Infoset vs pure syntax permathread (was Re: [xml-dev] The subsetting has begun)

On Tue, 25 Feb 2003 01:00:09 -0500, Gavin Thomas Nicol <> 
wrote:

>
>
> Right... but FWIW. XPath, and Schema are really tied closely to the 
> syntax of XML too. I don't think many many people would say that XPath is 
> a globally applicable addressing mechanism for structured data

XQuery in both theory (the spec) and practice (e.g. Microsoft's stuff that 
Dare was showing at XML 2002) operates off a data model (aka Infoset, more 
or less) that encompasses both XML and the relational model.  Since it 
incorporates XPath by reference, XPath is therefore a globally applicable 
addressing mechanism for data that can be represented in the 
datamodel/infoset.  What's wrong with that logic?

>
>> I suspect that the infoset is of little interest to ordinary programmers
>> doing ordinary work, who will work either at the level of the syntax or
>> of some particular API.  -Tim

I don't know how to reconcile that statement with the popularity of 
XPath/XSLT and the widespread enthusiam for the *basic idea* of XQuery (net 
the mumbo jumbo we like to complain about, which has more to do with the 
PSVI than the Infoset).