Re: [Add] [Ext] Mozilla's DoH resolver policy

[Brian Dickson <[e-mail]>]

On Thu, Apr 11, 2019 at 1:51 PM Adam Roach <[e-mail]> wrote:

> On 4/11/19 3:34 PM, Paul Hoffman wrote:
> > On Apr 11, 2019, at 8:46 AM, Peter Saint-Andre <[e-mail]>
> wrote:
> >>> If Firefox is to be taken mainly as a real-life example on this
> >>> mailing-list, is there a suitable Firefox-specific place to discuss
> >>> related issues?
> >> The Networking:DNS component is the one to watch:
> >>
> >>
> https://bugzilla.mozilla.org/buglist.cgi?product=Core&component=Networking%3A%20DNS&bug_status=__open__&list_id=14662850
> >>
> >> We don't (yet) have a discussion list for "Firefox Doing DNS".
> >>
> >>> (Assuming Mozilla is actually interested in that.)
> >> Feedback is always welcome. :-)
> > Do you mean "feedback on the ADD list" or "feeback in the Bugzilla
> tracker" or both?
> >
> > --Paul Hoffman
>
>
> I think we'd like to prevent the "Applications Doing DNS" list from
> becoming just the "Mozilla Doing DNS" list. To the extent that you'd
> like to engage with a broader slice of Mozilla than just those of us who
> participate in the IETF, I'd encourage you to take Mozilla-specific
> concerns to the mailing lists I pointed to earlier.
>

Hmm...

I disagree on multiple distinct fronts.
One is "Note Well", whereby all the IETF protections/restrictions are
applied to IETF-hosted mailing lists and all other IETF activities. I
believe there is benefit in keeping this discussion here.

The second is, even if some of these appear to be Mozilla issues, those are
mostly because they're visible (not sure if "first" applies, but it could
as well).

Whatever surfaces here, is valuable to capture here, as a way of helping
inform anyone else that tries to do similar stuff (whether in a browser, or
in any other app) in future.
At a minimum, being able to search this mailing list is probably very
helpful.

The third issue is that whatever is done, might well be appropriate to
"standardize", at least in the sense of establishing expectations for how
any given vendor should do whatever is being done.

Ideally, the outcome could be actual interoperability between apps and some
common "substrate", whether it is implemented as an API, a protocol (if we
can consider "loopback" as qualifying for "on-the-wire" treatment), or a
configuration file or set of files.

Note that DNS itself standardized on the "master file format" for zones,
even though that was one specific implementation of not-on-the-wire data
(plus some metadata?). So, from the DNS perspective, it has significant
precedence which is maintained, including on such things as "presentation
format" (which often is a 1:1 mapping for "master file format").

Brian