Lists Home |
Date Index |
We've had this discussion before.
1. A GUI sees the http:// turns it blue, and makes it clickable. People click
away. It isn't the specs' fault but it happens. If the implementor changes
that behavior, does that eliminate the possibility of dereferencing it?
2. Are systems using the URI and dereferencing it? In other words, is
the silence of the spec on the issue a tacit permission? If so, there
is an architectural problem here and people are righ to gripe about it.
As long as identity and location are conflated (the web architecture)
this issue will never go away. Since that is an accepted tenet of
web gospel, right or wrong, anyone who uses a URI will have to explain
it away or set firm policy every time for what an application language
requires for conformance. If they don't see variations of item 1.
From: Jonathan Borden [mailto:firstname.lastname@example.org
Jeff Greif wrote:
> The unfortunate problem with RDDL and all the alternatives except (1) is
> that you cannot tell without looking whether there will be something at an
> NS URI, and if there is, whether you need to know about it.
Suppose you simply decide never to dereference a URI regardless of what it
looks like. How does the existence of a document which _can_ be dereferenced
affect your life? As has already been said, you can always use google.
Indeed a tenet of the RDF framework is that URIs are opaque. One derives the
properties of a resource by what is said "about" the resource, not by
dereferencing its URI. One never _has_ to dereference a URI to use RDF. Nor
does one have to dereference a namespace URI _ever_ if all one want to do is
use XML namespaces as a way to disambiguate element names. What could you
"need to know about"?