Lists Home |
Date Index |
"Thomas B. Passin" <firstname.lastname@example.org> writes:
> Tim Bray wrote:
> > On Mar 5, 2004, at 6:19 PM, K. Ari Krupnikov wrote:
> >> The question boils down to whether an HTTP URI is opaque or if the
> >> question mark has predefined semantics. (The specs would have me
> >> believe the former, but the practice seems to be the latter).
> > 404. The URI is opaque. There is no such resource. -Tim
> I don't think so. Rfc 2396 says
> "The query component is a string of information to be interpreted by
> the resource."
> This shows that the query string is not actually part of the resource
There is further support to this notion in the fact that relative URIs
are interpreted relative to the last slash before the "?"; all
slashes in the query string are ignored. If the document at
http://example.com/archives/lists/xml-dev?date=2004/03/06 contains a
link to "../newsgroups", I'd expect it to resolve to
http://example.com/archives/newsgroups. Tim, are you saying it should
resolve to http://example.com/archives/lists/xml-dev?date=2004/newsgroups
> In this case, there _is_ a resource, but there is nothing to return
> about the view (or whatever) of the resource indicated by the query
> I think a better return would be a 204 -
> 204 No Content
The server may wish to make a suggestion about other resource on it
has that may be of use. It may even say what it thinks the error was:
if you requested information on the 44th US president, the server may
tell you that the valid range is 1-43.
> An HTML or XML document saying there was no result and why would also
> seem to be OK, but 204 ought to be better.
I disagree. It's like saying 404 doesn't need a response body.
Elections only count as free and trials as fair if you can lose money
betting on the outcome.