Re: namespace usage as assertions

> You are blurring the question for which this list was defined:

I agree that the ultimate questions are whether relative syntax is to be
accepted at all, and if so how to interpret it. I was digressing  to try to
clarify the point that was being questioned,  which the meaning of one of
the proposed interpretations. Explaining it doesn't mean I consider it
preferable.

>> "The URI is just compared literally". This continues to bias me heavily
>> away from the Absolutize behavior.
>Au contraire, it *is* the Absolutize behavior;

Literal comparison of the declaration as written matches both the Literal
and Forbid cases. (Once you Forbid relative references, you have only
absolute URI or URI+, which can be compared as strings independent of any
other meaning or context.)

Absolutizing before comparing matches the Absolutize case. I still consider
this approach more confusing and cycle-consuming than helpful. Even should
the Semantic Web become a reality someday, there is as yet no evidence that
relative references to namespaces will be a useful tool in getting us
there.

The Deprecate/Undefined solution promises only that comparison of
declarations which are already absolute will yield results "as if" compared
literally. Relative syntax will have unpredictable results.


The DOM currently expects to handle the namespace name only as a string,
passing the buck for what format that string takes and how it's derived
back to the middleware and users. Depending on what happens, it may be
unfortunate that this field is currently named "namespaceURI".

______________________________________
Joe Kesselman  / IBM Research

Received on Thursday, 29 June 2000 12:36:47 UTC