Re: [dnsext] Announcing public suffixes as a DNS record

[Phillip Hallam-Baker <[e-mail]>]

On Fri, Feb 1, 2013 at 11:58 AM, Paul Hoffman <[e-mail]> wrote:

> On Feb 1, 2013, at 8:30 AM, Phillip Hallam-Baker <[e-mail]> wrote:
>
> > Has anyone written a draft to propose such a record or is interested in
> helping to make such a proposal?
>
> draft-sullivan-domain-origin-assert deals with the core issue, and does so
> in healthier way than simply reproducing the Mozilla list (which is often
> wrong due to insufficient input and validation).
>

Well it is not just a question of 'the Mozilla list' there are multiple
lists in use maintained separately :(

The idea would be to migrate away from that position and do so by pushing
the responsibility for maintaining these lists onto ICANN and the operators
of large public spaces like blogspot.com


>
> I still would like to see draft-sullivan-domain-origin-assert adopted as a
> standard.
>

Andrew's proposal is essentially the opposite of mine. Instead of
assertions of the form 'this is a gap in the trust model', Andrew proposes
that the default is disconnected and trust is only established if there is
an explicit link.

That is definitely the right way to do this particular feature for
JavaScript if starting from scratch. (OK no the right way is not to do it
at all but..) But I don't see a viable transition strategy from where we
are to where we want to be.

Happy to be proved wrong here...

-- 
Website: http://hallambaker.com/