I confinrm lot of problems with rounding issues in javascript, the tests should ALL be rewritten to allow solution in a range and not exact number to cover all the floating point math implementation

It looks like some work was being done here...

https://github.com/Codewars/codewars-runner/issues/9

It is possible to support scala directly? Considering you are supporting already Java (and so you must have a JVM running somewhere) it should be easy

I suggest for us users to have the ability to filter through our Katas(completed, unfinished, authored, etc...) by language. In case we'd like to re-do a Kata that we did in some specific language or if we simply want to show some katas we did in a language that we haven't used as much as the others.

I suggest to change the way kata difficulty is chosen.

For me, it is generally more comfortable to answer the question "Is this kata easier than that one?" rather than "Is this kata 6 kyu or 5 kyu?".

Also, some katas are currently misplaced in levels: I really think that http://www.codewars.com/kata/where-my-anagrams-at is much easier than http://www.codewars.com/kata/unary-function-chainer, yet the former is 5 kyu, while the latter is currently 6 kyu (and I think it was 7 kyu not long ago).

Currently, this relies on mean of user assessments during beta. I think this is not quite adequate, and this leads to almost no new high level katas.

For example, if Alice, Bob, and Carol mark a kata as 1 kyu, and then Eve marks it as 8 kyu, kata will hardly be purple. No one accounts for possibly wrong assessment by Eve.

I suggest another system. It should be based on comparison of current kata to some other katas that user have solved recently. User must answer the question: "if compared to %KATA% you solved recently, do you think this kata is harder, same level, or easier?"

After that, it will be possible to sort katas by their difficulty by calculating the relative difficulty using pairwise comparison of katas.

Top 1/128 of this sorted list can then be assigned 1 kyu,

   top 1/64  of the remaining katas  - 2 kyu,

   top 1/32  of the remaining katas  - 3 kyu,

   top 1/16  of the remaining katas  - 4 kyu,

   top 1/8   of the remaining katas  - 5 kyu,

   top 1/4   of the remaining katas  - 6 kyu,

   top 1/2   of the remaining katas  - 7 kyu,

...and the rest will be 8 kyu.

As currently we have 821 approved katas, this will result in:
  6 katas in 1 kyu category,
 12 katas in 2 kyu category,
 25 katas in 3 kyu category,
 48 katas in 4 kyu category,
 91 katas in 5 kyu category,
159 katas in 6 kyu category,
240 katas in 7 kyu category,
240 katas in 8 kyu category.

Also, as new katas emerge, old katas may drop or raise their ranks accordingly.

This can also be done in a language-specific fashion, as for different languages difficulty of a certain kata may vary.

Another benefit of this method is that kata's difficulty assessment can be ongoing, not just during the beta process.

Can you support directly Typescript? It should be very easy, just compile with tsc the code submitted and run it in node as you do with javascript, probably it would require just some minutes of work to implement it