Perl 6 - the future is here, just unevenly distributed

IRC log for #webwork, 2013-01-29

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

All times shown according to UTC.

Time Nick Message
00:55 advengers joined #webwork
00:58 advengers Hello. I would like to ask if there is any way I can combine popup questions and open ans_rule using something similar to the MultiAnswer package? Any help is appreciated. Thanks very much.
01:25 aubreyja joined #webwork
01:42 mgage_ hi
01:53 djun Hi Mike
01:54 mgage_ hi
01:55 djun I chatted with Bruce Dunham in stats last week.  He's on board with the idea of hosting a code camp at UBC in the summer.
01:55 mgage_ excellent
01:56 mgage_ the more the merrier.
01:56 djun Should be fun :)  I have to run now, but let's chat later, and maybe I can pick your brains re: getting things organized on my end
01:56 mgage_ and I think with a few substantial pushes we can have a dramatically different webwork for next fall
01:57 djun Yes, looks like some good work already in the pipeline
01:57 mgage_ kk -- we need to brainstorm about which projects to pick -- that will have something to do with a planned roadmap and I know you have ideas about that.  -- then that gets modified by who can come and what they are passionate about at the moment.
01:57 djun sounds good
01:57 mgage_ tyl
01:57 mgage_ tttyl
01:58 djun bye for now :)
03:51 aubreyja joined #webwork
04:21 rbeezer joined #webwork
13:27 mgage_ joined #webwork
13:49 goehle joined #webwork
14:54 dpvc joined #webwork
15:00 mgage_ dpvc: hi
15:00 dpvc Good morning.
15:01 mgage_ we had this question yesterday -- I think the answer is no, but I wanted your opinion:   Hello. I would like to ask if there is any way I can combine popup questions and open ans_rule using something similar to the MultiAnswer package? Any help is appreciated. Thanks very much.
15:01 dpvc mgage_: or should I do it this way (with your userid)?
15:01 mgage_ either one works -- I have my irc set so that it pings whenever anyone posts
15:02 mgage_ I'm also logged in both here at home and in the office -- that's why the tail
15:02 dpvc I'm not sure I understand the question.  Can you be more specific about what you want to do?
15:03 dpvc What is in "open ans_rule"?
15:03 mgage_ it was a question in the irc yester day from  advengers  -- I believe he wants to place an ans_rule in a popup menu
15:04 dpvc Huh?  He wants to have students enter answers in a text area within a popup menu?  That doesn't' make sense to me.
15:05 dpvc Or do you mean a pop up window of some kind?
15:05 dpvc (Popups current mean menus)
15:05 dpvc (I mean "currently" not "current")
15:05 mgage_ ok -- I didn't really understand the post either -- so I didn't try to handle it -- I guess we can wait until he drops back in
15:06 dpvc I'll look in the log and see if I can make sense of it.  But if that is really what he is after, then the answer is "no", you can't have other input elements embedded inside a pop-up menu in the current implementation.
15:07 dpvc We currently use HTML <select> elements, and those can't contain other input elements.
15:08 mgage_ I gave you the complete quote from the log so I don't think you have to check further.  And that confirms what I thought that you can't imbed elements in pop-upmenus
15:10 dpvc I'm wondering if he simply meant could you have a MultiAnswer that includes popup menus and other answer types.  In that case, the answer is "yes"
15:10 dpvc You need to use the parserPopUp.pl file, which implements MathObject popup menus.
15:11 dpvc They are not as sophisticated as the traditional ones, but may be sufficient for his needs.
15:12 mgage_ good to know.  unfortunately I don't know advengers other identities
15:12 mgage_ on other news -- I just updated test.webwork  to ww2.5.1.3 and it is running smoothly
15:13 dpvc That is good.
15:13 dpvc I'm still trying hard to catch up on the several months of emails that I have been unable to get to.  Over 470 in my inbox.  Yikes!
15:13 mgage_ I'm trying to cut down on the number of branches so webwork2/webwork2-dev has a devel branch.  All developers should pull from that and return features to that via a pull request  -- this is pretty much your suggestion.
15:14 dpvc OK, sounds good.  I'm still hoping to write up a more concrete proposal.
15:15 mgage_ I think this will work well for the interface between the Integration group and the developers --- I think more will have to be done to get sufficient testing of the integrated version(s) -- but that can wait -- particularly since I don't yet have any ideas I really like
15:15 mgage_ headed for school
15:15 mgage_ ttyl
15:15 dpvc I think one of the difficulties with the approach that has been in use is that you can't tell if a branch is stable or not.
15:15 dpvc OK, see you!
15:16 mgage_ that kind of communication is certainly a problem.    -- the other problem is that there are degrees of stability -- and that is even harder to communicate
15:16 dpvc Right!
15:22 aubreyja joined #webwork
15:56 mgage at work
16:23 aubreyja joined #webwork
16:23 aubreyja_ joined #webwork
16:51 rbeezer joined #webwork
17:29 mgage aubreyja: hi
17:30 mgage aubreyja_:
17:58 djun joined #webwork
19:01 dpvc joined #webwork
19:55 goehle hey
19:55 mgage hi geoff
19:56 goehle I have a pg question
19:56 goehle also I was wondering if you finished 1.5.3
19:56 goehle I know you wanted me to look at it
19:57 dpvc What is the PG question
19:57 goehle as for my pg question, is there a way to check and see if a set of parametric equations is correct independent of the parameterization
19:57 goehle one of my problems is only recording correct answers for a particular parameterization
19:57 djun joined #webwork
19:57 dpvc Only for special cases, like lines.
19:58 aubreyja Hi all
19:58 dpvc Or if you have special restrictions on the form of the parameterization
19:58 goehle this is a line !
19:58 goehle in 3d
19:58 mgage ww2.5.1.3 is ready -- you can pull it from openwebwork/webwork2-dev
19:58 goehle great
19:58 goehle I'll take a look at it (probably over the weekend)
19:58 dpvc Then yes, you can use parserParametricLine.pl for that.
19:58 aubreyja mgage, mgage_ - any clue about Arnie's bug? I'm pretty certain that webwork2-dev and webwork2 are the same
19:58 mgage you can also just look at it on hosted2   the maa10x courses
19:58 aubreyja brb
19:59 dpvc Note, however that this does expect the line to be linear in its parameter.
19:59 mgage no -- my first bet is that there is a .conf file that isn't synced with the .conf.dist file
19:59 dpvc If you want to allow really arbitrary parameterizations, that is hard.
19:59 goehle that should be enough.  Mostly people are picking the points in the opposite direction and getting the negative parameterization
20:00 dpvc Then parserParametricLine.pl should do the job for you.  See http://webwork.maa.org/pod/pg_TRUNK/macros/parserParametricLine.pl.html for details.
20:00 goehle hmm
20:01 goehle so this is using mathobjects
20:01 dpvc Yes.
20:01 goehle and uses the parametric line mathobject already in the problem
20:01 goehle the issue is that its asking for the x, y, z formulas as 3 separate answers
20:01 goehle Library/UMN/calculusStewartET/s_12_5_8.pg
20:02 dpvc Do you want to keep it three separate formulas?
20:02 goehle I guess thats the question.  The problem already asks for the vector form
20:03 dpvc Then it should not be a hard conversion to MathObjects.
20:03 goehle and then it asks for three formulas, to double check moving from vector form to non-vector parametric form
20:04 dpvc In that case, you could use a MultiAnswer object with the three parts, and then put them together in the checker to form the parametric line object and use that to check the answer.
20:04 dpvc It sounds harder than it is (I think).
20:05 dpvc I don't see any other natural way to do it, since the check requires knowledge of all three parts to see if it is a valid parameterization of the line but different from the professor's answer.
20:05 goehle yeah
20:05 goehle exactly
20:06 goehle so I need a way to get all three parts,
20:06 goehle assemble them back into a vector parametirc line
20:06 goehle and then compare
20:06 dpvc That is what the MultiAnswer object provides.
20:07 dpvc It gives you access to multiple answer blanks at once.
20:07 goehle I'll give it a shot
20:07 goehle thanks!
20:08 dpvc You would get the three student answers, and form a Vector() from them (which would return a vector Formula) and compare that to the correct ParamatricLine object.
20:08 dpvc Something like this (off the top of my head):
20:09 dpvc $ma = MultiAnswer("1+4t","3-2t","5t")->with(
20:09 dpvc checker => sub {
20:09 dpvc my ($correct,$student) = @_;
20:10 dpvc my $student_line = Vector(@$student);
20:10 goehle now I need to compare $student_line to the ParametricLine that was passed in
20:10 goehle but its not cmp right?
20:11 goehle or is it?  Its not clear to me
20:11 dpvc my $correct_line = ParametricLine(Vector(@$correct)))
20:11 dpvc return [1,1,1] if $correct_line == $student_line;
20:11 dpvc return [0,0,0];
20:11 dpvc }
20:11 dpvc );
20:12 goehle ok, so $corect_line == $student_line will check for equality of the underlying objects?
20:12 dpvc Yes.
20:12 dpvc It is exactly the same check that would be performed for checking a student answer to a ParamatricLine (including all the fuzzy tolerances).
20:13 dpvc Then you use $ma->ans_rule (three times) to get the answer rules for the three components, and finally
20:13 dpvc use ANS($ma->cmp) to get the answer checker at the end.
20:14 dpvc I'm pretty sure that will work.  The ParametricLine(Vector(@$correct)) is the only part that might need to be adjusted, but I'm almost certain that will be OK.
20:15 goehle lets try it out then (i;'ve been implementing so far)
20:16 goehle hey looks like it works.   Thanks.  I wouldn't have known about the operator overloading of ==
20:16 goehle or multianswer
20:17 dpvc These are two of the advantages of MathObjects over the older techniques.
20:17 dpvc All the operators are overloaded, so you can add formulas, for example.
20:18 dpvc The overloaded == makes it easy to write custom checkers that use the same tests as the standard answer checkers themselves
20:18 goehle hmm
20:18 goehle seems to be marking everything as correct though
20:18 dpvc (rather than having to do things like $f->cmp->evaluate($student_ans) like in the old days).
20:19 dpvc OOPS, try using (1,1,1) and (0,0,0 rather than [1,1,1] and [0,0,0].
20:19 dpvc That is, "(0,0,0)"
20:22 goehle hmm, now I'm back to the original behavior.  It will accept one particular parameterization, but not with a -t instead
20:22 dpvc OK, let me check it out...
20:23 mgage by the way Davide, I don't think anyone would mind if you updated the example http://webwork.maa.org/wiki/MultiAnswerProblems  -- what were you going to use to compare answer types  (rather than ref())
20:23 mgage -- when you get a chacne
20:23 mgage chance
20:23 dpvc the classMatch() method.
20:24 mgage sounds good
20:24 dpvc or perhaps typeMatch(*).
20:24 dpvc (oops, no star intended)
20:24 mgage I suspect they hadn't noticed the existence of those features
20:24 mgage it would be good to have them on display on the wiki
20:25 mgage students incoming :-)
20:25 mgage ttyl
20:25 dpvc They are in the new MathObject documentation.
20:26 goehle changing my $student_line = Vector(@$student);
20:27 goehle to my $student_line = ParametricLine(Vector(@$student)) seemed to do the trick
20:27 dpvc OK, great.
20:27 goehle nice
20:27 goehle good to know
20:28 dpvc The object types all have a ranking, and they get promoted up the ranks if needed, but I guess ParametricLine isn't above Formula.
20:29 dpvc I thought it was, but it seems not.  So forcing both to be the same avoids the problem of promotion.
20:29 goehle ah, so it was using vector formula to compare still
20:30 dpvc Right.
20:56 djun joined #webwork
22:20 rbeezer_ joined #webwork
22:52 djun joined #webwork
23:01 aubreyja joined #webwork
23:02 djun_ joined #webwork
23:46 goehle left #webwork

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