Re: [TLS] [tls13-spec] relax certificate_list ordering requirements to match current practice (#169)

["Ryan Sleevi" <[e-mail]>]

On Tue, May 12, 2015 2:38 pm, Martin Rex wrote:
>  One of the worst problems in your proposed text is the replacment
>  of the term "self-signed [...] root certificate authority" with
>  two bogus occurrences of the term "trust anchor".
>
>  Trust anchor is a _strictly_ local concept, so this term
>  MUST NOT be used in the description of the TLS Certificate PDU.
>
>  self-signed root certificate authority is a simple technical term,
>  that does *NOT* imply trust by any party.
>
>  -Martin
>

Martin,

You will find the reasoning already provided in past messages for why this
proposed terminology change. "self-signed root certificate authority" is,
unquestionably, NOT a simple technical term (considering how many times it
has confused people, I should think this obvious). Please do re-read the
message is which I provided the justification for this change, and
hopefully provide counter-factual evidence if there is any.

Considering that "trust anchor" is a term extensively used in RFC 5280, I
also fail to understand your objection about it being a "local only"
concept. This is entirely in line with the current language where you can
omit certificates that you know your peer does not need.

I'm happy to work to resolve the objections, but I can't help but feel
like some, like this, aren't well justified or supported.