Re: Ephemeral XML?

Ralph Ferris <ralph@fsc.fujitsu.com> wrote:
> I don't consider my proposal to be "namespace pollution." I consider it to
> be method over-riding, in the manner of object-oriented programming.
> [...]
> One solution is to define these tags as architecural forms, and make these
> architectural forms part of a default set that's used by XML.
> [...]
> Pre-defined interfaces with over-rideable methods are what make component
> software work. Call these "component DTDs."

This sounds like a good approach, as long as the XML spec
also mandates some form of declaration to enable specific
architectures.  IOW, predefined element types are OK, but
we should make sure that documents are only interpreted
with predefined semantics if the author explicitly asks
for that to happen.

This may not seem like a big deal if there is only one
predefined element type ("use any GI you like, as long
as it's not ALINK") but if, as I suspect, there end up
being many different XML sub-architectures each with its
own collection of forms, protecting authors from unforeseen
nameclashes will be important.

--Joe English