xml-dev - Re: Elements cannot be described more than once in DTD, right ?

[Len Bullard <cbullard@hiwaay.net>]

Pavel Velikhov wrote:
> 
> Well, theoretically they are pretty serious problems. If you want to
> take a union of DTDs, say when integrating two XML databases, you might end up with a DTD that
> allows a lot of junk elements, while the original DTDs were strict enought. In the
> database setting, when the DTDs are used as database schemas, this leads to all sorts of
> problems.

Maybe.  I don't take DTDs that seriously and that doesn't mean I don't
use 
them.  Take the example of combining two databases.  Look at the
relational 
representation.  To update or append, you must first map names then
identities.  
When you map names in XML, the namespace works.  Why?  Because an
instance 
aggregates fine as long as the local content model is valid.  The
namespace
identifier does its job of uniquefying the string values of the element 
and attribute members.   So far so good.  

We tend to regard DTDs as static defintions, monoliths.  In effect, 
if we use aggregation in instances and still want to use DTDs, we 
have to consider DTDs as dynamic structures which modify their 
productions to cohere with the aggregate instance.  This reciprocity of 
instance and defintion will indeed make the DTD slacker or tighter 
with regard to any instance of itself but not of the validatible
instance.  
In essence, the evolution of the instance and the DTD are tightly 
coupled.  That is useful.

The problem is not creating the database, it is keeping the 
schema up to date with the forms of the instance which evolves 
in response to process.  If processes are scheduled and the 
behaviors are well-formed, then the DTD changes according to 
regular productions which are known.   What is useful then 
is to define those productions and by them, order the evolution 
of the DTD in accordance to the requirements of the enterprise.

len



xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev@ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ and on CD-ROM/ISBN 981-02-3594-1
To (un)subscribe, mailto:majordomo@ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo@ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa@ic.ac.uk)