I’m about to agree with BurningBird (which I’m always happy to do since she’s right so damn often) but in a way that neither of us is going to find very satisfying.

IMO, she’s right to point out that something important has already begun:

My idea of semantic web is if I can look for a poem that uses a metaphor of bird as freedom, and get back poems that have bird as metaphor for freedom. But you know, I don’t have to go everywhere in the web to look for this — if I could just do this at something like poets.org, or among the poetry weblogs I know, I’d be content.

I don’t have to scour the complete world wide web today. I don’t have to get every interpretation of every poem that has ever used bird as metaphor today. I can start with a small group of people convinced that this is the way to go. And eventually, other poetry fans, and high school sophmores, will also see the benefit of doing a little bit of extra work when putting that poem online, aided and abetted by helpful tools. It’s from this tiny little acorn, big mother oaks grow.

Yes, over time we are developing schemas that make particular domains of discourse more useful, more searchable, more automatable. And, yes, over time we are hooking these together so that various domains can operate as something like a unified information space. We’re not just doing this on the Web. Without anyone declaring a new standard, business cards started including email addresses and even referring to phone numbers as V, F and C without a single piece of legislation declaring that we do so. And we map domains all the time also, as when I ask you what “C” means on your card, and I say “Aha! We call ours mobiles, not cell phones!”

So, if the semantic web means only that we’re learning to understand ourselves better on the Internet, or even that we often adopt similar terms and rhetoric, then, yes, the Web is constantly semantically webbing itself. And if the semantic web means that we are formally knitting together, in an ad hoc way, the various standards we’re adopting, then, yes, the web is semantically webbing itself.

But, I don’t think this is what most people mean by the Semantic Web. I think they have two other implications in mind.

First, they think that this semantic webbing process is going to continue until the Web is a single “information space.” But we’re not going to get close to that because ultimately the semantics of the Web is human language and understanding. And if we did get close, we’d pay a price for it: Repair manuals for aircraft are close to being a single information space because the manufacturers adopted a uniform DTD and a reduced language set. That’s how it’s done and it’s not what any of us want the Web to become. [Actually, I’m not sure they ever did adopt a uniform DTD.]

Second, the proponents of the Semantic Web aren’t simply cheering on the attempts to come up with useful domain-specific metadata standards (such as XBRL). We all like standards that help. But the supporters of the Semantic Web aren’t saying simply, “Standards are good!” They are suggesting that when these standards are put together, they will form something more than their parts. They will be machine readable and we will see marvels of automation. But history has shown us that it’s really hard to get domain-specific metadata to work together. Maybe this time it’ll happen. Maybe. But that it’s happened in this or that domain should not lead us to generalize about it happening generally.

So, I’m feeling whipsawed. Either the Semantic Web promises something grand and unifying and transformative or it refers to the growth of standards. If the former, it’s not just implausible, IMO, but is actually based on an overestimation of the ability and desirability of disambiguating language. If the latter, Shelley’s definitely right to lower-case it.

Categories: Uncategorized dw