Re: [TLS] WG adoption: draft-nir-tls-rfc4492bis

[[e-mail] (Martin Rex)]

Stephen Checkoway wrote:
> 
> Martin Rex <[e-mail]> wrote:
>> 
>> So any *new* document should advise implementors to remove any such
>> erroneous and counterproductive "same algorithm checks" that might
>> accidentally got implemented by an unexperienced implementor who didn't
>> realize that this must be a defect in the spec and is backwards-incompatible
>> to SSLv3.
> 
> It makes sense that a new document shouldn't mandate unnecessarily
> restrictive behavior.

Such a restriction would be in direct conflict with rfc2109 section 6.

  http://tools.ietf.org/html/rfc2119#section-6

The original wording can be seen as a may "unless otherwise specified",
because it comes entirely without rationale, without any indication why
this would be a problem and with no security consideration whatsoever.


> 
> SSLv3 compatibility isn't high on my list of concerns though.
> I know you disagree, but we should stop using SSLv3 for all
> of the reasons specified in draft-thomson-sslv3-diediedie.

This is only relevant when looking at that statement from a viewpoint
of formal correctness of an implementor, who might have been reading
at rfc2246 in Summer 1999, when pretty much the entire installed
base was SSLv3-capable.

This wording, if it was meant to be real, would have required
a TLSv1.0 server implementation to negotiate SSLv3 (and not be allowed
to use TLSv1.0) if the server cert did not fulfil that requirement.

Such a "requirement", especially when it comes entirely without rationale,
is clearly bogus and void.


-Martin