- From: Al Gilman <asgilman@iamdigex.net>
- Date: Sun, 26 Aug 2001 09:42:45 -0400
- To: "Larry Masinter - LMM@acm.org" <lmnet@attglobal.net>, "Aaron Swartz" <aswartz@upclink.com>
- Cc: <uri@w3.org>
The tdb form would certainly appear to be useful in creating the sort of URIs one needs for the "predicate" URI in an RDF (subject, predicate, object) triple. It fits the 'test case' reference in EARL to a T, likewise the usage surrounding a namespace in a language specification which has a proper URI. The prohibition of future times is unfortunate, a mis-feature. We can't do Business without such forward references. The 'that described by' inversion is of interest whether the 'by' is a dated citation or not. for the same reasons. There are business applications for "that described by <reference> of date." That is to say the version in force at the date in which the constraint is evaluated. If this violates a semantic assumption of URNs, then the tdb form should be a new root scheme which may take an URL or URN (encoded as required) including a duri as its 'by' argument. The encoding could be its own I-D as it deals with the general matter of embedding URIs in URIs and the dated case and the X3D case are not the end of this. Is there no sufficiently general way to work the date stamp into the resulting URI form as a parmeter ";as-of=dateStamp" ? The parallel to tdb at the root scheme level would be "The Resource Itself,' indicated by a 'tri' scheme token. Either tdb: or tri: takes an indicatedBy=encodedURI field which takes an optional 'as-of' dateTime parameter. There is no way to take a search URL and make it name an identity, simply by restricting it to a certain date. Search URLs restrict the scope of a service. The identity of the _thing_ or things referred to is determined by the operation of the service, not the requester providing the URL. So it is not clear that the 'duri' form is safely a URN either. Both the 'tdb' inverse application operator and the 'as-of' dateTime restriction are applicable to arbitrary URIs, but the former should be a prefix and the latter a suffix because the former inverts the sense of the reference while the latter narrows it. Al
Received on Sunday, 26 August 2001 09:22:37 UTC