In general, if you want us to remember to do something about your comments,
you need to open or update Trac tickets (discussion here before doing so is
fine, of course).

On 5 March 2015 at 21:39, Stephen Kent <[email protected]> wrote:

>  Overall comments
>
>  The correct expansion of the acronym CA is certification authority, not
> certificate authority. I believe I noted this typo in my comments months
> ago. I won’t cite each instance of it the document.
>

I have just done this, so no need to open a ticket.


>
>
> There are a number of instances where the document cites normative
> references, but the cited documents appear as informative references in
> Section 10.2
>

And I am doing this right now, so again, no need for a ticket.


>
>
>
>
>
>
> *Abstract*
>
>
>
> Reword the opening sentence; it’s five lines long!
>
>
>
>
>
> *1. Informal Introduction*
>
>
>
> Why is this still an “informal” introduction? Doesn’t a standards track
> RFC deserve a real introduction :-)?
>

It is informal in that the treatment of the subject is informal. The
remainder of the RFC is the formal part.


>
>
> public CAs -> Web PKI CAs? (since the scope, is TLS)
>
>
>
> each chain is rooted in a CA certificate … -> each chain terminates with a
> certificate verifiable under a CA certificate …
>
>
>
> “provide evidence to TLS clients that the chain has been submitted …” just
> TLS clients, not monitors?
>
>
>
> The text still talks about TLS clients requiring that all certificates
> they accept having been logged.  Since TLS client behavior is out of
> scope, this statement, which motivates the use of logs, is out of scope as
> well. (we can’t have it both ways, i.e., defer definitive statement about
> client behavior and then cite such behavior as a motivation for the
> proposed mechanism.)
>
>
>
> This section still discusses monitoring of logs, but defers specification
> of exactly how that monitoring is performed. Again, if monitor behavior is
> out of scope for this document, it is inappropriate to make vague
> statements about how such behavior justifies the development of a
> specification for logs.
>
>
>
> The phrase “domains for which they are responsible” is vague. It suggests
> an attempt to convey how monitoring works, while deferring the real
> definition to another document.
>
>
>
> “The checking operation is asynchronous to allow TLS connections to
> proceed …” The reference to asynchronicity suggests that this is describing
> TLS client behavior, which is out of scope.
>
>
>
> The last paragraph in this section asserts that it is efficient for
> (unspecified parties, formerly known as Auditors) to detect log
> misbehavior, but does not provide evidence that such detection is, in fact,
> efficient.
>

However, the remainder of the document does. If we turn the introduction
into a formal treatment, it becomes the whole document. And then we need to
write a new introduction.


>
>
> This section refers to certificate mis-issuance, yet defines it nowhere.  
> Since
> mitigating problems caused by mis-issued certificates is the rationale
> cited for CT, this omission is serious. And, I note, that I have offered
> text to define this term.
>
>
>
> *2. Cryptographic Components*
>
>
>
>
>
> As I noted in my initial review of 6269-bis, I think most of the text in
> 2.1.1-3 belongs in an appendix.
>
>
>
> It is common practice in the IETF today to place crypto algorithm specs in
> a separate document, to better accommodate changes to such specs in the
> future. This suggests removing the specs of hash and signature algorithms
> from 2.1.4. There also is a need to describe how changes to the algorithms
> will be accommodated by CT.
>

This point has already been addressed in the next version of the I-D.


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

Reply via email to