Re: Insertions

Tom Hillman <tom@expertml.com> writes:
>  data: @xmlns, value+",".
>  xmlns: +"http://example.com/data".
>
> If your output is serialised text, this sidesteps some of the namespace issue.
>
> But if your output is XDM nodes, this should throw well-formed-ness
> errors as namespace declarations are not attributes, and attribute
> names starting with [xX][mM][lL] are forbidden (as I understand it).

Right. If we decide to pursue insertions, and if we allow them to create
attributes, I suggest we forbid attributes that begin with “xmlns”.

> It's only limited to the default namespace binding if colons are
> disallowed in rule names, right?

Yes. When I first started using ixml, I thought it was obvious that
nonterminal names should allow “:” and have the “natural” effect of
creating namespaced elements. I’m much, much less enthusiastic about
that now. Partly because I don’t think vxml ever needs to have
namespaces in it and partly because I think we’d be inviting a whole
bunch of issues around when, where, and how namespace declarations are
produced in the result.

There will always be things you can’t do *directly* with vxml output. I
think I’m ok if “make namespaces” is one of them. I could probably be
persuaded that a mechanism for creating a single, default namespace
declaration on the root element was worthwhile.

                                        Be seeing you,
                                          norm

--
Norm Tovey-Walsh
Saxonica