Re: Are *relative* URIs as namespace nemes considered harmful?

James Clark wrote:

> The problematic case is when you have
> two URI references that are identical when compared as strings but refer
> to different resources (because they have different base URIs).  This is
> like having obj1 == obj2 but not obj1.equals(obj2).  I would call that
> inconsistent.  By only comparing absolute URIs (either because you
> forbib relative URIs or because you absolutize first), you guarantee
> that if two namespace names compare equal then they refer to the same
> resource.

But can that case arise in the namespace spec itself?  Comparison occurs in only
one context: verifying the uniqueness of attributes.  And that comparison
necessarily occurs within a single element, thus ensuring that both elements being
compared have the same context and therefore would absolutize identically.

If I've overlooked something, do you have an example showing how the uniqueness
test can give misleading results because of different absolutizations of the
attributes being compared?

Paul Abrahams

Received on Tuesday, 23 May 2000 09:16:46 UTC