RE: Clarification of URI vs. Resource

My understanding of the problems DocuShare was encountering was they were
using URLs similar to:

http://demo.opentext.com/livelink/livelink?func=doc.browse&nodeid=22278

For access to all of the content in their repository.  Presumably this was
because they employed an implementation strategy of using a single CGI which
acts as a gateway to the DMS.

But, according to RFC 2396, everything to the right of the "?" in a URI is
to be interpreted by the resource, which is identified by the text to the
left of the "?".  My interpretation of this is that there is a single CGI
resource which interprets the text to the right of the "?" as a query on a
DMS, and returns the desired document.

But, WebDAV assumes that if you're editing a resource, that each resource
has a URL.  This condition appears to be violated by the use of "?" in the
DMS query URLs, since there is really only one resource here, the CGI
resource.

To me, this suggests that a DMS which previously to employed a CGI script as
a gateway to the DMS will now also need to add a module which maps URLs to
DMS items.

- Jim

> -----Original Message-----
> From: w3c-dist-auth-request@w3.org
> [mailto:w3c-dist-auth-request@w3.org]On Behalf Of Larry Masinter
> Sent: Friday, November 13, 1998 3:28 PM
> To: Slein, Judith A; ejw@ics.uci.edu; WEBDAV WG
> Subject: RE: Clarification of URI vs. Resource
>
>
> > It wouldn't mean you couldn't mention URLs, but just that you
> couldn't say
> > that URLs are members of collections, or design the protocol so that
> > anything depends on URL syntax for figuring out collection
> > membership.  Now that I sit down to look at it, I see that this
> > would require changes to the protocol, so never mind.  For example,
> > MKCOL now determines the parent collection as well as the name of
> > the new collection from the request-URI. And of course there would
> > have to be a new method for navigating up.
>
> This was the problem the DocuShare group was having in trying to
> implement WebDAV for DocuShare.  The problem is that you can't really
> add these features after the fact as 'optional' components, because
> clients will be released that will ASSUME that the URL syntax determines
> (the possibility of) containment.
>
> Larry
>
>

Received on Friday, 13 November 1998 19:10:35 UTC