[noise] Protocol Names

[Jason A. Donenfeld]

On Tue, May 30, 2017 at 2:26 AM, Trevor Perrin <trevp at trevp.net> wrote:
> Thanks for drawing attention to the test vectors format.
>
> Looking at that, it feels like a better argument for continuing down
> the road of making the Noise protocol names parseable with a precisely
> defined syntax.
>
> At that point, we could simplify the test vector format by removing
> the redundant "dh", "cipher", "hash" etc parameters from the test
> vectors, since the test code could just parse the name.  Make sense?

I agree with that wholeheartedly. It allows for very little room for
error and mismatches.

Also, I think the JSON drum beat is a potentially dangerous one that
we should avoid if possible. "Defin[ing] a JSON schema to encompass
all possible handshake parameters/choices" sounds like Enterprise Hell
that we'd really benefit from avoiding.

Jason