RE: [xml-dev] Should we include the XML prolog in the XPath data model,

["Michael Kay" <mike@saxonica.com>]

> As I wrote earlier, XML prolog information could be 
> properties of the document node in XPath (2.x). It's upto the 
> applications to make use of this information.

Why would you want to tell the application about the encoding of the
document entity, without also revealing the encoding of every other entity?

How could you supply information about entities without including entities
in the data model? If you included them in the model, how would they relate
to the other objects in the model?

It's wrong to make these things properties of the document node when they
are actually properties of an entity. I'm not saying it would be wrong to
provide information about entities in the model, but it would add a lot of
extra complexity for a questionable return.

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