Mind if I ask what this enormous API complication would entail (beyond, say, 
implementing ML-DSA signatures in TLS)?

The complexity would be:


  *
Including the code points
  *
Including calls to the crypto engine when those point points are referenced

And, well, from the TLS side, that's about it as far as I can see.

If you disagree about the need, well, that's quite fine (and that is quite a 
reasonable viewpoint).

However, we tried quite hard to make it as simple as possible for an 
application (in this case, TLS); it's difficut to see how it could be simpler.  
I'm just tired of people who dislike a proposal to automatically raise the 
"complexity" objection, when the complexity is really quite minimal (and zero 
if you decide not to implement it).

________________________________
From: Salz, Rich <[email protected]>
Sent: Wednesday, April 15, 2026 10:37 AM
To: [email protected] <[email protected]>
Subject: [TLS] Re: Composite ML-DSA

I am also opposed to composite signatures for TLS. I don’t see a need that 
justifies the enormous API complications that would ensue. And, frankly, much 
more likely to have implementation bugs for the first couple of years.


_______________________________________________
TLS mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to