Misleading instructions for set builder notation

[dlglin]

This is related to https://github.com/openwebwork/mathquill/issues/5

There are a number of PCC problems that ask for set builder notation, and include the instruction "The | symbol is above the keyboard's Enter key." (e.g. Library/PCC/BasicAlgebra/LinearInequalities/GraphToSetAndIntervalNotation10.pg).

If MathQuill is enabled then typing the pipe character on the keyboard will default to absolute value and autocomplete a closing magnitude bar. This gets marked wrong, so these instructions will lead students astray.

There are a few possible remedies:

• Update MathQuill so that the pipe key on the keyboard can be used for vert. I suspect that this isn't easy, as in most cases the desired behaviour for pipe is absolute value. Can it do both, or would it need to default to one or the other? Can that default be changed on a question-by-question basis?
• Disable MathQuill for these questions. That also seems less than ideal since students might need help with other symbols (e.g. \geq).
• Change the instructions. This is also problematic as the correct instructions depend on whether MathQuill is enabled. If MQ is active then students need to either type vert or use the toolbar. If MQ is disabled then they can simply type |. Maybe just delete that line altogether? Can there be a conditional in the code that checks if MQ is enabled and displays the appropriate instructions? Is that a bad idea?

[dpvc]

Maybe the solution would be to use the colon notation rather than the vertical bar. I think the set-builder macros allow for either one (and if not, it would be an easy fix).

[gajennings]

The colon idea works here, but I recall seeing other cases where MathQuill misinterprets students' input, so it might be better to show MathQuill users how to fix that in every case.
For example one could add the line "MathQuill users: if MathQuill misinterprets your input then use MathQuill's "Tt" button to enable "plain text" mode." to the various helpLink files that assist with data entry.

[dlglin]

Tagging @Alex-Jordan since this arose from PCC problems.

I'm of two minds with respect to promoting the plain text mode in MathQuill. While it allows for backward compatibility with older problems, it's also a kludge to tell students to use it. I believe that it is already possible to disable MathQuill from within a problem, so problems for which it doesn't make sense to use an equation editor (e.g. where a student is expected to type text) should have MQ disabled. In the long run problems where MQ is misinterpreting input should be looked at to see if the problem and/or MQ needs to be fixed.

With that being said, this may be a losing battle. In fact, this could be the case that proves that. Is there always a right thing to do with MathQuill? Is it feasible to code all of these cases? If not, we may be back to telling students to use text mode.

The other thing to be careful about with instructions is that students likely don't know what MathQuill is, and I don't like the idea of students having to figure out whether or not they're using MathQuill. We should be trying to word this in a way that makes sense to a student who doesn't know any of the WeBWorK jargon.

[Alex-Jordan]

My local colleagues would not accept using a colon instead of a vertical bar. Any tie there is a compromise with the notation you type compared with the notation they want to promote in writing, it is a problem.

We have been updating these exercises to disable MathQuill. But that is with our versions that we maintain, which is not the same as the OPL versions.

[pstaabp]

Perhaps we could build in a MathQuillHints or MathQuillOps flag that could pass information to mathquill. This could include disable, as well as vertical_bar_is_abs to false to not complete the right |. And then other features can be added.

Then in a cmp call, these options could be passed, like

cmp(MathQuillOpts => {vertical_bar_is_abs => 0})