python new test framework is required. updated in this fork

Python: Random tests are vulnerable to input modification

My program sometimes passes the tests and sometimes not, but the code is the same. Is this a problem with the test cases?

Python:

• Tests use rounding as a workaround for equality of floats

General:

• Description should use Markdown instead of HTML formatting, and possibly some KaTeX for math notation.

The instructions don't cover the case of an empty input string well enough.
From the description I would assume that rule number 5 applies (because all other rules don't match and because an empty string has no vowels).

No random tests.

Ruby 3.0 should be enabled.

Link needs to properly inserted.
Wikipedia link states:

When referred to as the divisor function, it counts the number of divisors of an integer (including 1 and the number itself)

and here:

where σ is the divisor function (or sum of all divisors)

Clearly sum != count, that needs to be fixed.

C# translation added!

• Tests are not enough stressing: brute O(N²) solutions pass the tests.
• This kata is far older than this one, but it is far less well conceived. I think ecolban's one should have the precedence.

The images in the description are broken.

I have several suggestions to improve this Kata.

I suggest that the title be changes to 'Variance in the length of words in an array'. This is a little clearer.

The section describing the requirement to limit the result to four decimal places was not clear to me. I suggest that in addition to updating the description that you also add a test case illustrating this requirement.

I am unfamiliar with the term 'aleatory variable' and prefer the term 'random variable'.

Thanks for your contribution.

Interesting kata! There are the following issues:

• the link to the other kata leads to nowhere.
• it should be clarified that here divisors include the number itself; that is, the traditional perfect numbers are defined by k = 2. An example larger than 1 would help.
• I would suggest to test for higher numbers. Now even the most unefficient algorithms easily pass.

Should have example test cases

Description is kind of hard to read, it also suggests one specific solution.