On the second point, it's as Blind4Basics says: I can't change the tests anymore. I definitely would if I could. The only thing I can do is delete the kata.

On the first point, I put a lot of thought into the description of the problem, and I'm satisfied with it.

That's a great idea. Sadly, I can't change the test cases any more.

Yeah, I think there's a fundamental problem with the "same" exercise sharing a rating across different languages.

That's because negative numbers would be out of scope. As the description says: "A string representation of an integer will contain no characters besides the ten numerals "0" to "9"." A negative number would also have a "-".

Those inputs don't contain any characters besides the numerals "0" and "3".

The kata is tagged "Big Integers," and the description includes precise and accurate information about the two arguments: they "will contain no characters besides the ten numerals 0 to 9."

Hmm, that's a shame. The error message should point out that the exercise requires a function which takes two strings and returns another string, not a number.

Nah.

I agree, but it's not something I can change.

The description is accurate. That it doesn't say everything it might is a feature, not a bug. Discovering and accounting for edge cases as you work through the exercise is the point.

I agree wholeheartedly, and I apologize. When I wrote the kata, I considered it an interesting exercise in javascript specifically. I don't know c# (that version was contributed by another user), but I can tell from people's reactions that in that language, the kata is ridiculous. I've tried multiple times to delete the c# version of the kata, but it's never worked.

Your solution returns a number, the description specifies strings.

I changed the tests to point out type errors.

I've added the arguments to the error messages to see if that makes the odder test failures more clear.

My experience in code challenges in interviews and the like is that constraints are added after you have a naive solution. I tried to mimic that style by saying nothing false (except 1+2 being 5, thank you for pointing that out), but communicating the challenge through successive, informative test failures. Where a person would say "what happens if", the tests say "strange, I expected this, but got another thing".

It seemed that this worked for you, at least, though it occurs to me the failure messages might be even more informative...