[xquery-talk] A doc()-map chain operator proposal

[Michael Kay]

> On 18 May 2015, at 16:56, Benito van der Zander <benito at benibela.de> wrote:
> 
> 
> Hi Michael,
> 
>> I think it would be quite inappropriate for any function or operator in XPath to attach special meaning to attribute of a particular name, such as @href. The only way it would make sense would be is if there’s a W3C defining the semantics of the attribute - e.g. @xlink:href - but almost no-one uses that.
> 
> Well, HTML defines its own semantics for all elements and attributes.

Yes, and that would be a possible way forward; though we’ve avoided providing XPath functions and operators that are specialized to a particular vocabulary in the past.
> 
>> Indeed, that is why XQuery does not allow a conformant implementation to add its own operators.
> 
> Too late to care about that, when there are already JSONiq's constructors.
> 

You can of course write implementations of languages that have the XQuery grammar as a subset, but they are not conformant XQuery implementations. Whether people care is another matter...

Michael Kay
Saxonica