On Fri, Oct 17, 2025 at 06:13:34PM +0000, John Gray wrote:
> From: Ilari Liusvaara <[email protected]>
> > Then another problem is that the LAMPS draft uses ASN.1 for RSA and
> > ECDSA, which greatly increases the complexity of encoding (I did
> > implement that stuff for a test, and the difference against custom
> > formats was just wild). 
>
> The LAMPS draft uses existing encodings of RSA and ECDSA.  Why would
> we need or want to invent a new encoding of RSA and ECDSA when every
> cryptographic library already supports them.  

The one I am using (for those two) does not (no, I did not write it).
:-)

And for ECDSA, IEEE is pretty standard format, and already used by
both COSE and JOSE. :-)

> Then people would complain that they had to re-encode the keys into
> existing encodings so they would be acceptable to their cryptographic
> libraries.

I think almost all libraries (also) have interfaces that take decomposed
things. And if not, re-encoding is much easier and safer than parsing
(e.g., size the buffers for the adversarial worst case).


> > I think that the current state of post-quantum signatures is so bad
> > that very few are going to use those without either some compliance
> > requirements or an imminent CRQC. And I do not think any compliance
> > regime is going to require hybrids (and hybrids will not improve
> > matters against imminent CRQC).
> 
> We already need to use them and have customers asking us for them.

I think most of that is due to compliance requirements. And the way
things work, "2035" might very much mean fire today.




-Ilari

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

Reply via email to