Don't allow 0th root about mathsteps

[hmaurer]

Oh yes of course; I'll check out some of the other issues.

[adamisntdead]

I am happy to pick this one up, if thats okay?

[hmaurer]

I am on this. Could you assign me to it? It's easier to filter :)

[evykassirer]

Actually I'd appreciate if you could work on a maximum of 1 small PR for now, since I want to make it easy for people to start on this project and it'll take a bit of time to find more small tasks for folks.

Would it be alright to leave this for another eager newcomer? I'm happy to discuss some other slightly bigger PRs with you if you want to find something else to work on :)

And, of course, thanks so much for your interest. It is so great to have someone else so on board 😀 😀

Also, I can't assign things to people unless they're on the collaborators list (have read and write access, which will probably stay Socratic people for now) - sorry, I know that's inconvenient. We might be able to find another system to help keep track of PRs for people - let me know if you have ideas!

[adamisntdead]

Just looking now, the nthRoot Code seems to be on the math.js side, so changing it would be changing the actual mathjs code. Is that correct, or am I missing something?

[hmaurer]

@adamisntdead Check out #29.

The operation nthRoot with base 0 would lead to an exception when trying to evaluate it with mathjs. For example, try this (at the root of the project after cloning the repository):

const mathsteps = require('./');
mathsteps.simplifyExpression('nthRoot(5, 0)');

In attempting to simplify nthRoot(5, 0) the program is going to evaluate it, and it will throw an error in our face: Error: Root must be non-zero. That's not great; what would be better instead would be to return a step with changeType NTHROOT_BY_ZERO or something (not sure...) and stop the simplification process. Essentially we're modelling an error as a changeType (I think? @evykassirer).

I hope this helps!

[evykassirer]

yay @adamisntdead thanks for taking this on!

What @hmaurer said sounds right. Ideally a student could enter anything and give something back. If we throw an error, the tech that's using mathsteps might not know what to do.

So if someone asked 2 + 4 + nthroot(4, 0), we should give the step:

2 + 4 + nthroot(4, 0) -> 2 + 4 + nthroot(4, 0) (ie no change)
changeType: NTHROOT_BY_ZERO_ERROR (or whatever you want to call it)

Checking for 0th root should look through the whole tree before we start, because there's no point in doing other steps (like adding 2 and 4 above) if there's gonna be an error anyways.

How's that sound? Please ask if this is confusing or if you have any other questions :) thanks again for volunteering to work on this 😸

(aside: later, it would be nice to give users of mathsteps a way to check if a changeType is an error, but you don't need to do that)

[evykassirer]

this has gotten more complicated with #129 (sorry for being misleading!)

BUT once that's all done, this will still be a nice small change. I'll just add the "blocked" label to wait until #129 and #29 are done :)