On 19 August 2014 11:36, Stephen Kent <[email protected]> wrote:
> Ben,
>
>> ...
>>
>>> The topic of how a CT-compliant TLS client deals with a redacted cert, of
>>> any type,
>>> is within scope for TRANS.
>>>
>>> What Chrome does is not a subject for TRANS, since you have already
>>> stated
>>> that Chrome will do whatever Google decides, irrespective of any TRANS
>>> RFCs
>>> :-).
>>
>> Google is obviously not unique in this regard - it's true of all
>> software, right?
>
> It's rather unusual for an author of an RFC to publicly state that he plans
> to ignore the RFC if it doesn't match his implementation plans.

As you are fond of pointing out, I am not the author, I am one of the
editors. I am perhaps overly honest in stating up front that I will
not unconditionally follow the consensus decisions of the WG, but
surely this is far from being a unique stance.

>>>> I am also mildly confused about how an RFC interacts with standards
>>>> that are not controlled by the IETF (i.e. the Base Requirements and
>>>> the Extended Validation requirements).
>>>
>>>
>>> Well, RFC 6125 is an example of a standards track RFC that talks about EV
>>> certs in the TLS context, so there is a precedent.
>>
>> As far as I can see only to say its out of scope.
>
>
> Only the first mention of EV certs appears in the section "labelled "out of
> scope."
> The reference to EV certs in Section 2.3.1 says that the RFC prefers use of
> the
> Subject Alt Name over the Common name for representing certain classes of
> IDs, but
> acknowledges that use of the CN-ID construct is OK for backward
> compatibility with
> "deployed infrastructure", citing EV certs.  In Section 7.2, there is
> another reference
> to EV certs, in the context of wildcard use. In that instance the RFC
> suggests that
> the guidelines published in 2009 allowed wildcards, whereas the RFC argued
> against their
> use except in one specific location.

It would be interesting to know if this is why EV now disallows wildcards.

> So, I see the RFC as an example of how an IETF standard can choose to
> accommodate, or
> reject, guidelines from outside the IETF. There are other RFCs that deal
> with this topic
> very directly, e.g., in the MPLS space. The bottom line is that the IETF can
> choose to
> view non-IETF standards as off limits, or can reinforce support for them, or
> reject them
> in the IETF context. I'm pretty sure we can find examples of all of these
> ways of dealing
> with other standards in the RFC collection. I suspect Russ Housely is
> especially well suited
> to cite examples based on his experience as IETF Chair.

OK.

_______________________________________________
Trans mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/trans

Reply via email to