- From: Roy T. Fielding <fielding@liege.ICS.UCI.EDU>
- Date: Tue, 23 Jul 1996 20:06:18 -0700
- To: Mark Nahabedian <Naha@ai.mit.edu>
- Cc: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
> I've been reading the draft HTTP/1.1 specification
> (draft-ietf-http-v11-spec-06) and have some comments.
>
>
> Section 3.2.1 General Syntax, page 16
>
> since <segment> is *<pchar>, it can be empty, implying that
> "foo/////bar" is a valid <rel_path>. Is this correct?
Yes.
> Section 3.2.3 URI comparison, page 17
>
> What is the intended behavior of the comparison for two
> different host identifiers (names or addresses) which refer to
> the same host? From the spec, one could conclude any of the
> following: (1) this issue was neglected; (2) the comparison
> algorithm is spared the effort and expense of doing DNS
> accesses; (3) the comparison algorithm is prohibited from
> doing DNS accesses. The comparison algorithm's behavior
> toward synonomous host identifiers, and the reason for that
> behavior, should be explicitly stated.
DNS accesses are irrelevant -- there is no such thing as synonomous host
identifiers because the hostname may affect the resource chosen. The
comparison is therefore limited to the contents of the URLs.
...Roy T. Fielding
Department of Information & Computer Science (fielding@ics.uci.edu)
University of California, Irvine, CA 92697-3425 fax:+1(714)824-4056
http://www.ics.uci.edu/~fielding/
Received on Tuesday, 23 July 1996 20:25:01 UTC