Is XPATH the way to search a data model?

[Martin @ ROE]

David Berry wrote:

> But once the XML has be de-serialised into an in-memory structure why not
> use class methods which have intimate knowledge of the semantic meaning
> of the class properties and can therefore do more intelligent decisions?

I think both ways (XPath & class method calls) are really the same if 
you mean using 'getXxxxx' class methods.  Both are just ways of 
navigating through the data model.

However I am sure we can expect our libraries to add more 'helper' 
methods that will do more than just blindly get aggregated/associated data.

> XML is just a dumb way of marking up text - why is it being seen as the
> answer to all VO problems?

It's worse; VOTable is seen as being the answer to all VO problems >:->

XML is more than just a text markup if used well, as it can make it easy 
to map model <-> serialisation.  But you're right, we should focus first 
on the data model.  If we need to look at the likely queries, we should 
really model those too before worrying about their implementation.

-- 
Martin Hill
Software Engineer, AstroGrid (ROE)
07901 55 24 66