Perl 6 - the future is here, just unevenly distributed

IRC log for #webwork, 2012-11-02

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

All times shown according to UTC.

Time Nick Message
00:12 ionparticle hi mgage, do you know of any bugs where students submit answers and the answers doesn't get recorded?
00:13 ionparticle to be specific, it's trying to add entries to past_answers in the database, but fails cause past_answers doesn't seem to exist
00:13 mgage there have been reports of this -- sometimes it has been determined that the student was actually working as guest -- I can't completely rule out the possibility that some results aren't recorded.
00:14 mgage that's a more specific bug than  I've heard before -- are you seeing an error in the log about this?
00:15 mgage also if one cares enough it is possible to examine the log data to determine whether a student has logged in to webwork and or submitted answers for  a problem during the time they claim to.
00:15 ionparticle might be a different one, I get an exception
00:15 ionparticle Can't call method "new" on an undefined value at /www_data/webwork/webwork2/lib/WeBWorK/DB.pm line 290.
00:15 ionparticle in WeBWorK::DB::__ANON__ called at line 112 of /www_data/webwork/webwork2/lib/WeBWorK/ContentGenerator/ProblemUtil/ProblemUtil.pm
00:15 ionparticle in WeBWorK::ContentGenerator::ProblemUtil::ProblemUtil::process_and_log_answer called at line 1244 of /www_data/webwork/webwork2/lib/WeBWorK/ContentGenerator/Problem.pm
00:15 ionparticle in WeBWorK::ContentGenerator::Problem::output_score_summary called at line 152 of /www_data/webwork/webwork2/lib/WeBWorK/Template.pm
00:15 ionparticle in WeBWorK::Template::template called at line 496 of /www_data/webwork/webwork2/lib/WeBWorK/ContentGenerator.pm
00:15 mgage sounds like you are using 2.5.2
00:15 ionparticle in WeBWorK::ContentGenerator::content called at line 200 of /www_data/webwork/webwork2/lib/WeBWorK/ContentGenerator.pm
00:15 ionparticle in WeBWorK::ContentGenerator::go called at line 393 of /www_data/webwork/webwork2/lib/WeBWorK.pm
00:15 ionparticle yes
00:16 mgage -- probably a new bug then -- prior to 2.5.2 past_answers were stored in a text file   -- not surprising since 2.5.2 is just beginning to get wider use
00:17 mgage send in an issue to github -- you can also alert Geoff Goehle (you know him) -- he wrote the new past_answers mechanism while implementing the essayQuestion type
00:17 mgage you know him?
00:17 ionparticle yes
00:17 ionparticle ok, thanks
00:17 ionparticle :)
00:19 mgage thank you -- glad people are looking at this -- there is a lot of good stuff in 2.5.2 (even more in the version in the github.com/mgage repository -- but the css for the new features is dicey enough that I haven't made it generally available yet -- need time/personnel to work on the css front end first
00:20 ionparticle what's wrong with the CSS?
00:22 mgage it's just messy -- Peter Staab and David Gage added the ability to display assignments by calendars for the student homepage (and more importantly for the homework editor page) -- probably not feature complete yet, but it also may or may not look good depending on your window size, etc. etc.
00:23 ionparticle ah
00:23 mgage I think it's a step forward in usability for most people in terms of visually understanding what has been assigned and when
00:24 mgage one of the things that was worked on (just not quite finished) at Fitchburg code camp
00:35 rbeezer joined #webwork
00:49 ionparticle https://github.com/openwebwork/webwork2-dev/issues/64
00:58 mgage ionparticle: ty
01:08 Paul_Pearson joined #webwork
01:09 djun joined #webwork
01:13 goehle joined #webwork
01:13 goehle hey mike
01:13 mgage hi geoff
01:13 goehle my guess is that the past_answers database doesnt exist for some reason
01:14 goehle although as long as someone has logged in as admin it *should* exist
01:14 goehle I guess its possible nobody has logged in as admin since the upgrade
01:14 mgage it's possible that hasn't happened
01:14 mgage ionparticle:   have you visited the admin course yet?
01:14 goehle another possibilitity is that the database.conf file is either old or customized
01:14 goehle and hasn't been properly upgraded
01:15 mgage kk -- in any case I gather that this behavior can't be replicated on your installation
01:15 goehle past_answer => {
01:15 goehle record => "WeBWorK::DB::Record::PastAnswer",
01:15 goehle schema => "WeBWorK::DB::Schema::NewSQL::Std",
01:15 goehle driver => "WeBWorK::DB::Driver::SQL",
01:15 goehle source => $database_dsn,
01:15 goehle params => { %sqlParams,
01:15 goehle non_native => 1,
01:15 goehle },
01:15 goehle },
01:16 goehle the database.conf file should contain that
01:16 mgage (--- i'm afraid I haven't had  a chance to check it yet myself and won't until this weekend.
01:16 mgage kk
01:16 goehle um, my current installations both work
01:16 goehle but I haven't touched htem in a while
01:16 goehle I'm certainly able to assign and grade hw
01:17 mgage I don't think these database parts of 2.5.2 have been touched so the next step is to make sure that the installation is up to date
01:17 mgage ionparticle:hi
01:18 goehle is it a new installation?
01:19 mgage yes
01:19 goehle but not an upgrade?
01:19 mgage no I believe it's an upgrade to 2.5.2   -- ubc has been using webwork for a while but they are just starting to work with 2.5.2 because they want essay answers (I believe)
01:20 goehle ah
01:20 goehle well my guess is that the stuff contained in database.conf is being masked by something
01:20 goehle maybe an old global.conf file
01:20 mgage but it's not uncommon to forget to update the database -- I do that myself quite often and you would think i would have learned by now
01:20 goehle I guess thats what I would do first
01:20 mgage probably not global.conf
01:20 goehle log in as admin
01:20 goehle check and see if that fixes it
01:21 goehle then check database.conf to make sure it has a past_answer entry
01:21 goehle then I guess I dont know
01:21 mgage that sounds reasonable.  thanks.
01:21 mgage we'll worry about the I don't know part after we've done these checks
01:21 goehle :)
01:21 mgage thanks again
01:22 goehle no problem
01:22 goehle I hope that helps
01:22 goehle I haven't had time to do much with webwork unfortunatelly
01:22 goehle I did want to ask a quesiton though
01:22 mgage sure
01:22 goehle are sets "supposed" to work properly if they aren't assigned to anyone
01:23 goehle in particular there was an issue the other day where if you click "try me" in the hmwk set editor
01:23 goehle when the set isn't assigned to you then you get an error
01:23 goehle I know the system autoassigns the set to you
01:23 goehle and I seem to remember a comment saying that it needed to happen or there would be wierd behavior
01:25 mgage there are a couple of cases where some initializations might get done the first time they were assigned to someone.  (I don't think this was *supposed* to happen, it just did) -- rather than track all of these down Arnie (and sometimes me) made sure that all (almost all) sets are assigned to someone.  If you build a new set from the library or from the homeworkset editor it automatically is assigned to the current user.
01:26 goehle if you use "create set" then it also automatically gets assigned
01:26 goehle well, the issue I mentioned is a bug.  I'll track it down next time I get a chance to do some coding
01:27 mgage the original assumptions in WW1.9 were that a problem was always in a set and a set was always in a course  -- this was modified/hacked to make the library work (hence the concept of unset) --
01:28 goehle yeah, I figured that was maybe the issue.  The error comes from the system looking for an undefined variable that must come from having a proper userset/userproblem entry
01:28 mgage yeah that was our attempt at a fix (or workaround) -- I think what is going to come out of the Fitchburg work on the homework sets editor is a more robust model of problem and set that will hopefully cut some of these bugs out at the ground floor.  meanwhile we can use the retro fit patches.
01:29 goehle yeah, I was talking with the cs guys about some of this
01:29 goehle now that there are versioned sets they were saying we should rethink how the classes are structured
01:30 goehle in particular everything should be a versioned set and then unversioned sets would be a subclass
01:30 goehle or something, it was a little while ago and we were out having libations
01:30 mgage there is also a habit of editing and locally modifying problems while still in the library and then adding them to the set  -- I tended to add candidate problems to the set, modify them there and then discard them if i decided not to use them.  -- that means there a a few more bugs that appear for the first method than for the second
01:30 goehle oh, I"ve never used the first method
01:30 goehle that seems dicey for some reason
01:31 mgage yep -- I completely agree with that  -- the whole gateway quiz was bolted on to a model that wasn't quite designed for it.
01:31 goehle it shows a little bit in the code.  There are places with some copy/pasting going on to deal with the sets and then versioned sets.  Then again I suppose thats par for the course.
01:32 goehle you cant rewrite your whole system everytime you want a new feature :)
01:32 mgage I thought so also -- but many people do it.  The main reason it's dicey is that it's not the method I used at the beginning so I didn't find the bugs first :-) -- in principle it shouldn't matter
01:33 goehle how do they assign them to the set after they have edited them?
01:33 goehle I usually use the "and replace in set soandso" option
01:33 goehle but that wouldn't work if you haven't even created the set yet
01:33 mgage that's true -- the fact that Gavin was able to attach gateway quizzes is something of a tribute to the original design -- but he had to work much harder than should really be necessary and the next database model should at least allow us to simplify the gateway code.
01:34 goehle oh right, the database.  I suppose that has to be properly set up before you can really start thinking about restructuring the object classes
01:34 goehle or at least set up with restructuring the classes in mind
01:34 mgage it's not that they haven't created the set, it's that they modify it from the library page (while it's attached to the "unset") and then add it to their real set.  -- the unset was also bolted on -- by this time it's pretty compatible but I still don't trust it completely.
01:35 goehle huh, so you modify it, save as a problem in the unset, and then when you "add to target set" it gets added?
01:35 mgage well -- I've kind of put off doing a lot of work refitting things if getting the database rationalized will make the work a lot easier (unless of course it's a feature/bug that needs to be fixed NOW)
01:37 mgage actually I'm not 100% sure how it's usually done -- I know that people play with variations on the problems in the library without placing them in sets first
01:38 goehle I guess I didn't realize that if you edit a problem you could save it to the library
01:38 mgage I'm not sure that a temporary edit would be moved to a set.
01:38 goehle do your course/templates problems appear in the library browser?
01:39 mgage you can't save it to the library (unless you have super permissions) but it will be saved to a temporary file while you are playing with it.
01:39 mgage except possibly for a development machine you should have things set up so that the library problems themselves cannot be modified by the webserver (no matter who is logged in to the webserver)
01:40 mgage the only way to modify a library problem is by command line access
01:40 mgage but you can do temporary modifications (think of the modifications as being held in memory -- although actually they
01:40 mgage are in a temporary file on the server.
01:42 goehle yeah
01:42 goehle taking a look now
01:42 goehle so if you click edit it
01:42 goehle then you can save as ain independent problem
01:42 goehle how do you view those?
01:43 mgage you would need to add those to a homework set using their path.  -- there is currently no way to view an independent problem
01:43 goehle huh, wierd.  I guess I don't understand how people are editing first then adding to the set
01:43 mgage (you can view those using a client editor easily enough -- and some version of that should be the way the next PG editor works -- the current one tries to do too much)
01:45 goehle so whats the word on library cleanup.  I was having trouble the other day because all of the problems I was adding to my homeworks didn't actually have randomization, just fixed numbers
01:45 mgage they might be editing the manually -- or they may have just been editing temporarily and ran into something that didn't work when you were saving to a temporary file (until recently auxiliary files weren't handled properly in that case for example -- I'm not sure they are always handled properly even now)
01:46 goehle interesting.  I'm not very good at predicting different workflows, so its interesting to see how other poeple do it
01:46 mgage I haven't heard much on the library cleanup recently -- I think eveyone has their heads down concentrating on mid-semester (true for me -- I'm on semi-leave and taking 4 CS classes -- just finished the last of my midterms :-) )
01:47 goehle oh nice
01:47 goehle I'm a little jealous
01:47 goehle My wife Naomi has been taking the MIT edX python course and it seems very interesting
01:47 mgage one my classes is HCI (human computer interface) -- and they keep emphasizing that "you are not the user" -- they over do it a bit I think -- but I understand their point.
01:48 mgage it is interesting -- a little humbling to be on the other side of the podium -- I'll probably be a bit more sympathetic next semester when I'm back at work.
01:48 goehle :D
01:49 goehle that sounds a little like the robot/exoskeleton analogy someone was describing to me
01:49 goehle its the difference between writing code which allows you to do things efficiently and effectively, but requires a lot of knowlege to run
01:50 mgage I tried to keep up with Keith Devlin's coursera  course on mathematical thinking -- mainly because he's been quite reflective about what MOOC's are and what they might be good for -- I'm afraid the local classes got priority however.
01:50 goehle and code that is much more autonomous, but maybe not quite as fexible
01:50 mgage yes -- the efficiency vs. ease of use comes up often --- the ease of use people tend to win, but I still have hold out a little for the
01:51 goehle I think MOOC's are interesting.  Naomi has taken a couple now.  I think that it takes a special kind of student to get a lot out of those courses
01:51 mgage efficiency group that wants to educate people on a "better" way to do things -- not just make it work "as they expect".  like most principles neither of these should be overdone
01:52 goehle sure
01:52 goehle that sounds like gnome-shell actually
01:52 mgage Have her take a look at Keith Devlin's blogs and see what she thinks
01:52 mgage or emacs
01:52 goehle sure emacs too
01:53 mgage http://devlinsangle.blogspot.com/2012/05/math-mooc-coming-this-fall.html
01:54 goehle Intersting, but I dont know that its going to be that big a deal
01:54 goehle people said some of the same stuff about correspondence courses back in the day
01:55 goehle cheating was rampant in Naomi's electrical engineering course
01:55 goehle and even then only a tiny fractoin of people actually finished it
01:55 mgage he has some interesting comments on that also   http://mooctalk.org/
01:56 mgage two comments:    moocs will replace textbooks not profs,   and   it's the social media (facebook etc.)  not youtube that makes moocs a current possibility   -- fleshes these out quite a bit but I think they are ideas worth thinking about
01:56 goehle I am definitelly behind the textbook thing
01:57 goehle it would make the "text" much more interactive and much less of a for profit thing
01:57 mgage and his blog -- and much of his videos emphasize the idea that you learn by interacting with your classmates and how (can) you duplicate that when there are 30k students
01:58 goehle its interesting.  I think a big problem is going to keep the quality of interaction high
01:58 mgage yep -- it's silly to think that it's more than an experiment at this point -- but it's a very interesting one
01:59 goehle it is interesting.  No doubt that it will lead to some real fundamental shifts
01:59 goehle I mean, the EE lectures that naomi watched were top notch
01:59 goehle and has some really nice interactivity where he would talk for a few minutes and then give you time to work on a problem and then talk
02:00 goehle I try to replicate it in my classes, but its tougher because we have a firm time constrain of 50 minutes
02:00 goehle I can imagine a world where I am some sort of super TA
02:01 mgage yes -- and even before exT the MIT lectures have been very good -- but as you point out there have been correspondence courses (or sunrise semesters for a long time without dramatic effect)
02:01 goehle well thats the thing, the edX courstes are MIT courses
02:01 mgage yes -- are you familiar with flipped classroom buzzword?
02:01 goehle that means they are great, but they cover in one course what we do in 3
02:01 goehle so its not like we can just point our students there and have them take off
02:01 goehle no, thats a new one
02:02 goehle ah
02:02 mgage -- it's essentially what you described -- students get the lecture on line or from text and you do interactive workshop/TA type things in the classroom itself -- the coursera stuff grew out of flipped classroom experiments at Stanford
02:02 goehle reading up on it
02:03 goehle sure
02:03 goehle its not a bad idea, since you aren't really taking advantage of my ability to quickly troubleshoot student work if I'm giving a lecture
02:04 goehle timeframes are a bit of a problem though
02:04 goehle most of my students spend more time on homework outside of class than they spend in class
02:04 goehle so I could help with some homework
02:04 goehle but certainly not all
02:04 goehle and at WCU it would be tough to keep kids motivated enough to watch their lectures
02:04 mgage yep -- it's hard to test quickly since I believe it involves a change in culture -- students have to get used to the idea of reading/watching ahead of time -- and I confess I'm not always doing that this semester :-)
02:05 goehle I should go back and be s tudent again
02:05 goehle it would probably bring some perspective back
02:05 mgage as I said it's been interesting ---
02:06 goehle I should probably get to bed, I've got class in the morning
02:07 mgage which reminds me -- I'd better get back to my homework -- I need to have a design document for my team for the software engineering class tomorrow. :-)    -- thanks for checking up on the bug -- I'll make sure John Hsu (ionparticle) gets the word that he should check his initializaiton of the site
02:07 mgage ttyl
02:07 goehle good luck with the database problem, let me know if I can help
02:07 mgage will do.
02:07 goehle left #webwork
04:03 Paul_Pearson1 joined #webwork
07:36 djun joined #webwork
12:16 goehle joined #webwork
14:27 Paul_Pearson joined #webwork
14:34 rbeezer joined #webwork
14:45 eee-blt hi
14:45 goehle hi
14:46 eee-blt i wonder if someone could recommend a url for a custom grader
14:46 goehle hmm. I'm not sure I understand the question
14:46 eee-blt i have a problem where students may or may not substitute to reduce
14:47 eee-blt a 2-variable function to a function of 1-variable
14:47 eee-blt i don't want to force them to do this substitution
14:47 eee-blt because it's actually easier to continue with 2 variables
14:47 mgage does this help?  http://webwork.maa.org/wiki/CustomAnswerCheckers
14:48 mgage there is more on answer checkers on the wiki as well
14:49 mgage the custom graders term in the WW world refers to the grading of the entire homework set based on answers to the individual questions
14:49 eee-blt mgage: thanks! i spent 15 minutes googling the website for 'custom graders' and got nowhere
14:50 eee-blt ahh, thanks, i admit i am doubly phased by american terminology around grading
14:50 mgage e.g.  http://webwork.maa.org/wiki/Volume1
14:51 mgage I doubt that it's standard -- WW has developed its own vocabulary to some extent -- there were no standard terms for online homework checkers
14:52 mgage there is also a "fluid grader" which is perhaps the most flexible way to add up the scores for individual questions -- but that is not your current issue
14:53 eee-blt fluid grader means that the subsequent answers are contingent on preceding student answers?
14:53 mgage no -- just that the answers can be weighted differently
14:53 eee-blt i see
14:54 mgage the terms for your concept are   either sequential problem or compound problem -- and these have been around for a while but are under continued development
14:55 mgage http://webwork.maa.org/moodle/mod/forum/search.php?search=compound+problem&id=3
14:55 eee-blt while i have your attention: is there a good reference for ww problem writing? a cookbook, if you will?
14:55 mgage e.g. https://hosted2.webwork.rochester.edu/webwork2/maa102/compoundProblemExperiments/5/
14:56 eee-blt i have found one, and while it is very helpful, it was written in 2002 and hasn't been updated
14:56 mgage the problem techniques section http://webwork.maa.org/wiki/Problem_Techniques
14:57 mgage and  http://webwork.maa.org/wiki/SubjectAreaTemplates
14:57 mgage and in general:  http://webwork.maa.org/wiki/Category:Authors  the POD docs give the more detailed descriptions and are sometimes the only source for the newer or less used features
14:59 mgage were you referring to Tom Shemanske's pdf tutorial?
15:00 eee-blt thanks. i have found increasingly that i am reading the pod docs, but i have wondered how up-to-date they are given ww uses git now and the pod seems to be in an svn repo
15:00 mgage that is a bit old, although still relevant -- where ever possible you should use the MathObjects methods -- they will be more robust
15:01 eee-blt yes, Tom Shemanske's WebWork Newbie Guide - version 1.7
15:01 mgage the POD docs are still up to date -- and I believe that they are reading the git repo through svn
15:02 eee-blt yes, it is old but i was surprised to see that the code essentially worked
15:02 eee-blt it's pretty rare in foss to have such backward compatibility
15:02 mgage it's quite useful for background, if you don't know perl, or if you want to understand WW in depth -- many of those techniques have been upgraded to MathObjects versions (written by Davide Cervone) that hide some of the complexity andy usually provide better student feedback.
15:04 mgage we've done a pretty good job of maintaining backward compatibility -- much of that code was written in perl 4 which didn't directly support object oriented programming -- that said I strongly recommend imitating code such as found in the problem techniques pages -- it's cleaner, easier to maintain and gives better feedback
15:05 mgage the downside of the backward compatibility is that people can copy the old code and often do.
15:05 mgage and while the backward compatibility is good, it's not perfect :-) -- sometimes things have to break to move forward
15:07 eee-blt well, i started out editing problems from the npl, but i found the code to be generally of low quality
15:08 eee-blt so i looked around and found his cookbook, and then i realized that ww has evolved well beyond the state in the cookbook
15:08 eee-blt so now i am casting about trying to find a better guide to the newer features
15:08 mgage we're working on that -- it's a fairly high priority project -- John Jones and Jeff Holt just received an NSF grant to make another pass at it.
15:08 eee-blt the npl?
15:10 mgage There are some in the library labeled with .mo  or MathObjects that are the better models to work with -- the problem techniques and subject area templates were written by Gavin LaRose at Michigan and Paul Pearson at Hope College and are good models and very helpful.  Looking under the school  directories can help -- the Union, FortLewis (where Paul was prior to Hope College) and Michigan directories have a high density of new, high qua
15:10 mgage problems
15:11 mgage yes -- recently renamed opl since it will soon be acquiring questions from outside the US and in languages other than English -- by the way -- which institution are you at?
15:12 eee-blt thanks for the pointers, i had noticed the fortlewis problems tended to be good
15:12 eee-blt i'm at central michigan at the moment
15:14 mgage you mentioned US language I thought you might have been from outside the country.  I'm sporadically working on keeping the localization project on track so that WW can be used more easily in other languages
15:14 mgage check out the Michigan and Union problems as well -- some of the Rochester problems have been rewritten as well
15:15 eee-blt yes, i am canadian and i worked in the uk for 6 years, and education jargon tends to be regionalized
15:15 mgage true enough -- we're getting a lot activity in canada -- quebec is working hard on french localization
15:16 mgage the english cegep are also using it
15:16 eee-blt this is nice to hear
15:17 mgage it's fun for me too -- I grew up on the St. Lawrence river (south side)
15:17 eee-blt whereabouts?
15:17 mgage Canton, NY
15:17 mgage due south of Ottawa
15:18 eee-blt i did my undergrad in ottawa and phd at queen's in kingston
15:19 mgage I'm at U. Rochester now -- my parents are still in Canton so I go back frequently
15:20 mgage here's a collection of "model" problems https://test.webwork.maa.org/webwork2/model_Calculus_1/
15:21 mgage there are more being developed http://webwork.maa.org/wiki/Category:Model_Courses  John Travis at Mississippi college and Jason Aubrey at U. of Missouri are playing a large role in this
15:22 eee-blt thanks again
15:22 mgage np -- need to head in to school.    ttyl
15:22 eee-blt ttyl
15:35 eee-blt i wonder, since ww uses irc, if this link could be added to the wiki: http://irclog.perlgeek.de/webwork/
16:21 Paul_Pearson joined #webwork
17:16 djun joined #webwork
17:33 Paul_Pearson1 joined #webwork
18:40 Paul_Pearson joined #webwork
21:15 djun joined #webwork
22:52 rbeezer joined #webwork
23:46 aubreyja joined #webwork
23:46 aubreyja joined #webwork
23:59 Paul_Pearson joined #webwork

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