Comments on 28022000 draft: Lack of expressiveness in DTDs

I would like to discuss again the following problem regarding the
(lack of) expressiveness of the DTD content model of XML 1.0:

In our "Object" element (and in some other places) we allow any 
well-formed XML element, which need not essentially be defined in
the grammar given by the XML document's corresponding DTD.

But how should we express this "freedom" with the limited means
of the DTD content model for an element?

* (#PCDATA) limits the allowed content of the element to character
  data only; no tags are allowed.

* (ANY) limits the allowed content to any elements DEFINED WITHIN
  the DTD; so this is also too restrictive.

In opposition to these restrictions, XML Schema allwows us to declare
exactly the content model we have in mind by using the <any> wildcard
element (see [1]).

Possible solutions (?):

* Skip DTD definitions and only use Schema definitions (very radical
  suggestion, I have to admit)

* Incorporate some clarification into the draft that our intended content
  model for "Object" (and some others) cannot be expressed by the limited
  means of XML DTDs, but only the approximation which can be currently
  found in our DTD definitions. (I have seen a red-colored sentence in the
  introductional part of the 18022000 draft, but it has been removed again;
  cf [2])

In any case, I think (ANY) would be a better approximation than (#PCDATA),
which is currenlty used.


Gregor

----
[1] http://www.w3.org/TR/xmlschema-1/#declare-openness

-- 
---------------------------------------------------------------
Gregor Karlinger
mailto://gregor.karlinger@iaik.at
Institute for Applied Information Processing and Communications
Austria
---------------------------------------------------------------