Re: RDDL: new natures

  Elliotte Harold wrote:

> Jonathan Borden wrote:
>> After alot of useful discussion on the TAG, it has been pointed  
>> out to me that some of the URIs that we originally selected for  
>> RDDL natures don't make sense. In particular we were using  
>> namespace URIs as the RDDL natures of things whereas the RDDL  
>> nature of something is really a class or group that it belongs to.
> I must have missed something. Why is this considered necessary? The  
> namespace URI seems like a perfectly natural way to identify a class
> or group that it belongs to.

I thought it was perfectly natural also.

This is why it is problematic:

A class has a (i.e. one) set or group of members. If a namespace URI  
identifies a class then what is the set of members:

a) the set of names in the namespace (for example)
b) the set of documents that validate to a given schema (for example)

If we are using namespace URIs as natures then these two get confused.

For example:

does this identify EITHER

a) the set of names in the XML Schema namespace


b) the set of possible valid XML Schema documents.

>> In response to the TAG request I've updated 
>> natures to deprecate the old nature URIs and suggest new URIs.  
>> This is all of the form:
> I wish there were a more formal procedure for updating RDDL. Even  
> if I were convinced that using namespace URIs as natures was a bad  
> idea (and I'm not), I still wouldn't want to change the natures at  
> this late date. I'd like to raise a formal objection to this, and  
> request that at least the existing RDDL natures be maintained as  
> is, unchanged, and without any new values for the same natures.

The procedure has been to discuss this on XML-DEV. More recently  
discussion of RDDL has taken place on WWW-TAG. These discussions have  
gone on for over two years including these specific issues. I can  
assure you that no changes to RDDL are being done rashly nor quickly.

Furthermore as has always been the case, the RDDL natures given in are merely guidelines or a place to put  
"well known" natures.  Nowhere in the RDDL specification (http:// does it prevent you from using the "old" natures --  
although I suspect that the TAG may have something more pointed to  
say about this.

In any case RDDL natures have never been intended to be a fixed set  
of values.

As always I am willing to listen to discussion.


Received on Sunday, 10 December 2006 14:01:19 UTC