Lists Home |
Date Index |
>>I'm just trying to persuade Patrick not to use a syntax that's
>>similar-enough-to-XML-to-be-confusing as the input to his processes
>>in the examples that he uses.
> In my defense, I am trying to persuade Jeni to not see a data model
> as a limitation on a particular serialization syntax.
I don't think that I do. I'm quite happy for XML to be interpreted as
the Infoset, as the PSVI, as the XPath data model, as the DOM data
model, as the LMNL data model, indeed as any data model anyone wants
to use with it! XML is a syntax, that's all.
> As I said, yesterday and I suppose it bears repeating, JITTs can use
> standard, valid, well-formed XML documents and syntax for many
> things. It can also use XML syntax that violates the XML data model
> but I fail to see why that is confusing?
I'm not sure what you mean by "XML syntax that violates the XML data
model". There is no *the* XML data model -- XML is just a syntax. But
there are very clear rules about that syntax -- the well-formedness
rules. My point is that if a document breaks those well-formedness
rules then it isn't in XML syntax. It's confusing to label documents
that aren't well-formed XML as "using XML syntax".
> That a serialization syntax is based on a particular data model is
> fine. But the interpretation of that serialization syntax should not
> be bound to the data model of its origin.
I agree with this statement entirely.