I'll add it to the test cases, thanks!

Fixed, thanks!

The Python version simply did not have the same example code that the JS and CoffeeScript ones did, because I did those as language-specific, so I had to explicitly add it later. It has been modified to reflect this.

It looks like when I was creating the new description I accidentally deleted the API specification and didn't notice. Thanks for pointing this out.

Sorry for the late response. It seems like you're returning an object rather than a number.

Because that's a Python only kata. I was going to add that as a separate one. Only Python supports operator overloading, so I don't want to make it a drastically different kata as a result.

(and for what it's worth, it would be somewhat inaccurate to implement __mul__ as the dot product, given that the dot product returns a scalar, not another vector!)

Fixed! Turns out, you can put as many spaces as you want in single backticked markdown, but only one will show up unless you use triple backticks... What a weird problem.

Where are you seeing this, specifically? Every instance in the code has two spaces where it belongs.

The thing is that the test environment expects that you throw an error, but you can't really tell it what kind of error it's looking for, so it thinks that the ReferenceError is the one it wants. That's just a quirk with the test environment.

That board is unsolved because it isn't full. Once X goes again then it's a draw. It's all semantics, though.

And uhh I'd have to look at your code? for-in loops loop over keys, whereas the other one loops over numerical values, so I'd have to see the code to make a statement.

Ah, then yeah, the server might be a bit slow.

Yeah, that could be it! Also make sure your code isn't too inefficient ;)

Yeah it would appear that throwing an error within your equals() method will mess up the testing framework.

Huh, totally thought I did mention that! I added that in the description. Thanks!