Dare Obasanjo aka Carnage4Life - Questioning RDF
In a recent post Dare Obasanjo writes:
Uche is one of the few RDF heads whose opinions seem grounded in practicality (Joshua Allen is another) so it is definitely interesting to see him begin to question whether RDF is the right path.
I completely agree with both statements:
- Uche is one of the few RDF heads whose opinions seem grounded in practicality
- it is definitely interesting to see him begin to question whether RDF is the right path
In fact just this last weekend Uche advised me on a particular topic regarding a sub-topic related to the semantic web and as such I am now indebted for two conferences worth of drinks. Now I have no idea how much Uche can drink. If he's anything like me that wallet of mine is in for a good 'ol fashioned whoppin... [UPDATE: It occured to me that this might come across the wrong way... I "owe" him two conferences worth of drinks not because that was the "deal" go into seeking his advice but because I felt compelled to show at least some sort of gratitude for all the time, energy, and potentially gobs and gobs of cash he saved me with his advice. Uche always has and I am sure always will help a fellow hacker understand things better if the opportunity presents itself to do so and doesn't have any sort of "now you owe me" attachments to anything he might do or say. Just wanted to make sure I didn't put his advice into the wrong light... Uche is a *VERY* giving person as I'm sure many of you are aware.]
But its a whoppin I am willing, able, and most definitely happy to take compared to the whoppin' it would have taken had I not looked to him for advice and received it with the open honesty and willingness that I have come to truly appreciate in Uche. As Dare states, Uche has his head grounded in practicality. He's one of those hackers in whom when he speaks, I simply shut-up and listen because I know what he has to say is based on experience and understanding, both of which come from working in the hacking trenches, building the technologies that many of us have come to simply take for granted.
Moving on to the next paragraph in Dare's post he states:
I definitely think there is some merit to disconnecting RDF from the Semantic Web and seeing if it can hang on its own from that perspective. For example, XML as a Web document format is mostly dead-on-arrival but it has found a wide variety of uses as a general data interchange format instead. I've wondered if there is similar usefulness lurking within RDF once it loses its Semantic Web baggage.
Hmmm... Well, either I don't understand what he means by "XML as a Web document format is mostly dead-on-arrival" or I completely disagree. But its not the kind of disagreement thats devalues the content in the rest of his post and I have often and quite recently had the same sort-of thought in:
I've wondered if there is similar usefulness lurking within RDF once it loses its Semantic Web baggage.
Here's a crazy idea that I posted to the LLUP working group list a week or two back regarding the ‘atom:category’ element which allows for the attributes ‘term’ ‘label’ and ‘scheme’ without definition for what these must be:
Has the Atom working group thrown in a possible RDF-replacement easter egg given that <atom:category term="foo" label="bar" scheme="http://foobar.org/Hello_Resource+Property+Value">MyCategory</atom:category> could easily be used to represent all three of these values with the added bonus of a text node made available between the start and end tags of this element that can be used as the basis in which the contained attributes can be applied?
Don't get me wrong. I'm not making accusations that anything evil or backhanded has taken place. In fact just the opposite. In a world where many of us can see great advantage to the RDF-styled RPV and yet great disadvantage to the implementation to RPV that the current state of RDF has put forth, it seems that the Atom Syndication Format could have contained within it exactly what is needed to do RPV right, ridding itself of the complexities contained within RDF, focusing instead on simplicity. In other words:
Could <atom:category term="foo" label="bar" scheme="http://foobar.org/Hello_Resource+Property+Value">MyCategory</atom:category> give us a good starting point in which we can start over with a new RPV-based description language, from scratch, this time based on a solid, well-defined syndication format that, when implemented according to the specification, dump's the loosey-goosey'ness of RSS and gives us a solid foundation in which has the potential to last for generations with need for little, if any, changes? Furthermore, can this be be done in a simple, easy to understand and implement way using the above mentioned atom:category element and its associated attributes?
I think it can.
[NOTE: Think of it this way: Isn't the general idea driving the semantic web built around bringing context to the content in which it is associated with? And isn't "category" a simple and easily understood term in which we can use to "categorize" the content that is represented, using either a pre-defined categorization such as what we find in DMOZ or something completely dynamic, similar to the way Technorati has made popular use of the term 'tags'?]