- From: Roy T. Fielding <fielding@liege.ICS.UCI.EDU>
- Date: Sat, 01 Jun 1996 03:07:22 -0700
- To: Paul Leach <paulle@microsoft.com>
- Cc: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
> Section 3.2.2 says:
>
> Note: Although the HTTP protocol is independent of the transport layer
> protocol, the http URL only identifies resources by their TCP location,
> and thus non-TCP resources MUST be identified by some other URI scheme.
>
> I don't think this is really true. "http:" URLs identify resources by
> DNS names, and IMHO it is the presence of an A record for that DNS name
> that signals that one must use TCP. In the future, the presence of an
> AAAA (IPv6) record would presumably mean to use TCP/IPv6 to access the
> resource (if this isn't true, isn't someone on the IAB going to kill
> us?).
Yeah, but that is still TCP.
> Similarly, if one were to invent an RR for IPX addresses or ATM
> addresses (if they haven't already), one could presumably run HTTP/SPX
> or HTTP/ATM (or HTTP/X.25,..., etc.) with "http:" URLs.
The http URL ultimately identifies a resource on a server that is
listening to TCP port (80 | what is given), because only that server
can map the abs_path part to an actual resource. How you get there is
a different matter.
> In any case, I'm not sure what compelling purpose this note serves, and
> rather than debate it, it might be better just to nuke it.
It serves as a reminder to implementors and (hopefully) as a way to
nuke people who say "HTTP can only be run over TCP because that
is the only protocol supported by the http URL", which explains the
"and thus ..." above.
...Roy T. Fielding
Department of Information & Computer Science (fielding@ics.uci.edu)
University of California, Irvine, CA 92717-3425 fax:+1(714)824-4056
http://www.ics.uci.edu/~fielding/
Received on Saturday, 1 June 1996 03:18:29 UTC