Re: Do you or Dont you buy Tim Bray's Namespace Validation Algorithm?

Mark Tucker (mct@foyt.indyrad.iupui.edu)
Thu, 10 Sep 1998 15:53:26 -0500 (EST)


John Cowan cowan@ccil.org wrote:

mct> 4. Do the same to the DTD's that you read in.

jc>This is the sticky bit. Tim Bray phrases it as "If necessary, rewrite
jc>the DTD so that the same URIs have the same prefixes in DTD and
jc>instance."
jc>
jc>However, there is no formal way to express what prefixes in the DTD
jc>refer to what namespace URIs. In essence, you must know that in
jc>advance.

Hmm, isn't there some place at the top of the Document defining
the DTD to write
xmlns:BK=uri:books
and in so doing define the meaning of the prefixes this DTD uses?

jc> For validation to be a mechanical process, you must have
jc>some way of recording the frozen namespace-URI map for the DTD.
jc>

The namespace_prefix<->URI map for the DTD would be embedded in the DTD

The per-validation normalized_prefix<->namespace mapping
is maintained by the document validator process, and only needs
to live for the validation of a single document.

If you mean the mapping that tells validator processes that
'The DTD known by "uri:FrenchLibrary" is the following ....'
then that looks like a resource management issue that can be handled
by a higher level.

-- 
==============================================================
Mark Tucker			tucker_m@regenstrief.iupui.edu
Regenstrief Institute		phone: (317) 630-2606
1001 W. 10'th St; Indianapolis, IN; 46202-2859;	fax: (317) 630-6962