Semantic Web

Sorry, I broke some URIs

By Kjetil Kjernsmokjetilk. Tuesday, May 30, 2006 2:10:47 PM

Since the launch of the Opera Community in September 2005, all members of the community has had an URI identifying them. Having URIs for everything is important, and even though you can't be retrieved over the network the URI is useful to identify you as a person.

Furthermore, we know that Cool URIs don't change, but I'm afraid I just changed them...

You see, the final, little part of the URI I gave you was very badly designed. It had a lot of unnecessary complexity, and with complexity comes bugs, and in fact, it turned out that some of them weren't even valid, since I hadn't taken into account that some usernames begin with a number.

It wasn't with a very light heart that I changed the URIs, but as I discussed this with other members of the FOAF community, it was clear that it was better to do it now, and deal with any problems that may occur with a process known as smushing, than dealing with the complexity that I had, especially as more people start using the URIs for real applications.

For most users, this change will not mean anything. The usual FOAF tools continue to work, it is only if someone has been indexing the URIs it will have any impact. I don't know if anyone has been doing that, but I would be interested in hearing about it!

UPDATE: I was a bit uneasy about this, so I found a way to at least attempt to unbreak most of the URIs, by using the sameAs property from OWL on those valid fragment identifiers from the previous version. That should give us the best of both worlds!

Field to reference external RDFPerl and RDF talk at the Nordic Perl Workshop

Comments

Trond K. Pettersentrondp # Friday, June 30, 2006 9:46:15 AM

"the URI is useful to identify you as a person"

Does it identify you or the resource available at the URI? What if I want to point to the *resource* at that URI (e.g. a HTML document, or whatever - if not a URL) and NOT to the person? Unfortunately, RDF can't help you there...

Kjetil Kjernsmokjetilk # Monday, July 3, 2006 1:28:22 PM

It identifies you as a person. If you want to point at the information resource, you use the URI of the information resource. In this case, you just remove the fragment identifier, and that's it, that's the information resource. So, RDF does all this. :-) Or, the URI work at W3C does all this. For more about this, I suggest looking through httpRange-14 TAG finding, and of course the Webarch document.

Trond K. Pettersentrondp # Tuesday, July 4, 2006 8:24:52 AM

"you use the URI of the information resource"

But what if I want to point to *exactly* that URI (the one without the fragment)? Perhaps from a different system... Anyways...I'll look into those documents...

Kjetil Kjernsmokjetilk # Tuesday, July 4, 2006 10:03:43 AM

Hmmm, I don't quite understand the question, but I think that the issue that you're getting at is the distinction between information resources and non-information resources. Indeed, that was not clearly defined until relatively recently, as correctly pointed out by Lars in one of his RDF vs. TM documents, but it is now. Sort of. People still complain... :-)