- From: Michel_Dumontier <Michel_Dumontier@carleton.ca>
- Date: Wed, 08 Aug 2007 13:53:31 -0400
- To: "Booth, David (HP Software - Boston)" <dbooth@hp.com>, wangxiao@musc.edu, ogbujic@ccf.org
- Cc: public-semweb-lifesci hcls <public-semweb-lifesci@w3.org>
> > Sure, HTTP URIs can be used as identifiers, but why would I > > mint arbitrary HTTP URIs when I can use a scheme that has no > > resolution protocol implicitly or explicitly associated with > > it? > > Indeed, in this case there would be no benefit to using HTTP URIs. But > the situation you describe is also not a practice that we should be > recommending, because it inhibits the use of those URIs by others. That's not at all true. I don't need the web (or http URIs) to share or reason about RDF/OWL ontologies. I publish non-http URIs in my OWL documents for others to use, make reference to, and integrate data with. Don't be fooled by thinking that URIs need to be dereferenced. It has its uses, but it also requires institutional commitment, and with all the broken links on the internet, I hardly believe that commitment exists. -=Michel=-
Received on Wednesday, 8 August 2007 17:53:54 UTC