About a ago I took the now traditional RDF-newbie step of getting het-up about HTTP-14.
Anyhow, I’m older and wiser (or at least more tired). One thing that did stick with me from the above-linked ramble is that it would be really useful to have a neat and simple way to refer to the topic of a page. Many normal humans don’t get data issues like URIs, unique-primary keys and jokes about “This is not a pipe.”. I’m interested in ways to enable stuff for the 99% of web users who are not data nerds.
Part of the thinking on this was a good friend who suggested that we should just give up on distinguishing between URI and document URL and let common sense figure out what we’re talking about. I kicked around some really dumb ideas of extending the URI scheme in some way to do this. Today I hit on a much much more simple solution.
Here’s the URI for a music video I like:
http://primarytopic.net/?url=http%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D4nigRT2KmCE
of the species Womp Rat, which doesn’t rate a page to itself in wikipedia so doesn’t get a dbpediaURI:
http://primarytopic.net/?url=http%3A%2F%2Fstarwars.wikia.com%2Fwiki%2FWomp_rat
me:
and more usefully, the SVG tutorial at the 2002 WWW Conference. This one actually indicates the primary topic of a fragment of a document. That works fine.
OK. You can mint your own URIs, but this means less URI proliferation than each service minting URIs they don’t really need to be resolved to their own triples. I reckon this will work very nicely with http://sameas.org/
My anticipation is that it’s useful to ask people for data about, say, an event or organisation by asking for the URL of the thing and then using this to construct a URI. If someone else is building data using the same system then your data links. Which is what we’re about, right?
Suggestions very welcome. I’ve kept this dead simple for now as it doesn’t need to be comlicated. To save everyone license headaches I’ve made the whole thing CC0.
It’s a great idea. The only thing that would really annoy me to use it is the nasty looking URIs. If I was hand-coding something somewhere along the line, the last thing I’d want to write is all that URL encoded nonsense. I know there’s the tool at the website but it’s still a hassle. If there was an extension that allowed ‘nice’ URIs (e.g. http://primarytopic.net/WompRat) that redirected or something, it’s be great. Still a good idea, though.
It also turns out the idea isn’t new, and was set up at http://thing-described-by.org/ but nobody seems to use it or know about it.
If you want to make a tinyURI service, be my guest, I’ll happily set it up if you code it, but it doesn’t solve any problem I currently have. 🙂
For me it defeats the point which is letting 2 people mint the same URI seredipitusly and skipping a sameAs step.
I was thinking that for certain well known sites, it could add a sameAs to the more commonly used URI. eg. twitter->semanticTweet, wikipedia->dbpedia etc.