i.e., RegExp.prototype.test is not a pure function: it modifies its this. It's documented, but I'm unsure it's a good thing ( this of course really means I think it's not ).

And as you put it "can be approached differently" :)

Since it is not possible to find some good, hard measures of what "good code" is - all submissions that run the test suite are pr. definition "good". Though, I admit, some entries are on the edge.

But even the edge solutions are of value. Some people are very good at finding the shortest route to getting through the test cases - so when writing kata test cases you are forced to be very particular which will highten your awareness of what can be bypassed.

So, in all - every submission has a value - depending on your perspective, you can learn from each (unless it's very similar to your own solution, that is :))

JavaScript allow for unicode characters in identifiers - even written as \uNNNN.

Essentially this solution is a cheat which you will see when you 'translate' the uni code characters following the assignment (first line).