>I do have to disagree with using "absolute"
>uri's for a namespace...I always thought that a namespace was just a
>unique identifier for elements within a document

Two years ago there was a _MASSIVE_ debate in the W3C about this. The
conclusion was that relative namespace URIs were officially Unsupported and
NOT recommended until/unless the W3C does reach consensus on how they're to
be interpreted.



Quick summary of the issue:

There are three

1) "It's just a string. It should be compared as a string. Absolute is
preferred because -- after all -- the purpose of using URIs was to bring
these strings under existing policies of ownership of parts of the URI
space, but if you insist on using relative it'll be compared as a string."

2) "It may some day point to something, so making it relative might be
useful and it should be absolutized before being compared/used."

3) "The namespace is involved in defining semantics of the document. The
concept of having a document's semantics change because you happen to copy
it from one place to another is _dumb_, and relative should be forbidden
completely."

Compromises existed that would satisfy any two of these camps. None existed
that would satisfy all three. The decision was made to declare relative
namespace URIs Unsupported -- nobody is required to accept them, and nobody
who does accept them is required to interpret them in any specific way,
which makes using them quite unwise -- until and unless the Semantic Web
task force reports back with some other recommendation.


Until then: Don't. Or don't complain when it breaks.



---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to