I had expected that real software would start to be developed alongside the
specs. This is true for XML1.0 where we have a very large number of parsers
but very much less true for namespaces, XLink (which is over a year old),
schemas, etc. So yes, the message is that if you actually want XML to do
something you have to hack it yourself - don't sit around waiting for
others. In particular beware of NOTEs to the W3C where there is no evidence
of anyone doing any hacking.
I think we will get there eventually - but I'm an optimist
P.