Friday, June 03, 2005

I wonder how many of the BBC's Backstage prototypes are using RDF.

Unfair question of course. Because if the BBC aren't using it, then however much demand there is for RDF it can't be done. OTOH, if they are, then potential consumers will have to use it.

At first glance at the feeds available there seems to be a mix. Some are RSS 2.0, some are RSS 0.9, some look like custom XML chunks. So maybe RDF will enter the mix (possibly via RSS 1.0)

I guess this is another way of hingeing the semantic web debate.

The anti-SemWeb intuition is that content trumps format. In other words : content is made available, and if there's someone who wants to take it and combine it with something else, they will pay the price of scraping, or take responsibility for the semantic inference from the original meanings to their meanings. They will pay for "interpretation"

Furthermore, if there's no content, and no desire, then the ease of combining, has little importance. The decision to combine is made by the consumer based on the value of the data. This is pretty much the message of interop can procede without pairwise co-operation.

The semweb intuition is that the interpretation must come first. That the meaning that will be shared between producer and consumer should be decided by the producer, at the time of production. And this will make the cost of re-use and combination so cheap that all kinds of new applications will be able to spring up.


Let's take that economic metaphor a bit further.

Any economic transaction has a transaction cost.

Now, that cost includes all kinds of things : discovery of the right partner, possibly including competitive signalling of fitness, co-ordination with that partner etc.

Now, in the case of a "semantic" web (lower or upper case), one of the costs is the acceptance of responsibility for interpretation. Who will decide that my "John Smith" is the same as your "John Q. Smith"?

Can that responsibility be modelled as a cost?

Well, there's clearly a cost if it's wrong. And who will suffer (pay the cost) if the semantic interpretation is wrong? The consumer.

This seems to put a strong bargaining position in the hands of the producer. It's a case of caveat emptor. The consumer of the information will need to pay the cost of giving semantic interpretation to it if the producer chooses not to. If the product is compelling, the user will pay. And non SemWeb formats will proliferate.

What changes this situation? Competition.

As the number of producers increases they will be competing for consumers. Part of that competition will be the value of their content. But as the balance of power shifts, convenience for the consumer will become important. Well marked-up, and particularly, semantically marked-up, data which is easier for the consumer to process will start to differentiate the producers.

And then the consumer can push the cost of "interpreting" back up the supply chain.

Now, obviously, this only goes if there's some benefit to the producer of getting consumed.

Let's have another look. We'd expect the semantic web to take off where there is competition among producers and an incentive for them to want to be consumed.

Clearly RSS is one of the most evolved "markets" for information. And here we see two kinds of RDF getting established : RSS 1.0 and Atom. The only other application of RDF I can think of which has generated much public excitement is FOAF. Once again, it's in competition (with Friendster, Tribe etc.). And once again there's some incentive to be consumed : you get more famous and meet more people.

Where am I going with this? Not entirely sure, but here's the obvious, although slightly surprising, conclusion :

Ads in RSS will drive the adoption of the Semantic Web

Think about it. Ads in feeds are just the beginning. Soon, there'll be all kinds of advertorial spewed out in syndication format. At which point the advertisers will have a huge incentive to make sure these chunks of text are routed to the right places. They'll need to be marked-up more smartly, they'll need to maximize the ease with which they can be cross-reference to other information so as to survive longer in the noosphere.



No comments: