Metadata and understanding

The solution to information overload is more information ... so long as that more information is metadata. We didn't drown in information the way the info fear mongers predicted in the early 1990s because the information--more of it than anyone foretold--didn't look the way we'd been warned. We'd been told to expect a flood of water. All water looks alike. So, the more water, the more you have to wade through to find what you want.

But information doesn't really look like water, except at the level of bits. True enough, we now deal daily with more ones and zeroes than we have commas to print out. But when some of that information is metadata, the information isn't transparent and undifferentiated like water. It's more like suitcases with lots of handles. Or like maps. Or like, well, make up your own metaphor. With metadata, we can find what we need in the oceans of bits.

The key to success was, of course, that we didn't have to make up all that metadata as a separate exercise. For example, if 100,000 people have linked to an article about bird flu, each of them linked because she or he had something to say. But each of those links is metadata. A clever app such as Google can put that metadata together and use it to influence the ranking of that page when someone searches for "bird flu." We have been rescued from a watery info death by our inadvertent metadata.

Traditionally, metadata has been used to help people find data. The metadata on a card in a library catalog is there to help patrons locate books on shelves. Metadata could do little more because metadata was a reduction of information: A 600-page book is boiled down to a few facts that fit on a 3x5 card.

But in the digital world, there's little reason to boil things down. Metadata can be expansionist. It can include the entire text of the book. It can include the text of a biography of the author of the book. It can include a map of the places the author of the biography of the author once visited. Why not? Suppose someone wants to find the book written by the subject of the biography written by that guy who used to live in Slough?

But now metadata can serve a different role. It can do more than locate lost data. It can help us understand the data that we find.

In fact, we've always relied on metadata for understanding, so the difference is in the expectations and demand for metadata. For example, if you want to understand James Joyce's Ulysses, you'd better get yourself some books about the book. And, if you hear a quote out of context and are puzzled by how someone could believe such a thing, it'll help to know that it in fact was a joke uttered by Jon Stewart. So, metadata has always aided understanding. But now we want and need more.

In part this is because of the quantity of information on the Web. We need metadata to evaluate what we should spend our time evaluating.

In part it's because the editorial barriers are down, so we are left on our own to figure out whom to listen to. To do that, we usually need more than the simple words on the screen. We need metadata about the person (or the pseudonym of the person) who wrote those words.

In part it's because we're no longer satisfied with the voice of authority speaking as if we should believe it simply because the voice said so. "I am an authority" is its own metadata, but it's much thinner than the contextual metadata we gather by learning about the speaker. We're getting more sophisticated because the new world of the Internet makes it possible for us to bring to bear the skills we honed in the real world for picking up cues about the people we hear. In the broadcast world, we had a small choice among speakers, and were given little to go on beyond the fact that the voices had made it through the broadcast filters. Now we demand more to go on. We need it if we are to make worthwhile judgments. Two conclusions: First, give us more metadata than you think we need. We'll figure out what to do with it. Efficient, controlled searches are often not as valuable as letting us wallow in metadata. Second, don't let your business speak as if all we need is the booming authority of its voice. We want to know about the people speaking. Without that metadata, we won't believe a word you say.