In favour of the sloppy semantic web
Posted May 9, 2010on:
Filtering and sorting all of the information online is a major problem for our age: agreed.
The semantic web can give context to all the data we have: coding the relationships between it, helping us to find it easier. This is a good thing, right?
But then we might consider questions about categorisation, labelling, power, agreement.
Abraham Bernstein and David Karger argue for the ‘sloppy, scruffy’ semantic web. That instead of an ontology (a formal, shared conceptualisation and associated vocabulary) a little structure can go a long way. I must agree.
Tim Berners-Lee states in this video that he ‘can’t imagine what people will do’ with the semantic web. His engineering rationale has his sights set on a robust solution, not as much on what will come after. This is perhaps understandable, given the success of the www. However when he developed the web it was with sharing in mind; in developing the semantic web, what is it? In one word? Is it control?
This isn’t an argument against control; but for remembering pluralism, even if it leads to messiness. For the alternative is the closing down of alternatives, which is worse. Crucial for semantic web engineers to consider is not only what users will do with it, but:
a) what users may not do with it
b) what users can possibly do about, what they realise is impossible
A system for linking all of the information in the world together, and making those links make sense, implies agreement on a vast scale. In constructing this innovation, openness and the capacity for its evolution are essential. Chantal Mouffe’s words seem apt to recall here, that, “no regime, not even a liberal one, can pretend to have a privileged claim on rationality.”