Re: [apps-discuss] "finding registered domains"

It's very helpful, thanks.

So for the sake of looking through all of our options, could you (or
anyone) comment on why we might not want to do a downward tree walk looking
for a first administrative boundary record?  It seems like those walks are
limited to only a couple of queries at most and could be cached for a very
long time.

It's also been suggested that maybe the solution for this should come from
WEIRDS.  The reason that might be is that we're trying to impose onto the
DNS some structure that isn't there (I hope I'm paraphrasing my DNS
colleagues properly there), so maybe the answer to this question should
also come from outside the DNS.  Is that something that deserves serious
consideration, or is it off the mark?

-MSK

On Mon, Mar 11, 2013 at 5:08 PM, Andrew Sullivan <ajs@anvilwalrusden.com>wrote:

> On Mon, Mar 11, 2013 at 04:23:25PM -0400, Murray S. Kucherawy wrote:
> >
> > I'd like to understand how these are related, or instead how they are
> > actually totally different problem spaces.  I think this is important to
> > figuring out a path forward.
>
> In my opinion, these are in fact the very same problem, just with
> different expressions.
>
> At bottom, the issue is always that there is an attempt to read from
> the hierarchical delegation space of the DNS some sort of
> administrative commonality that just doesn't exist.
>
> In general, domain names just represent the labels of every node down
> the DNS tree from the root zone.  So, the name
> label1.label2.label3.example.com. names several levels of the DNS
> tree.  And labelA.labelB.labelC.example.com. also names several levels
> of the DNS tree.
>
> What we _cannot_ tell from any of that is who is in control of which
> parts.  This is true regardless of whether there are any delegations
> (SOA records) in the tree anywhere along that FQDN.  You can tell that
> example.com in both FQDNs are under the same control (because they're
> the same node), but you can't tell whether example.com is controlled
> by the same people who control .com (compare this with the co.uk/.uk
> case).  Similarly, you cannot tell whether
> labelA.labelB.labelC.example.com. is controlled by the people who
> control example.com (compare this to web hotels, for instance, which
> have of course mostly fallen out of fashion for this reason).  There
> are some historically good pretty good rules of thumb, but that's all
> they are.
>
> The "cookies problem" is just a special case: the cookies spec had
> peculiar rules about TLDs, and now has some special handling of
> sibling and children relationships, but it's still both too broad and
> too narrow.
>
> Does this help?
>
> Best,
>
> A
>
> --
> Andrew Sullivan
> ajs@anvilwalrusden.com
> _______________________________________________
> apps-discuss mailing list
> apps-discuss@ietf.org
> https://www.ietf.org/mailman/listinfo/apps-discuss
>