Ranks can't be changed for old katas with high completions

Not a kata issue. Your code is just reversing the text, and not reversing each word in text. Note that word is composed of any non-space characters.

??

fixed~

The example conversions have the units flipped. Should be ex. 24C = 20.1R

Haskell fork with random tests.

Haskell fork that removes random testing from the sample inputs, among other tweaks that are detailed in the fork's discourse.

Enabled

Every translation now has approximate equality on the level of test checker, resolved

Done, I think I can close out this issue then ;^^

Added random tests to C#.

nahhh, if the kata was translated to C#, the signature would be CelsiusToRomer, doing this explicitly for each language with different naming conventions is not ideal as of current standards and is hard to maintain overtime.

So, the ideal approach would be to completely remove the function names, like just writing you've decided to write a function that takes...

As for the examples, the same generalized approach applies , something in the form of Input --> Output is sufficient. If solvers need to get a clearer idea, they may refer to sample tests which serve such purpose.

Added the different signatures for the first and reworked the example for each language for the second proposals, please check

Description should be language-agnostic in the meantime

Talking about these -->

celsius_to_romer()

For example: celsius_to_romer(24) should return 20.1.