- From: Bijan Parsia <bparsia@isr.umd.edu>
- Date: Fri, 5 Mar 2004 22:18:59 +0100
- To: public-webarch-comments@w3.org
(This is just on the part of section 2 up to 2.1) """Parties who wish to communicate must agree upon a shared set of identifiers and on their meanings.""" This is false. A baby communicates distress and discomfort to his or her parents without there being any identifers, or even any identification going on on the part of the baby. I might be able to communicate that this large bolder crushing my leg should be removed by the stout and helpful non-english-speaking lass beside me by making somewhat spastic gesticulations. Or, in a more structured way, I might point at the bolder, or wap the bolder, and make a little rolling motion with my hands. """The ability to use common identifiers across communities motivates global identifiers in Web architecture.""" This makes me feel parse-challenged. The *ability* to use common identifiers motivates global identifiers? Why? How? On the one hand, I value the effort at terseness the TAG clearly made in this document. But, reading this I feel lost. This *little* bit of narrative is so elliptical to leave me no better off. The third paragraph seems to provide *some* explication, in that I guess I should be motivated by the power gained from the vastness of the choice (though, frankly, I hate tons of choice). "Thus, Uniform Resource Identifiers ([URI], currently being revised) which are global identifiers in the context of the Web, are central to Web architecture." (Editorial) I'll just point out that four paragraphs later: """The scope of a URI is global; the resource identified by a URI does not depend on the context in which the URI appears (see also the section about URIs in other roles). """" This reuse of "context" in contrary ways is confusing. """"Constraint: Identify with URIs The identification mechanism for the Web is the URI."""" Presumably this isn't *quite* right, as there is a need for some idenification mechanisms that are not URI based in order to associate (some, at least) URIs with resources for subsequent reidentification. Also, for example, host names identify things very critical to the functioning of the web, and yet, aren't URIs. Etc. """A URI must be assigned to a resource in order for agents to be able to refer to the resource. """ Even restricted to software agents, this is false. _x foaf:mbox <mailto:bparsia@isr.umd.edu>. Allows an OWL Reasoner to refer to me (since foaf:mbox is an InverseFunctionalProperty). (While there was a URI involved, it wasn't assigned *to me*.) I can make or refute assertions about me in this way. """Resources exist before URIs; """ If URIs are strings, and string are abstract mathematical entities (i.e., a kind of data structure) independant of their physical instantiation, then, reasonably, URIs have always existed, so any particular URI has existed before some recently come into existent Resources. I'm not even sure of the point of such metaphysical statements. Or imagine I have, oh, a programming language where I have URI objects (a subclass of String). Let's say I want to use a URI to identify some other objects in my system. Does this claim require that (in pseudopython): my_object_uri = URI('http://blahblah.com/blah') #The URI now exists! my_funky_object = FunkyObject() #Now the Resource in question exists. my_object_uri.assigned_to(my_funky_object) is broken in some way? Why would this matter? """a resource may be identified by zero URIs."""" Ah, this is what you mean? It's not very happy either. I take it you mean that some resource might *not* be identified by *any* URI. Cool. And given my above example, it might still be possible for agents to refer to it. Naturally, it's often a good idea to give various resources a URI! For example, I don't think it's possible (or, at least, easy) to *link* to something in a machine readable way in HTML. So, give such resources URIs, please. I think it's quite possible to make the sensible point without appeal to broken metaphysics. Actually, the rest of the paragraph seems quite good and sensible. """Principle: URI assignment A resource owner SHOULD assign a URI to each resource that others will expect to refer to."""" I would recommend the TAG study FOAF because that community has made a different choice (i.e., to rely a lot on inverseFunctionalProperties). Aside from that, I think this principle misses an important point: Formats and protocols should (often?) be designed to use URIs. This encourages URI assignment by adding value to such assignment. Cheers, Bijan Parsia.
Received on Friday, 5 March 2004 16:18:56 UTC