On Tue, Jul 14, 2015 at 09:47:43PM +0200,
George Michaelson <ggm(_at_)algebras(_dot_)org> wrote
a message of 136 lines which said:
I think this is a politically naive request, and enacts a poor
architecture of name to locator models, and misunderstands both the
nature of domain names in general, the DNS specifically, and the
role of the omnibox and URL bar in a modern browser, which
inexorably relates to calls to gethostbyname() on the contents of a
URL.
I would love to see a more detailed analysis of the nature of DNS but,
in the mean time:
I think it would be a mistake to proceed with .Onion, and I think
the IESG would be well advised to re-consider the special use names
documents, because they are a poor fit for the modern world.
The "special use names document", RFC 6761, has already been used (for
.local) and nobody objected at this time.
since the rest of it is political and weakly argued, many would
chose to ignore it
No, there are practical reasons to "register" .onion:
* the forum of X.509 CAs decided to stop issuing certs for names
"unregistered" so if onion sites want to use HTTPS, they need some
form of recognition,
* the special handling by software (section 2 of the I-D) will help to
prevent leaks in the DNS (which are bad for privacy).