Another interesting interview over at ACM Queue: Tim Bray, mostly on XML and RDF.
The XML serialization of RDF is horrible; it’s a botched job.
Amen. Shame that.
[RDF] quickly became mixed up with a whole bunch of classic KR (knowledge representation) people who wanted to go refight the AI wars of the ’80s. … You know, KR didn’t suddenly become easy just because it’s got pointy brackets. Doug Lenat has been off working in the desert on that for decades and nobody has ever made a buck on it yet, as far as I know.
There surely has been a buck made on KR, albeit not on Lenat's work (the Cyc project). Representing a significant amount of human knowledge formally is way too expensive to give a positive ROI, perhaps, but formal logic based languages have great promise in far more situations than they are currently used in. The current state of RDF is that it is too limited for it to be of much help there, and this is a pretty unsurprising result of its history. What was supposed to be a language for encoding metadata has been dragged into being a KR language, but it remains a seriously limited one. In the mean time it attracts a load of funding at the expense of other work.
Limitations of RDF as a KR language are limitations, too, as a metadata framework. After all, what use is a metadata framework intended for deployment on the Internet that can't represent provenance of data? This is being worked on, of course, but if RDF had been based on an existing KR language which supported multiple contexts, it wouldn't be an issue now.
Comments