RE: [xml-dev] XHTML 2 Working Group won't be renewed?

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

 > I keep seeing the idea being pushed of defining some sort of umbrella namespace that can be utilized to "wrap" other namespaces, so that you can essentially work with a default namespace for multiple tags.  

 

Yes, I've pushed similar ideas in the past. I've always thought there was room for a concept of a "namespace path" where a local name does not have to carry an explicit prefix to determine what namespace it belongs to, but is instead resolved by some kind of deterministic search algorithm: for example, if there is a set of namespaces in scope, look in all of them, and if the name is defined in more than one of the namespaces, give precedence to the namespace of ancestor elements in order of containment. This depends on some kind of mechanism where the set of local names in a namespace is closed and is defined somewhere.

 

Then you can start to have a concept of a document type or media type (or perhaps a "super-namespace") mapping to a set of namespaces and a search order, so once the document/media type is known, it doesn't need to contain any explicit namespace declarations at all, unless they are actually needed to resolve ambiguous local names, which will be rare.

 

(Java manages this perfectly well, where package names only need to prefix a class name if the class name is ambiguous within the set of packages defined in the document header).

 

A nice feature for incremental adoption is that a document with implicit namespaces in this form could still be well-formed XML 1.0 and could be mechanically converted to a document using explicit namespaces conforming to XML-Namespaces 1.0 or 1.1, and its infoset would be compatible with existing infosets.

 

Regards,

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