RE: Thoughts on namespaces

The instance model of the two is almost identical, modulo the small
differences I cited.  The Japanese proposal suggests in addition a
mechanism for combining DTDs.  I don't make any proposal for or against
combining DTDs.

Well-formedness and validation are two different levels. The instance
model permits well-formedness and unambiguous, infinitely extensible
names. The DTD model would be responsible for integrating several
namespaces to define validity. As the Japanese proposal shows, this is
not difficult. Given that it is not difficult, and that validating DTDs
can be created today (without waiting for a standard mechanism for
combining DTDs) the exact details of the DTD issue are not an impediment
to resolving the instance model. 

The proposed instance model simply says that, regardless of how the DTD
is phrased, names consist of two parts: the namepace and the specific
name.

Received on Monday, 26 May 1997 17:47:15 UTC