• Not an exact duplicate in my view, this opens up more solutions

  • The kata you linked is different from this one but I'm pretty sure that finding groups of consecutive values has been done before.

  • There's no reason this error would show up unless you've changed the function name. If it refuses to work no matter what, try resetting the trainer.

  • the fractions involved turned what was a nice simple (and satisfying) kata into more of a hassle

    Why? If not for the output formatting crap, which is annoying regardless of how convoluted the task is, this would have been a nice and simple task even with the additional log(value, base^n) -> log(value, base) / n simplification requirement.

  • I agree with uttumuttu, the fractions involved turned what was a nice simple (and satisfying) kata into more of a hassle. And as said, you could relatively easily make a follow up "hard" version with fractions, expressions of logs as inputs (rather than a single logarithm) natural logs, etc.

  • In addition to this, regardless of types involved, the specification should be more explicit about the types it expects. Currently the return values in the examples aren't quoted to indicate the expected return values are strings nor does it state strings are expected anywhere.

    I'd go pretty far in saying input/output types should be specified in exceedingly obvious ways, but at the least, I think one shouldn't have to 'try and see' or look at the sample tests to figure it out. The description ought to have all the information needed and be as explicit as possible—except possibly in the case of some puzzle kata.

  • Really, write actual specifications.

    Examples can clarify specs, but in no way replace them or otherwise introduce new ones. Examples should also follow specs.

    Specs are leading, examples are fluff. The kata should still work without any examples!

  • @Djacon: not better, no. You should follow uttumuttu's advise, on this.

  • Please, remove the parsing requirement, and make the input a pair of integers: base and value. The output should ideally be a 4-tuple of numbers too.

  • Examples are not specifications.

  • What matters here is to actually specify the task/rules to apply

    That's exactly the point. There can be multiple possible simplifications depending on how far you're willing to go, and without clear specifications it's impossible to tell what's expected.

  • the calculator will print 3/2 + 1/2log3(2) instead of 1 + log9(6)

    3/2 + log9(2) is also a valid result, and since both log9(2) and log3(2) can't be calculated exactly, I'd not say that either option is better than the other...

  • Loading more items...