Re: [DNSOP] Questions about draft-adpkja-dnsop-special-names-problem-00

[Stephane Bortzmeyer <[e-mail]>]

On Wed, Nov 04, 2015 at 12:20:27PM +0900,
 Stephane Bortzmeyer <[e-mail]> wrote 
 a message of 73 lines which said:

> draft-adpkja-dnsop-special-names-problem-00 raises several issues,

And I forgot one of the most important ones, but I remembered it
during a discussion over sashimi this evening (the sashimi were good,
thanks).

The entire section 2, about "switches" is questionable because using
.bit or .onion is not only to change the *resolution* protocol but
also (and specially) to change the *registration* process.

These are two different systems. Of course, they have some links (the
fact that domain names are organized into a tree is used by the DNS
protocol for fast resolution) but not identical. The current version
of the draft says "any TLD registered in IANA-maintained root-zone
(use DNS)" which is not quite exact. Names registered in the
RFC2826-root are often looked up through the DNS but not always (some
people use local hosts file or LDAP to do it). And, more important,
some TLDs outside of the RFC2826-root do not always indicate a
switch. This is the case of .bit (if you already know Namecoin, you
can skip the next paragraph).

Namecoin uses a blockchain to store registered names. That way, you
can have meaningful names without a registry. Because few clients
speak the Namecoin API, most of the times, name resolution is done
through the DNS: you set up a local authoritative name server to
export data from the blockchain into a .bit zone that you load.

This example clearly shows that the TLD is not a "protocol
switch". That's because Namecoin is intended to address perceived
problems with the registration system, not with the DNS.