Uche,
you wrote:
"
It's probably the use of "documents" above which prevented me from
thinking of yours and Michael's documents in the same way. As I
understand Michael's quote, it's not just a convenient view ver existing
documents, but fundamentally redesigning what "XML document" means, at
least for this limited scenario. Rather than documents I would think of
the component bits as more like general entities or XInclude fragments.
In my opinion what Michael is proposing is in effect redesigning the
fundamental behavior of an XML parser whereas what your super-documents
idea is doing is layering some additional, useful capability on top of
XML parsers as they already exist.
"
I have an inkling what you mean, although I am not sure about the significance of distinguishing documents from fragments. And I wonder if it is not in fact better (!) to let those fragments be conventional documents and let the view decide whether they play the role of a fragment or the role of a document. Just as you said: layering the super-document structure on top of what already exists (the document model). Has a document any additional weight, compared to a fragment? My uncertainty could be summarized like this: what are the advantages when switching over to a model of distributed fragments, compared with keeping the document model and adding the super-document model? Perhaps we should really take some time (weekend and more) to
consider these matters carefully, interesting and far-reaching as they are.
Hans
