Re: Agenda for TC 2008-06-04

On Jun 10, 2008, at 11:02 PM, Michael Schneider wrote:

> Ivan Herman wrote on Wednesday, June 04, 2008:
>
>> 3. If we decide to have a separate XML Namespace URI, I also proposed
>>
>> http://www.w3.org/2002/07/owl/xml
>>
>> as an alternative which, though a bit long, makes the relationship
>> between the two much clearer.
>
> This URI looks fair to me.

Fair?

> It's clear that OWL's XML syntax is meant. And if
> someone knows the OWL URI, then it's easy to remember what the XML  
> URI is.

Yes, but please remember it's not just remembering, it's learning,  
understanding the difference and then having to use at least two owl  
prefixes (owl and owlx, or default and owl...x) and possibly entities  
as well (in addition to rdf, rdfs, and xsd...hork).

> What I would like to see in the longer term is some general naming  
> conventions
> provided by the W3C, which help working groups to rapidly find  
> proper names.
> This whole discussion about this XML URI takes much too long. And  
> it is by far
> not the only naming issue we have to resolve...

Might be nice, but I'm skeptical its going to happen (or that I'd  
like it if it did :)). The TAG does a bit of this, but, if I may  
speak a bit frankly, there are different groups who want to make the  
namespace (and similar) URIs technically significant in a variety of  
ways that does not have broad consensus (e.g., GRDDL; it's not just  
me...most members of the W3C, I'd be happy to bet, paid no attention  
to GRDDL; you regularly see GRDDL denigrated (along with other  
heavily evangalized semweb tech, see <http://krijnhoetmer.nl/irc-logs/ 
whatwg/20080606#l-361> for a recent example)).

In simple point of fact, proliferating namespaces is relatively high  
cost. Software has to be revised; users (and developers) have just a  
bit more added to their conceptual load. I still, frequently enough,  
get people who think RDF *has namespaces* (it does not). There is a  
tendency to think that adding another namespace (or uri prefix) is  
the default option. It shouldn't be. Anytime to add or change  
fundamental syntax in a non-explicitly functional way you run a big  
risk in imposing benefitless costs (often when the cost isn't paid by  
you!). Thus, such changes should require high bar justification.

(This is one reason I've never suggested moving *everything* into the  
owl or to a new owl2 prefix, i.e., subClassOf, range, domain, type,  
etc. In a very clear sense it would be "better", certainly easier (if  
considered in isolation). But transition and ongoing interop cost  
make it utterly prohibitive.)

Cheers,
Bijan.

Received on Tuesday, 10 June 2008 22:52:42 UTC