Re: [xml-dev] Which is more declarative? More XMLish?

[Hans-Juergen Rennau <hrennau@yahoo.de>]

The direction of thought I have in mind has nothing to do with "constrained vocabularies". It has to do with an extension of our thinking about XML: the introduction of "semantic identity", as opposed to node identity; or to put it simply: the view that complex XML elements describe a thing (its properties), and that this description may be destributed over any number of element nodes. Introduction of a standardized way how to identify the thing - e.g. NIEM's @structures:uri attribute - enables the (semantic) equivalence of any distribution of a thing's description over one or more nodes. As a special case, the equivalence of a fully expanded child element and a child element containing nothing but its identity attribute becomes well-defined (assuming the content of the fully expanded version is found in other elements with the same identity value). As you see, the artificial "infinity" of the XML representation may vanish.

With kind regards,

Hans-Juergen

PS: I wonder if the W3C working groups have considered extending the XDM by a new node property which is an element node's "thing identifier" (IRI) - similar to NIEM's @structures:uri. 

Michael Kay <mike@saxonica.com> schrieb am 23:53 Samstag, 2.Dezember 2017:

On 2 Dec 2017, at 22:18, Hans-Juergen Rennau <hrennau@yahoo.de> wrote:

Michael Kay wrote: "the result of serializing the map as JSON and then converting the JSON to XML will be an infinite XML document"

But this assumes a certain way of representing JSON data in XML (or, more generally, "information objects" in XML),

Yes, I was talking in the context of XPath 3.1 and this phrase was intended to mean "applying the function serialize($map, {'method':'json'}) and then applying the function json-to-xml()".

Of course the XML representation of JSON delivered by this function is not the only one possible. In particular, it is chosen ("arbitrarily" if you like) as a format that gives a lossless XML representation of any JSON text. For specific constrained JSON vocabularies and data models (e.g. the Linked Data model), you will always be able to devise something that's more suited to the particular model. For example, you can always do better if you know that all the keys in the JSON text will be valid XML names.

Michael Kay

Saxonica