@bozh, I agree 👍 str.each_char seems like a better option.

Done.

On it.

Added a Javascript translation!

Added a python translation!

Fixed. I've added one test for checking whether projects are randomly chosen, and another to check that ordering is shuffled.

Ah, got what you mean. The randomness is choosing projects is tested, but not necessarily the randomness in order. Will fix.

Hmm. Wouldn't the above test cover that? Do you have any suggestions on how the test can be improved?

Interesting, didn't know that Clamp existed!

There is a test for randomness, it runs the same function 100 times and makes sure the number of unique solutions received is greater than 1.

Did your solution happen to pass that test without adding any randomness?

I can't make fundamental changes to the question anymore, however I've worked on the description to make these points more clear. Marking as resolved.

Thanks a lot for the feedback @seasidetesting, I'll try to improve this soon :)

How, for example, would you negate an object such as Object.new?

Every object has a true/false value in Ruby - doesn't always make sense, but the concept exists :)

However, what is expected to happen if someone uses the keys :foo and 'foo' in a Hash?

Hmm, that's valid. HashWithIndifferentAccess spoilt me :/

I think interpreting hash['!key'] as !hash[key] is a bit confusing
What is supposed to happen if someone used '!key' as the key?

Ah, yes - I kinda agree here. I'd never design a DSL like this, must've gotten lost in trying to add some complexity to this kata. I'm not sure if I can go back and remove that though?