# IRC log for #webwork, 2013-07-01

All times shown according to UTC.

Time Nick Message
04:35 rbeezer joined #webwork
04:39 mgage joined #webwork
05:36 mgage joined #webwork
14:09 mgage joined #webwork
14:32 aubreyja joined #webwork
15:11 mgage joined #webwork
15:19 goehle joined #webwork
16:34 mgage joined #webwork
17:16 goehle hey mgage
17:17 mgage hi
17:17 goehle how was the code camp?
17:17 goehle I didn't have time to read all of the emails thouroughly.  So anything I am supposed to look at that I haven't already responded to
17:18 mgage it's gone well.  we still have a few people in town.  I'm out at the ubc campus with travis and djun -- they are giving a video class which is the first session of the modelCourse class to be held next week.
17:18 goehle nice
17:19 mgage look through them again.  I pulled a fix which I think takes care of the complete invisibility of hidden courses on the homework set list in the math3 theme.  I haven't checked the fix closely but I think you'll find that fix in devel.
17:19 mgage develop
17:19 mgage branch
17:20 goehle I submitted a fix for that
17:20 mgage yes -- and I pulled it -- just haven't had a chance to test it yet.  I'll do that this morning while they are running their conference.
17:21 mgage you can pull the latest develop branch and use that for a while.
17:21 goehle it should be fine.  There was the same issue on homeworkseteditor2 that I fixed a while bake
17:21 goehle back
17:21 goehle I didn't think to check the first homework sets editor
17:21 goehle what do you want to do about mathview?
17:21 mgage right.
17:23 mgage That's something to work on.  I looked around a bit to see what was clobbering it but didn't see anything right away and had other duties.  If you can fix that in the develop branch that would be great.  My first guess is that it has to do with the various ways in which jquery is called.
17:23 goehle I can poke at it
17:23 goehle shoudl I mess with the mathview.js code
17:23 goehle or try to layer something onto the math4 theme?
17:23 mgage that would be great can you see the problem?
17:24 goehle hmmm
17:24 goehle well as far as the visibility of the icon, I did figure that out
17:24 goehle the mathview code does some wierd stuff with putting an image inside a width:0 div
17:24 goehle but a little css tweak makes the icon visible
17:24 mgage You can do either -- what ever is cleaner.  MathView has been handed to us -- the guy who customized it for us is patrick Belly and he said he wouldn't have time to work on it anymore but could answer questions.
17:24 goehle unfortunately it all looks kind of off
17:25 goehle I sent you a screenshot
17:25 mgage I think it's ours to mess with if we wish -- so go ahead.    Is he using zero width to hide the icon?
17:26 goehle yeah
17:26 goehle well
17:26 goehle sort of
17:26 goehle the icons visibility is also set to hidden
17:26 goehle i'm not entirely clear what that div is even there
17:26 goehle but it might be used for something
17:26 goehle if I mess with the mathview code do you want me to fork it
17:26 goehle or should I make something that works in both math3 and math4
17:26 goehle ?
17:28 mgage I think we can create our own version.  I don't think it matters whether we have a separate repo for it on openwebwork -- you can do what you think best
17:28 goehle ok
17:28 mgage if possible it should work on both math3 and math4    --
17:29 goehle right now its pretty messy, so I'll treat the fact that the icon is broken as a "feature" until I get it nicer
17:29 mgage I don't think it was completely done in any case, and some of the css might have been a little weird.
17:30 mgage there is a feature on the back end that works for now but we should look at later.  It is using the AlgParser module to read student answers and translate them to tex but it might be better to switch that over to using mathjax.
17:30 goehle does that do the same thing?
17:30 goehle I thought this was so that students could build equations that pg can read
17:31 mgage similar -- the algParser was done by someone else and was more primitive -- it did do some highlighting of errors.
17:31 goehle yeah
17:32 mgage it's used to take what students type into the answer blank and add that to the tex which gives the formatted equation.
17:32 goehle ?
17:32 mgage david gage is around online as well if you need comments about javaScript
17:33 mgage there is an entry blank at the bottom of the widget where you can type and have what you have entered formatted
17:33 mgage you type in using calculator notation
17:33 goehle ah
17:33 goehle is [ a, b [
17:34 goehle standard notation?
17:34 goehle its one of the options in mathview
17:34 goehle I think its supposed to be [ a , b )
17:34 mgage that's for later -- it works well enough now -- Patrick couldn't get MathObjects to work for some reason
17:35 mgage actually   [a, b [ is a common alternate through out europe, or at least in France.  Patrick is from Quebec
17:35 goehle ah
17:35 goehle is there a reason that a lot of times the formula has a
17:35 goehle *
17:35 goehle in front?
17:36 goehle like when I type log
17:36 goehle it comes up *log
17:36 goehle when I type log(2)
17:36 goehle I get
17:36 mgage no -- not that I know of -- that (or a centered dot) is needed between numbers
17:36 goehle log((2))
17:36 goehle so no star but extra ()
17:37 mgage we noticed the last one as well -- I thought the extra star an indication that more needed to be added -- but see what you think.
17:37 goehle ok
17:39 mgage I just forwarded the last message I got from Patrick.  Once you get things figured out a bit you can email him and see if he wants to comment.  There might have been a reason for some of his choices.
17:39 goehle thats a good idea
17:39 goehle I need to play around with it a bit
17:40 mgage go for it.  I think it's a really good start for those that want a math editor, but it's just the first proof of concept model.
17:40 goehle yeah, it seems a little rough in places
17:41 goehle its not clear to me what I"m going to do about personal preference type issues
17:41 goehle like [ a , b [
17:41 goehle or e^a creating exp()
17:41 goehle (my students dont know exp, for example)
17:41 goehle or arcsin
17:41 goehle vs
17:41 goehle sin^{-1}
17:43 mgage also send a note to david gage that you are working on it -- he has some ideas about js widgets (javaScript components) that might influence your work
17:43 goehle ok
17:44 goehle I was also thinking about putting some of the bootstrap stuff in there
17:44 goehle it will make it look more consistent with the rest of the  page
17:44 mgage I think the basis of some of this is from jquery.  Then the customization is done in operations.js
17:45 goehle thats good to know
17:45 goehle I hadn't looked at operations.js
17:45 mgage we don't need to do this on the first pass, but we should create a couple of different common notations groups that people can choose from
17:45 goehle right
17:46 goehle ah
17:46 goehle ok
17:46 goehle operations.js
17:46 goehle has the meat of the actual elements available
17:47 goehle then mathview.js (and jquery) are used to make the editor window and assemble everything
17:47 mgage workflow for fixing this feature:    set a branch in your git on your computer that tracks develop:
17:47 mgage git branch -t develop_mathview   origin/develop
17:47 goehle have I not been doing this?
17:48 mgage (assuming your origin points to http......
17:48 mgage then checkout develop_mathview   ….. do work   and
17:48 mgage git push https://github.comgoehle/webwork2    develop_mathview
17:48 mgage then send a pull request from develop_mathview
17:49 mgage I think you've been doing things pretty close to this -- I'm just reviewing since we still don't have a document where this is all written down.
17:49 goehle ah ok
17:49 goehle i'm not sure I always have my local branch track develop
17:50 goehle but I do always pull develop into the local branch before pushing it to my github and doing the pull request
17:50 mgage The main useful part of this is locally developing on a new feature branch  AND creating this feature branch in your github.   then the commits for this feature to develop are sorted independently from other feature work that you submit and it's possible to accept one feature but not another.
17:50 goehle now, I shouldn't do the pull request until I have things relatively stable
17:50 goehle right
17:51 goehle right now I have a mathpets branch, I'll have this branch, and then I have a misc branch where I do small fixes
17:51 goehle or should all of those small fixes get their own branch
17:52 mgage I've been doing the small fixes on the develop branch -- or at least merging them into develop and then issuing a pull request from my develop to openwebwork develop
17:52 mgage I don't know if that is best or not
17:52 goehle creating lots of small branches is a pain
17:52 goehle especially for one or two commits
17:54 mgage for the features -- if you submit  from a math_pets branch and from a mathview branch then they show up separately in the pull request queue.  So people can look at them, pull your mathpet feature and play with it, make suggestions, you can make fixes, and none of this interferes with your mathview pull request
17:55 mgage so this means you can submit when things are working and you want comments and feedback.
17:55 goehle I'll be sure to keep those separate then
17:55 mgage that's the idea.  we're still working on it.
17:55 goehle I've been trying to do this.  My validation work was on a different branch than my achievement_items work
17:56 goehle and I think math4 used to be a third branch
17:56 goehle although its been merged into everything by now
17:56 mgage it also looks like at one point you merged the master branch changes into your develop -- I'm not sure, but I think that is something we should try to avoid -- I think it's one of the things that causes annoying conflicts when you merge release branches which branched from develop back into master.
17:57 mgage yep -- I think you've actually been doing this pretty well. I'm just reviewing.
17:57 goehle hmmm
17:57 goehle shouldn't we be periodically merging master and release/2.8 into develop
17:57 goehle to keep any bug fixes that happen in those two branches in develop
17:58 goehle that was what I was doing there
17:58 goehle I fixed something in master
17:58 goehle and then (instead of making the fix twice) I pulled master into develop
17:58 mgage I've also been trying to delete a feature branch once it is merged into develop -- that way there aren't two versions that need to be synced.  I'm going to delete release/2.7 on openwebwork but haven't quite had the courage to do it yet. :-)
17:59 goehle once they have been merged they kind of fall into the background as long as all of the other branches are ahead of htem
17:59 goehle so keeping it around isn't the worst thing
17:59 goehle just dont accept any pull requests for it
17:59 mgage we should periodically merge release/2.8 into develop -- although I'm hoping to keep the life times of these release branches much shorter so that won't be that necessary.
18:00 goehle I think that testing is always going to be a bit drawn out
18:01 mgage for master - if there is something that really needs fixing --- we create a hot fix branch from master --- fix things --- and then merge that branch back into both master and to either release/2.8 or to develop. (If release/2.8 is around merge to release/2.8 and then merge that do develop)
18:01 mgage I'm not so sure.  I'm going to try to make it go much faster. we'll see.
18:02 goehle ah ok
18:02 goehle good to know
18:02 mgage develop is actually pretty stable at the moment -- I'm using it --- some new features are a little weird but the old ones aren't broken.
18:02 goehle yeah
18:03 goehle Ive actually got develop on our new server since I want to use math items and I like the latest iteration of math4
18:03 mgage and I'm holding off merging really big changes into develop until I'm sure they are fairly stable.
18:03 goehle that seems like the right thing
18:05 goehle btw, someone from our graphics design department
18:05 goehle is going to use mathpets as the main class project over the next year
18:05 goehle so hopefully that will produce lots of good quality graphics that we can use
18:07 mgage cool
18:12 mgage djun said that we should look at mathdox.org based in belgium -- it's possible that there is a javaScript based equation editor that is even better than MathView
18:14 goehle so what is mathview supposed to be fore
18:14 goehle for
18:14 goehle I'm just trying to get the goals straight
18:17 mgage there are many that ask for an equation editor to enter equations.  -- I think it's from two sources -- those teaching precalculus and below have students who have trouble with calculator syntax entry of math .  my experience is that students quickly abandon the editors for most uses and prefer to type directly since it is faster.
18:19 goehle so we might not want something full featured
18:19 goehle if this is for beginning students
18:19 goehle it just needs to do fractions and functions and the basic stuff
18:19 mgage The second is that instructors ask for this as a feature.  Even though eventually they don't end up using it very much.  In both cases I think it's more a perceived need than a real one  but providing it means we don't have to argue with potential adopters about why it's not really needed and there are few cases where it really is beneficial.
18:19 goehle http://stackoverflow.com/questions/7433540/embeddable-wysiwyg-equation-editor
18:20 goehle should we be trying to satisfy both goals with the same thing?
18:20 mgage I think we want something moderately full featured -- but still simple enough for precalculus students to use.
18:20 mgage I suspect that we can write something common with customizable pluggins
18:20 goehle I guess I dont understand what the instructors want it for
18:21 mgage well -- it is something that is constantly asked for by people who first start using webwork
18:21 goehle but what do they want to do with it
18:22 goehle just create "calculator format" equations using a wysiwyg editor?
18:22 mgage bill wheeler implemented a version back in the late 1990's using design sciences's MathEQ -- and eventually took it out because it wasn't used much.
18:22 goehle interesting
18:22 mgage yes -- use wysiwyg to create input to webwork
18:22 goehle the site i linked to above said there is a mathview based editor in the works
18:22 goehle is that true?
18:22 goehle mathjax
18:22 goehle rather
18:24 mgage I don't think they care how the equation makes it into webwork precisely, but going through calculator syntax is good enough.  I think math view does use mathJax -- let me look at the link
18:25 goehle hmm.  I can certain work on making mathview prettier.  If I was to go shopping for a new editor or really redo a lot of mathview I would need some user stories.  I've never felt a need for this kind of editor and havent found that my students need it either
18:26 mgage Davide has a prototype editor using mathJax -- I think you needed to enter latex symbols for his version yet.  I don't know what his plans are.
18:28 mgage I kind of agree with you on this.  For the moment let's just 1. make it look a bit prettier, 2. make it work properly with the themes 3. if it's not too much trouble make it customizable in some way so that we can have both the french conventions and the english conventions for symbols (check with Davide about this -- that might already be built in to mathJax)
18:28 goehle ok
18:28 goehle that I can work with
18:29 mgage That's plenty -- and the last doesn't have to be easy to change -- it could just be a case of having some alternate operations.js files
18:30 goehle I could always make it so that the operations.js file is an easy to change string
18:30 goehle (maybe even something that webwork provides)
18:30 goehle then you oculd have
18:30 goehle operations-en.js
18:30 goehle operations-us.js
18:30 goehle or whatnot
18:33 mgage yes
18:59 goehle hey mgage
18:59 goehle quick question about intended behavior
19:00 mgage kk
19:01 goehle this is for pgEditor3
19:01 mgage k
19:01 goehle when you click new version (and any of the options) and then take action
19:01 goehle it opens a new editor window in a modal
19:01 goehle shouldnt it just go to that webpage (no modal, no new window)?
19:02 goehle on the other hand append
19:02 goehle opens a rendered version of the problem in a modal
19:02 mgage yes -- I think that would be good -- you would get back to the editor with the edit button.  -- the modal is good for View.
19:03 goehle or should that open a new editor
19:03 goehle ?
19:04 goehle I feel like append should just open a new editor
19:04 goehle and things that open a new editor
19:04 goehle should open them in the same window (no modal)
19:04 mgage I think append should become attached to the new file (it doesn't currently) -- I think it should open an editor window -- but one might have different opinions about that.
19:04 goehle ah
19:04 goehle ok
19:04 goehle so append doesn't become a new file
19:05 mgage yes that seems ok to me -- the biggest thing is to get attached to the new file
19:05 goehle so in some sense opening a view modal makes sense
19:05 mgage append does make a new file, but the editor stays focused on the old one
19:05 goehle but thats what new version does
19:05 goehle you can use new version to create a new file and append
19:06 goehle actually append doesnt make a whole lot of sense to me
19:06 goehle why would you use that option if it doesnt create a new file
19:06 goehle unless you want multiple copies of a problem in a homework
19:08 mgage I don't think it's been used much.  One use would be simply to add this problem file to another homework set.  -- not sure it's needed.  (my bet is that you could hide it and then see who screams for it. :-) )
19:12 goehle Eh, I guess there is no reason to step on anyones work flow
19:13 goehle although it is a bit confusing
19:14 goehle I'll hide it in pgeditor3
19:14 goehle and see how things work
19:14 goehle as a test run
19:14 mgage I'm fine with that.
19:15 mgage it's about the only way to find out if it's important to someone
19:17 goehle yeah
19:17 goehle and its always available in pgeditor2
19:18 goehle if you really care
19:21 hij joined #webwork
19:24 hij Hello, I have a question about references. I have a macro file that I am writing. Part of what it does is take data sets which are pointers to arrays and keeps track of them. I use the following subroutine:
19:24 hij sub push_stat_data_set {
19:24 hij my $data = shift; 19:24 hij push(@dataSets,$data);
19:24 hij }
19:24 hij When I call it using this:
19:24 hij $pointer = \@data1; 19:24 hij push_stat_data_set($pointer);
19:24 hij It works, but when I use this:
19:24 hij push_stat_data_set(\@data1);
19:24 hij It does not work. It stores a ref that is different from \@data1.
19:25 hij Am I doing something stupid or have I just lost control of my mind?
19:26 mgage those should both be equivalent. I'm assuming you are not in a .pg file (there you have to use ~~ instead of \ )
19:27 hij This is in a perl file that is in the macros directory. I am saving it using the file manager in webwork.
19:27 mgage kk -- it's pure perl then.
19:27 hij Sorry - the subroutine is in the perl file that is in the macro subdirectory.
19:27 hij The calls to it are from a pg file.
19:28 mgage that's not a problem
19:28 hij I am not familiar with the "~~" syntax?
19:29 mgage those constructions should be the same.   hidden character could be a problem, or it could be a genuine bug in that version of perl -- have you just tried this with a file from the command line?
19:29 hij Yes, I have tried it from teh command line, and it was fine.
19:31 mgage \ is reserved by tex in the .pg files   basically all of those are escaped when the problem is rendered -- if you really want to use a backslash, say to dereference an array, then you need to use ~~ which gets turned in to \ and passed to perl.
19:31 * hij just found http://webwork.maa.org/wiki/Basic_Perl_syntax#.UdHZAHWfHds
19:31 * hij nods
19:32 goehle hey mgage.
19:32 mgage I don't have an explanation of why it wouldn't work fine in a .pg file  -- you can try ~~ to see what happens
19:32 mgage hi
19:32 hij ugh -- That seems to have fixed it. I am an idiot.
19:32 goehle is it an issue that Revert is basically useless in PGEditor3
19:32 mgage np
19:33 goehle since the main editor page doesnt get updated
19:33 hij Thank you!
19:33 goehle it never really appears
19:34 mgage yes -- I was going to go look into that a bit more -- the user needs some feedback and should be returned to the editor page with the original contents.
19:34 mgage main editor page
19:36 goehle right
19:36 goehle since the main editor page is never reloaded it could be misleading whats going on
19:36 goehle on the other hand it would be tough to refresh the main editor page
19:36 goehle based off of the results of the save
19:38 mgage can you just restart the whole editor?  you do need to pass the success message back in, but since the .tmp file has been erased the editor will grab the original file.
19:38 mgage the point of revert is to stop using the .tmp file and erase it -- and return to the original file.
19:39 goehle right
19:39 goehle so revert shoudlnt open up in a modal either
19:39 goehle ok
19:39 goehle actually this might be oke then
19:39 mgage (by the way -- all of this can probably be done more simply now -- e.g. using local storage and things like that -- this was hacked for an earlier version of the web. )
19:40 goehle so (assuming the new window option isnt checked) new version will replace the current editor window
19:40 goehle which may now have a revert option
19:40 goehle and once I change revert, using that option will again replace the main editor window
19:40 mgage right -- revert should not open in modal -- it's a toss up whether it should open in an editor or in the view of the original problem.
19:40 goehle I think it should open an editor
19:40 goehle since the idea is you are reverting the pg code
19:40 mgage kk
19:41 goehle but thats just me
19:41 mgage let's go with that -- it's closer to what I implemented for editor1
19:45 goehle wait now I'm confused
19:45 goehle when does revert appear?
19:46 mgage it appears whenever a .tmp file exists.    Those files are created when you do a View (that allows you to make temporary changes and view them -- only the user who creates the tmp problem can view it. )
19:47 mgage I think the rules are that they go away whenever you do an update, or new version
19:47 mgage if my memory is correct the button just checks the existence of a .tmp file
19:48 mgage using local storage instead of the .tmp file would make everything much simpler, but that is for later
19:51 goehle hmm
19:51 goehle ok so it is still a little broken then
19:51 goehle since viewing a temp file
19:51 goehle since view doesnt refresh the main editor window
19:51 goehle so the revert button wont appear
19:53 mgage yes -- it's broken.  I don't have an immediate solution.
19:53 goehle hmm
19:53 goehle yeah thats a bit tough
19:53 goehle although you had the same issue if you used "Open in new window"
19:53 goehle so its not new
19:54 mgage right -- you didn't get revert until you pressed "edit problem" in the new window
19:55 mgage we could trigger an xmlHTTPrequest event that checks for the existence of a .tmp file, shows the revert button  we could trigger that event not just when the page loads but also when the modal window closes.
20:01 goehle and then that would unhide the revert option if there is a temp file?
20:17 mgage yes -- that's the idea
20:51 goehle hey mgage
21:24 hij left #webwork
21:27 goehle left #webwork