Re: [TLS] DSS with other than SHA-1 algorithms

[Peter Gutmann <[e-mail]>]

Geoffrey Keating <[e-mail]> writes:

>For future work, I would ignore DSS.

I missed a chance to comment on this earlier, but when someone asked "why 
would you need a TLS 1.3 when TLS 1.2 is infinitely extensible", my response 
would have been "we might need a TLS 1.3 precisely *because* TLS 1.2 is 
infinitely extensible and therefore potentially infinitely complex".  TLS 1.3 
would have the one or two standard cipher suites that everyone uses anway, the 
one or two standard hash algorithms that ..., and a fixed way of doing things 
that would satisfy (literally, from the figures we've seen posted here) about 
99.995% of all users while vastly reducing the complexity (and thereby attack 
surface) of implementations.

Speaking of fixing TLS 1.2 problems, what's the next move for the action list 
of TLS 1.2 issues I posted a week or so back?  Can I get the edit token for TLS 
1.2bis, or should I do a distinct draft that'll then be folded into a future 
TLS 1.2bis?

Peter.