I’ve been playing with mash-ups lately for the book, and at one point had to slap myself in the face to get me to Stop! Stop! Not another service!
straup at Flickr’s announcement of “machine tags” is significant, because, as he demonstrated, it really is the same as RDF, except without the scary name (and we’ll shoot the first person who mentions reification). Of course, now I’m looking at my mash-up examples for the book and thinking, like jello, there’s always room for more.
Speaking of integrating services and data, I still like RDF as XML. I can do things with it, such as load it into the browser XML Parser and manipulate the data using DOM methods. Unfortunately, I have to copy the RDF file, such as Dan Brickley’s FOAF file, to my home directory before using Ajax–it’s not packaged correctly for cross-domain browser access. It wouldn’t be difficult for any RDF/XML source to be packaged as end points for cross-domain access. Leaving aside issues of trust.
Danny, who points to a nice semantic/scripting challenge (but…iPod?), asked about RDF Turtle notation and Ajax, and sure you could use Turtle in XMLHttpRequest (XHR) requests, or as endpoints and dynamic scripting. All you have to do is either return it as text for XHR, or as a valid parameter in an endpoint (wrapped in a function call, and used with dynamic scripting). What we need, is an transformation between Turtle and JSON, and return Turtle formatted as JSON (and we have it). But I like RDF/XML because I can just cram it into the browser’s parser and use the DOM. Either XML or JSON works for me.
The Flickr API’s “machine tags” works, too, basically flattening triples and squeezing them skinny thangs into a JSON response; The API provides an endpoint, too, so that you can call it from the browser. If you’re as curious like me, as who would use the dc: namespace at Flickr, click the dc: button in the example page I linked earlier, and you’ll see the most recent cases. From the pictures, it looks pretty much like everyone.
Let me say that in the crowded field of photo services, Flickr just got all pretty and sparkly, and is still *Queen of the game.
Sparkly…sparqly…say…that gives me an idea…
Yup. There’s always room for more.
Update Try out the end button, which pulls in the dc:subject from my RSS 1.0 file. Click on an option, and it searches for all the matching photos in Flickr. Of course, I’m the only one who has used dc:subject with Flickr…still.
Quick note: The example application I linked works on most browsers, but this is just a quick hack, for fun. It hasn’t been heavily tested other than me playing around, nor have I optimized the code. I haven’t tried it on IE 6.x or IE7 yet, me having ‘fun’ being the operative concept in this paragraph.
Bonus points: Kingsley Idehen: SPARQL, Tagging, Ajax,…
*What, you thought I was going to say King? Don’t know me well, do you?