> From: Michael Kay <mike@s...>

> Date: Thursday, 15 August 2013 6:14 PM

>

> I've always felt the right way for XPath/XQuery to achieve that is

> some kind of "abstract axis" notion where the relationship is accessed

> using an axis, say "bookings-for-room", and the implementation of the

> axis hides the implementation detail.

.. and ..

> From: Peter Hunsberger <peter.hunsberger@g...>

> Date: Friday, 16 August 2013 5:17 AM

>

> As I followed much of this discussion it struck me as a bit document

> centric, so I absolutely agree; "nodes on demand"... I have a project

> I'd like to embark on if I can find the time, which is driving

> something such as Saxon directly from a graph database.  Ultimately, I

> want graph traversal and graph composition. I see tools such as xPath

> and XSLT as pretty good competition for some of the current ways to do

> that.

Makes me think of the abstract axis as an RDF predicate, meaning we really can model "the universe" via RDF triples?? :)

-Gareth