- From: Tom Baker <tbaker@tbaker.de>
- Date: Tue, 6 Sep 2011 11:38:05 -0400
- To: Carlo Meghini <carlo.meghini@isti.cnr.it>
- Cc: public-xg-lld@w3.org
On Tue, Sep 06, 2011 at 02:15:33PM +0200, Carlo Meghini wrote: > Very interesting debate indeed. > > I am not sure I have followed all the developments, but here it seems to me > that the problem is NOT the "text string" per sé. A URI (in its abstract > syntax) is in fact a text string, and so is an ISBN. The difference between a > URI and any other type of string is that a URI has a meaning associated to > it, and this meaning allows an agent (for instance a piece of software), who > knows there is a URI in a certain place, to do something with the URI > (whether display it nicely or dereference it and get back a representation). > So, a text string is fine, as long as the string conforms to a syntax with an > associated semantics. Interesting point, Carlo! I guess I'd put the difference this way: that the URI lies in the managed, global context provided by the Domain Name System, within which the URI (i.e., strings that are URI) are globally unique. In the case of HTTP URIs, those unique identifiers are also associated not just with a meaning, but with a protocol universally supported by Web browsers, so that the identifier can resolve to documentation. Tom -- Tom Baker <tom@tombaker.org>
Received on Tuesday, 6 September 2011 15:38:42 UTC