>>>>> CodesInChaos <[email protected]> writes:
> For .onion or namecoin a real TLD would be a good fit. For most
> others not so much.
> In many such protocols the uri contains secrets, and you don't want
> to leak those via DNS.
I wonder what are the specific examples?
> They usually have no meaningful path portion either.
Note, however, that having one or more TLD's registered with
IANA doesn't (AIUI) necessarily imply that any access whatsoever
is provided via the DNS protocols. Instead, such a registration
may only reserve the TLD for a particular purpose, so to avoid
name clashes between different applications (uses.)
> If you want to unify them, extending magnet links sounds like a
> better idea than a TLD.
AIUI, the magnet: URI scheme is nothing more than a container
format for various URI's (both URL's, like http:, and URN's,
like urn:btih:), as well as some metadata (&dn=, &kt=.) As
such, it may (and, arguably, that was the intent) rely on a set
of separately-registered URN namespaces for content-derived
identifiers.
Unfortunately, neither magnet: URI scheme, nor the specific URN
namespaces for such content-derived identifiers, were ever
registered. (And working towards registering these seems like a
good idea, indeed.)
I see nothing wrong with using GNUnet or Freenet URI's (URN's)
within magnet:-scheme URI's, BTW. (Perhaps a feature to be
implemented in gnunet-uri(1).)
--
FSF associate member #7257
_______________________________________________
p2p-hackers mailing list
[email protected]
http://lists.zooko.com/mailman/listinfo/p2p-hackers
