On 3 Sep 2015, at 10:47, hellekin wrote:
On 09/03/2015 11:36 AM, Joel Halpern wrote:
Actually, DownRef won't cut it as far as I can tell.
The two documents are not stable. As a github reference,
they are simply "the most current version of foo".
Come on, GitHub is a corporation, it has NOTHING to do with it.
Pretty sure Joel was just referring to where the current documentation
is stored, not poking sticks at corporations.
Github repositories are sometimes pretty ephemeral as projects
restructure. RFCs live potentially for ever. There's some benefit in
references in RFCs being long-lived (it's a kindness to readers from the
future).
This particular spat is about the appearance of the URL, not its actual
permanence (unless Joel is one of those readers from the future). I
think Joel is inferring that the reference might be left dangling based
on what he expects from the behaviour of other projects hosted at
github.
What the onion folks said to me was that they were working on
creating stable, referenceable documents that explained how
this should work. I understand there is a time problem due to the
certificate. As a reviewer, I don't see how you can register
without a stable reference for the meaning of the registration.
The referenced Tor documentation does not affect the reservation
because
changes in that documentation won't affect the principles set forth in
the relation between Tor and DNS. These are orthogonal matters.
I agree. It seems to me that the matter before the IETF is to determine
whether tor exists, whether onion names are in use outside of the DNS
and whether cross-pollution between tor and the DNS is a big enough
problem to care about, not to provide an anchor for any particular
specification.
There's nothing in 6761 section 5 that demands a reference to a stable
specification. The fact that there's any reference at all is a courtesy.
Joe
_______________________________________________
DNSOP mailing list
DNSOP@ietf.org
https://www.ietf.org/mailman/listinfo/dnsop
