Re: [CFRG] Comment on draft-irtf-cfrg-hash-to-curve-10

[Daira Hopwood <[e-mail]>]

On 23/04/2021 20:30, Riad S. Wahby wrote:
> Hello Daira,
> 
> Thanks for clarifying your feedback.
> 
> I remain concerned about mixing implementation detail with high-level
> description. Here I am referring to using divsqrt in place of natural
> field arithmetic operations (sqrt, inversion, etc.) in the body text.
> Describing the algorithm independent of the implementation details is
> a way of specifying the mathematical properties of the algorithm, and
> having this specification explicit in the document has value, from my
> perspective.

The description of divsqrt (or sqrt_ratio) I gave *does* specify the
high-level mathematical properties of the algorithm, and keeps that
separate from how to implement it.

> But as I said in my prior email, it seems like refactoring Appx. G to
> use divsqrt and adding a few implementations of that function for the
> relevant cases (3 mod 4, 5 mod 8, 9 mod 16, and general, perhaps?) is
> a nice way of cleaning things up.

I think that if you explain the general case then the specializations to
{3 mod 4, 5 mod 8, 9 mod 16} are immediate and simple.

> And it seems like the same refactor
> applied to SvdW and Elligator in Appx. G would help, too.
> 
> This isn't something I can do in the near term, but I'm very happy to
> spend time on this once I've got some! I'm hopeful that's about three
> weeks from now, but I've been called an optimist before.
> 
> Thanks again for the feedback and best regards,

You're welcome.

-- 
Daira Hopwood