> 2) As an interpretation rule, all relative URI references are to be > absolutized relative to the fixed base URI > "http://www.w3.org/2000/Namespaces/base/" Thinking about this a bit more, it would be better if a different protocol or hostname part could be used such as "http://relativenamespacenames.w3.org/2000/Namespaces/base/" (or any other base that ensured that no relative URI from that base would ever produce a currently used URI on the W£C site.) Using the one you suggested (or the one I suggested earlier) would mean that xmlns="../../../1998/Math/MathML" would be a valid declaration of MathML, which isn't very nice. Having a different server would also simplify setting things up so that an attempt to dereference any such namespace name could succeed and return a text saying something along the lines of This is a resource corresponding to the namespace <whatever> DavidReceived on Thursday, 8 June 2000 05:54:53 GMT
This archive was generated by hypermail 2.2.0 + w3c-0.30 : Tuesday, 12 April 2005 12:17:24 GMT