WeBWorK Problems

Comments for Instructors

Comments for Instructors

by Nicole Wilson -
Number of replies: 4
I would like to use some type of comment to indicate to instructors (including myself as I am unlikely to recall each question) what the recommended settings are for the Weight, max attempts, show me another, and rerandomize after values.

Currently, I'm putting these in the COMMENT(' '); part of the code.
Unfortunately, these comments are only visible in the library browser and the settings for a question must be chosen in the Hmwk Sets Editor.

I believe everything seen in the Hmwk Sets Editor will be visible to students at some point. Short of embedding this data in the file name is there another way to make these recommendations to instructors that will be visible when they actually set the values?


In reply to Nicole Wilson

Re: Comments for Instructors

by Alex Jordan -
Is the context that you are guiding faculty within your own institution?

Personally I don't think these things belong with the .pg file. In one setting (one instructor teaching one course at one school) there may be "best" values for these things. But the problem may see use in another setting where different values are appropriate. Even within the same course, if you use a problem once in a regular homework set, and use it again in some sort of review, perhaps you would use different weights.

Basically I don't see these things as properties of the problem file, so they shouldn't be coded into the problem file.

Instead, I see these things as properties of that problem in the context of a particular homework set. So they should be written down as properties of that set definition file. To manage this, what we do here is make collections of default homework sets that all courses can access. In a set definition file, you identify that, say problem 12 will use this certain .pg file, will be worth so many points, will have this particular SMA threshold, etc.

If that approach is appealing to you and I can say more about implementing it, I'd be happy to.


In reply to Alex Jordan

Re: Comments for Instructors

by Sean Fitzpatrick -
Yes, this would be for problems in a local question bank. But we may eventually submit them to the OPL, so your point is well taken.
Instead of recommended settings, perhaps it is preferable to include information in the comments such as "not randomized" or "random, 30 versions" and then let instructors decide?

I'm involved with the project and also act as our WeBWorK server admin, and I can help with implementing the approach you suggest. We have a few different courses that we're writing problems for, so what I can do is have faculty participants create template homework sets from the problems we create, and then I'll create a model course for each subject with those set definition files.
In reply to Sean Fitzpatrick

Re: Comments for Instructors

by Alex Jordan -
Instead of model courses, here is an alternative we use.

Note that in your courses' template folder, there is a symlink called "Library" that goes to the OPL files. You can do something similar for a local collection of set definition files. Somewhere in your server, say at:
/opt/webwork/libraries/SetDefinitionFiles/

You could store .def files with some sort of folder and file naming structure that you like. Then you could make all of your courses have a symlink to this folder. When a user would go to import problem sets from set def files, it will search the course's templates folder and go down into subfolders (including symlinks) looking for them. It goes down to a default depth of 4 (I think) looking for .def files.

This way you could manage the communal set def files from one place, instead of propagating copies as courses are created. Just see that the symlink is in a course you intend to clone, and the symlink will be copied over as well.
In reply to Alex Jordan

Re: Comments for Instructors

by Sean Fitzpatrick -
OK, interesting. We'll have to play around with this.
The reason I was considering model courses is I set up WeBWorK courses every semester for the same collection of 5 or 6 courses. In the past I've copied templates from the previous iteration of that course, but this generates a lot of clutter over time.
It seemed to me that from an administration point of view it might be simpler to have a model course for each of the courses that runs regularly on our server, and then every time I create that course for a new semester, I could use the model for that course. This would mean that in addition to problem sets, things like the welcome message would be set in advance. (This is something I end up doing for most people when I create their course.)