Fixed :)

Thanks, I will fix it right now !

The sample test setup is wrong. You're using the old version from yesterday.

I shortened the description of irrelevant information.
Could you give area of improvement to make it more articulate?

Yes you are right this make more sense.
I changed the result expected from None to an empty string.

Description mentions a lot of irrelevant things, such as capital/small letters, non-letters, and the exclusion of j, v and w, but they do not matter in the tests.

If they are irrelevant then they should be removed from the description. (Especially since it's already very confusing)

Why encoding and decoding empty string should return None, when the natural result should be empty string?

Thanks !
I have re-published a new version of the tests with your feedback included.
As you advise, I have reduced the number of tests.

Python testing framework is not used correctly.
Please see Python authoring guide for examples.

2000 tests is quite a lot. Do you think this many is necessary? Commonly recognized practice is to keep amount of random tests around 100.

It cannot be split into "cats and dog", because there is only one "d", it could have been split into "cats an dog"

In C# I get a failure because the following is false? This can be split into "cats and dog"
----- (false, Kata.WordBreak("catsandog",new List {"cats","dog","sand","and","cat"})) ----

It works, but where do you have to go to translate a kata?

that means that you should unpublish, yours, actually (and possibly translate the other one to C#)

Thank you for your feedback! :)