Add scriptq question type

[seragunn]

The goal here is to add questions requiring some computation to verify.

[jverzani]

Cool. So you pass in an expression and not a function? Is there some reason that is preferable? I would have guessed the function might have f(this.val, ...) signature so that things are run in a different scope. (But I haven't really thought any of this through.)

Also, is it possible to use external javascript libraries in the evaluation of correct or not? I don't have any in mind, but I could see that being of interest.

[seragunn]

Also, is it possible to use external javascript libraries in the evaluation of correct or not? I don't have any in mind, but I could see that being of interest.

Yes! In the screenshot, the Permutation class is defined in a separate file.

So you pass in an expression and not a function? Is there some reason that is preferable?

The question is: how much of the JavaScript code should be written in your markdown/Julia versus in an external JavaScript file. If we're writing the question in Julia then you'd have to wrap the expressions in (str) => { ... } and that function would only be used for the contents. But if you write a closure in a separate JavaScript file then you could pass in closure(...). So it kind of forces you to keep the JavaScript in a separate file...maybe that's preferable.

[seragunn]

...Going to make a new branch doing it via functions for comparison

[jverzani]

Let me know what works out best from your point of view.

[seragunn]

I like the implementation in #49 better:

• it encourages you to keep most of your JavaScript code in JavaScript rather than within a Julia string.
• you don't need to have a specifically named correct: boolean variable at the end, just any function which returns a boolean. So a cleaner interface.
• for some short functions it can even be less verbose doing (i) => 43 <= i && i <= 48 rather than var correct = 43 <= this.value && this.value <= 48;.

I suggest reviewing #49 and closing this PR.