RE: [xml-dev] "self" axis for an attribute context node

> 
> Others might be able to explain the rationale behind this 
> design decision, if that's your question.

You'll have to ask James Clark.

I think it would have made much more sense for a NameTest to match all nodes
with the required name, irrespective of node kind. But XPath 1.0 was done in
a year, so not all decisions were perfect.

Michael Kay
http://www.saxonica.com/