Re: [DNSOP] kskroll-sentinel and unclear results

["Wessels, Duane" <[e-mail]>]

> On May 24, 2018, at 7:51 PM, Paul Hoffman <[e-mail]> wrote:
> 
> On 23 May 2018, at 11:49, Warren Kumari wrote:
> 
>> 
>> ​I for one would like to see proposed text - we can decide from that if it
>> makes things clearer.
> 
> The proposed text is at
>   https://github.com/APNIC-Labs/draft-kskroll-sentinel/pull/21
> It's an omnibus change, so you might want to pick up parts, but I think as a whole it deals with the above concerns in a consistent fashion.


Paul,

I took a look at your pull request.  I like the direction this is heading, and
I like the change from "invalid" to "bogus."  It leaves me with a couple
of questions though.

Your new second paragraph of section 3 says "this entire section is about
DNS resolution systems" and "we can classify DNS resolution systems into
five distinct behavior types" but the subsequent descriptions of Vnew,
Vold, etc still talk about a single resolver?

Similarly, in Appendix A should the text be changed so that, for example,
instead of saying "Bob is not using a validating resolver" it says "None
of the resolvers in Bob's DNS resolution system are validating"?

"All of Charlie's resolvers are validating..."?

"All of Dave's resolvers implement the sentinel method..."?


Nitpick: 

  "If a client directs these three queries to a DNS	
  resolution system where the resolvers have different propertied,
  the results cannot be determined."

Should be "properties" and I'd say "cannot be reliably determined."

DW