- From: Clemm, Geoff <gclemm@rational.com>
- Date: Mon, 26 Nov 2001 13:53:58 -0500
- To: "'WebDAV'" <w3c-dist-auth@w3.org>, uri@w3.org
I think Jim makes a good case here. It is particularly convenient that "foo:" is not allowed as a relative URI, since otherwise the use of "foo:" as the URI for the schema would be ambiguous. <humor> And this would retroactively give us a much better answer for why "foo:" is not a legal relative URI </humor> Cheers, Geoff -----Original Message----- From: Jim Whitehead [mailto:ejw@cse.ucsc.edu] Sent: Monday, November 26, 2001 1:26 PM To: Roy T. Fielding Cc: 'WebDAV'; uri@w3.org Subject: RE: RFC2518 (WebDAV) / RFC2396 (URI) inconsistency Roy Fielding writes: > OTOH, I have found at least one other example -- the "about:" URL in > Netscape Navigator -- so I guess this could be changed in a > future revision of 2396. A very principled case can be made for viewing the scheme namespace as a set of URIs. This URI space identifies URI/URN/URL scheme names. A scheme identifies a resource, the abstract concept of the particular identifier of locator space. For "http" scheme URLs, the URI "http:" identifies the abstract notion of locators for Web resources accessed via the HTTP protocol. The "dav:" URI identifies the abstract concept of property and XML element names used in the WebDAV family of protocols. Put another way, there is a set of resources: {{The abstract notion of locators for Web resources accessed via the HTTP protocol}, {The abstract concept of property and XML element names used in the WebDAV family of protocols}, {The abstract concept of locators for directory resources accessed via the LDAP protocol}, {The abstract concept of identifiers for people's email inboxes}} These resources map to the following URIs: {"http:", "dav:", "ldap:", "mailto:"} Once you accept that these are all valid URIs, then RFC 2396 can be viewed as needing revision to remove the (artificial) constraints on URI syntax that prevent the use of scheme name URIs. - Jim PS - So, yes, I am reversing my earlier opinion that RFC 2396 does not need revision.
Received on Monday, 26 November 2001 13:54:34 UTC