUTION -- I'm about to send you a pull request.
17:02 djun great
17:02 djun thanks!
17:04 mgage__ it was pretty much what you pointed out -- the preprocessing of the file needs to be repeated when you include new material using includePGtext().  When new capabilities were added to the preprocessing routine in Translator they weren't also added to includePGtext.  I've changed things a bit so that includePGtext can actually get access to the preprocessing routine initially used by the translator -- hence things will stay in sync.
17:06 djun good stuff :)
17:06 mgage__ I tested it on a unit test that I put up, but it will help if you can do some testing also.  Testing on problems that have been redirected in the OPL should bring any further issues to light.
17:06 djun Sounds like a plan
17:07 djun I have to say I think that redirecting from duplicate problems is the wrong solution
17:07 mgage__ ok -- I've sent the pull request.  Unless you find something in the next couple of days I'll pull the changes into the openwebwork pg-dev
17:08 djun i should have a chance to look at it by then, I think
17:08 mgage__ me too -- it was a hack that used the existing tools at the time, but it can be done more efficiently and robustly  using a pseudo directory stored in memory.
17:09 djun agreed - this is something that should be happening at the level of relations, i.e., in a database or something similar
17:10 mgage__ I'm starting to get interested in the idea of using mongodb -- with a json directory at each node you can use attributes to describe the character of the node (redirect, directory,  directory for a problem with auxiliary files , .pg problem   etc. etc.) and also of the edge leading to the parent (parent directory, original problem of which this is a variant -- or this is a correction ) and so forth
17:11 djun jason was saying he'd started on something like this a while ago, iirc
17:11 mgage__ I'd like to at least consider that for webwork2 -- but a good place to start is with the OPL since that is somewhat independent -- and it's easier to do experiments without breaking everything.
17:13 mgage__ I missed that -- but we did talk about various databases.  couchdb came up first, but it looks like there is more activity around mongodb.  I haven't looked closely at either yet, but aubreyja may have.  jason grout is the one who first mentioned mongodb to me since apparently they have been experimenting with that for sage
17:13 djun whatever we do we need to decouple the representation of problems (e.g. file in a filesystem) from the software
17:14 mgage__ yes
17:14 djun maybe i was mis-recalling what aubreyja told me… perhaps it was couchDB.
17:15 mgage__ we were talking about both.  I don't have enough data to make any decisions yet.
17:15 djun having file paths as (not really) unique id's is awkward :)
17:16 mgage__ I think when I get back I can rewrite the current database (adding one table) so that the directory search actually takes place in software -- and I might be able to add the redirects to that -- that would help proof of concept that this can be done and still retain current funtionality
17:17 mgage__ If it turns out to be fairly easy I'll do it.  if it's hard I'll wait until we make more substantial changes.
17:17 mgage__ This would also make the creation of gateway quizzes less awkward.
17:18 djun i've been writing down a few notes about problem representation, library organization and software requirements
17:18 djun had to put it on hold for a few days with some more pressing but mundane work here
17:19 djun but I'll post what I have
17:20 mgage__ kk -- no rush for me -- I won't be looking closely at stuff until Thursday or Friday
17:20 mgage__ ttyl
17:21 djun cheers!
18:55 djun joined #webwork
19:08 djun joined #webwork
19:22 mgage__ left #webwork
20:32 djun joined #webwork
22:06 djun joined #webwork
23:26 djun joined #webwork

| Channels | #webwork index | Today | | Search | Google Search | Plain-Text | summary