> what matters is that, once one is standard, the mechamisms to provide for
> symbol distinction / identity will already exist. it doesn't make sense to
> require that a second mechanism duplicate (part of) the behaviour of the
> first. any processor which can read a document which follows the scahem spec
> will already have to be "namespace aware". it makes much more sense to require
> of the first mechanism, that it offer an (additional) interface function of
> the form
> (document X qualified-name ) -> symbol.
> yes, this is missing in the present namespace proposal. which is reason to fix
> it (given that it's still in flux), not reason to do something else.
This argument proves too much. XSchemas mostly duplicate what DTDs
do, so why create them at all? XSchemas, however, are made of general
mechanisms and so provide the potential for extensibility, which DTDs
-- John Cowan cowan@ccil.org e'osai ko sarji la lojban.