Re: Syntactic extensibility

On 9 Sep 2008, at 15:29, Sandro Hawke wrote:
[snip]
> I think this goes against WebArch.

Even if so...so? :)

>   Folks supplying an extension should
> provide useful material to folks dereferencing the URIs used in the
> extenion, but they can't with this approach (unless W3C were  
> willing to
> be a registry of these x- extensions).

Wiki page?

Let's consider the alternative...we have URIs coined by extension  
writers that....tools can't specially recognize but people can look  
up. But why would they look them up if they didn't have their  
attention drawn to them some way?

(Consider the baleful harm of they way SWRL has worked, with  
implementors thinking they needed to import the SWRL ontologies, etc.  
etc.)

As a user, I'd rather get the warning, and if I did look it up, find  
out that it was an extension, then google for the extension or ask on  
public-owl-dev. (I think of search engines and mailing lists as part  
of how the web works :))

I believe, also, that RDF only requires a warning if you add terms to  
the rdf namespace...so people *can* do this already, but we have no  
canonical way of indicating what's a random extension and what's a  
sanctioned extension.

Cheers,
Bijan.

Received on Tuesday, 9 September 2008 15:15:53 UTC