Hi all, I looked up the RFC6979 implementation in pybitcointools (line 367) and it is very compact. Like you mention above, the implementation is made a lot easier if we're only considering the bitcoin usecase, so there is not that much complexity in terms of checking code.

I personally also really like the deterministic k for the reasons mentioned earlier:

• Can use a solid randomness source (like dice) for initial high entropy seed and not worry about PRNG,
• Reproducibility, unit testing and standardized test vectors.