Math: n-th root, logarithm with arbitrary base
Both the proposed implementations do fp rounding twice, and so produce an inaccurate answer. I think we should probably leave it to the user to define incorrect math functions, rather than bake them into the language.
2011/12/19 Axel Rauschmayer <axel at rauschma.de>:
But wouldn’t that rather be a reason for making these functions part of the core language?
2011/12/19 Axel Rauschmayer <axel at rauschma.de>:
But wouldn’t that rather be a reason for making these functions part of the core language?
MS have proposed log2, log10, log1p. This makes more sense to me. These already exist and are useful in other languages and are not too hard to implement.
If we go beyond that it would be nice to see a sample implementation that is not either trivial or inaccurate or both?
I haven’t seen these two functions among the proposed additions for Math (should these be in a math module?):
Have they been considered and rejected?
Axel