or "Aristotle's lost second book of Poetics: Comedy"
Alot of people have been crowing about how Flickr's Machine Tags are pseudo Semantic Web and how they could have been RDF.
Well, piddle, and tough luck. I'm quite excited by it all and here's why.
Machine Tags are essentially just words: strings of characters. They make no attempt at assigning meaning or giving Semantics, e.g.: they are words that you and I can use and interpret. Differently if we want to, together if we decide to agree to.
Just like we do when we speak or write to each other.
Ambiguity: the devil's volleyball. I spend my life wrastling ambiguity. It's a good thing.
What makes them slightly different from regular tags is that they allow one to add just a touch of context and thus a touch of disambiguation. Just enough to make them transportable (out of the Flickr domain) with slightly more ease and value* than single dimension tags.
flickr/tag/apple is local to flickr
flickr/tag/fruits:name=apple is still local flickr but:
fruits:name=apple is far more useful out in the enormity of the world wide web then just
apple
Flickr does not want to be "L'Academie Française", or the Encyclopedia Britannica. It's up to you (er us) all to be Wikipedia.
Back to my point though: Meaning is contextual.
Machine Tags are no more Semantic Web than English is Funny. They both *can be*...
(A major mechanism of humor is meaning derived from context, and not the words themselves. The delivery, the story/situation surrounding them, etc.)
Now before the SemWeb'ers and RDF'ers jump on me.: yes, SemWeb accounts for providing context via complexe namespacing URI schema mumbo jumbo voodooh la, but hell: no one *got* it. "dear users, please just get used to angle brackets" is a pipe dream. So calm down and watch what happens next. ;)
All Flickr has done here is given us an example of usage (with a leapfrog over the ramping up part of an adoption rate curve), a place to store them and a means to query for them.
What you do with it from there is your business. Make RDF with it if you want, but I am betting most hackers will simply follow suit and implement their own systems for storage and retrieval of namespace:predicate=value tags. And a bunch of other stuff they will have learnt to want to do when they tried to do anything with RDF (cause RDF IS awesome, it's just a P.I.T.A.). They already are. Some already have been for a while.
None of this means RDF is bad or dead. It's sort of like "do we use C++ or PHP for our web project? Hrmmm, depends!"
Flickr is already a, if not the, Golden Child of the so-called "Web 2.0", but in launching Machine Tags, they have fired the starting gun on what I'd hate to see named "Web 3.0": the lovechild produced when the dumbass naive teenager that is Web 2.0 grows up a bit and meets a Semantic Web that's learnt to relax and slip into something sexy...
Make. Whoopie. It's fun and healthy.
* ease and value: what happens when you lower the barrier to entry and increase the value of what's behind that barrier? The crowds come running. Identify what the masses might find valuable (or take a flying guess) and make it easy for them to get. Then install a toll booth. Profit.