Re: [xml-dev] RE: Namespaces Best Practice

[Sean McGrath <sean.mcgrath@propylon.com>]

[Jonathan Borden]
 >It seems to me that placing the prefix-URI bindings at the top would ease
 >this task, but if you would rather perform this somewhere else in a
 >document, then I see no problem. What ever is more clear for a person to
 >read. (this -ought- not make any difference to a well designed XML parser so
 >this guideline exists entirely to allow people, so inclined, to better read
 >XML documents)

Is it just down to human readability? The placement of the prefix-URI
bindings used in document branch X effects the contextual
dependencies of X. These contextual dependencies have a significant
effect on the extent to which fragment X can be treated as
a tree in its own right-rather than as a branch of another tree.

One of the beauties of tree structures is that the tree/branch distinction
can depend on your point of view. Many beautiful algorithms
exploit this self-similarity. i.e. algorithms that process a tree by
treating the tree as a collection of smaller trees etc.

Implicit in such algorithms is that the interpretation of a branch
does not require information from beyond the branch. prefix-URI bindings
of course get it the way as they they make branch scope vital to
the semantics of the branch. If the bindings can occur anywhere
in the hierarchy of the tree then the only way to re-introduce
the a-branch-is-a--tree-is-a-branch view is to expand all
the namespaces. Disagreements about how to do this
on this list lead me to believe that this simple sounding
task hides numerous wasps nests. This, I think, is
very unfortunate. Doesn't it force us to think of XML documents
monolithically? Doesn't it make the phrase "XML Fragment"
oxymoronic at anything but the triturated brain
PSVI level?

Does it mean the the cosy world of cut'n'paste style
XML processing used in, for example, Pyxie
http://pyxie.sourceforge.net/ is lost and gone
forever?

Ah, the years, the years,
Down their carved names,
The raindrops plough

Sean