This is what I’m on about when I talk about ontologies and object-orientation of knowledge. In science, the code is the knowledge. Unlike computer programming, the code is locked up PDF and XML formats, and behind firewalls and copyrights (at least in code you could write your own if you knew what you wanted.). On top of that, ontologies are a royal pain in the ass.
But it’s worth it – if the pain is big enough. Standard ontologies and common names mean we at lest get a speck of modularity in database aggregation, for example. You can do some cool stuff once you can aggregate databases.
It’s not worth doing this for the consumer web – that’s why so many people write cogent attacks on the semantic web. Small groups, tags, informal systems, google, they all work ok for the majority of humans.
But life scientists aren’t the majority of humans. They need modularity and need it now. The government is paying for open ontologies. Thus, the vendors and tools are coming. The pharmaceutical industry is laying bets. And as the geeks build enough basic validations, and the vendors make it easy, more fields in science are following.
Modularity may escape science and make it to all knowledge. I don’t know. Or honestly care so much. Object-oriented web content wasn’t needed, and I believe in the right tool for the right problem. Part of what gets me riled about the attacks on the semantic web is the dogged ignorance of the problem in science, and the dogged insistence that if we just used the web right everything would be peachy keen, wikipedias, sparkles and ponies forever.
It won’t. Science needs modularity and object orientation in knowledge, or we’ll be stuck in the knowledge equivalent of machine programming for years to come. That’d be a shame.