Friday, October 06, 2006

Meta Machine Music

Lately I have thinking a lot about Metadata, a term that is a fancy way to say "data about data", the kind of thing that makes libraries go. There are plenty of smart things to be said about this and the debates are many. I don't want to get into that now. Frankly, one of the things I like about blogging is it is a space to make quick and convenient observations. If you want the scholarship dig deep into my bookmarks and tags and you will get an idea of some of the debate. But what is cool abouy is that it is a metadata assembling machine. I am not the first to make this point, but one of the things I want to note is that it is a wonderful example of how metadata is not only folksonomically assembled but, like wikipedias, is an efficient manner in which cultural caital is assembled. One of the basic problems with metadata it is actually really ime consuming to create. The standards that are often debated by experts are quite wonderful, but they are also costly. For example, to gather a set of people who make these standards requires years upon years of collective education and expertise. The professionals who specialize in this spend thousands of dollars for their education just to get the right to enter into taxonomic debates. Put simply, you want to classify things in a library, you need an baccalaureaute and an MLS. Last time I looked, they weren't giving those degrees away.

The exact opposite is true when it comes to folksonomic formations of metadata. You have a computer, a general understanding of the term, and a mouse... you enter the game. The concern is accuracy, but early studies are beginning to reveal proof that this form of knowledge formation is fairly accuarate. From my perspective classifications are part of our work as humans and is always social. As I remind my students when studying genre, classifiying things is one of the powers God gives to man in Genesis. We have always understood this as a human task: men and women are interested in their nominalist potencies, indeed! Furthermore we like to do it. Making judgements and distinctions is a great social pleasure.

Which brings me to If you haven't tried it, go to it. It is a program that helps you design a "streaming radio station" by providing music and you provide rudimentary judgements. In essence, you are given a song which sounds like an artist or a song you have chosen. This "gift" is given to you based on musicological metadata which has been applied to the songs themselevs in what is called The Music Geome Project. You then make rudimentray "like it" or "don't like it" judgements These are then collected and your "station" is slowly assembled on the backend. It is a kind of collective enterprise between "experts" and us "folks" and produces an interesting compromise of sorts.

I find the whole experience with Pandora both oddly compelling but dissatisfactory: in many ways it runs counter to what I enjoy about being a fan, a facet that I will blog about later. That said, I look forward to talking with Tim Westerberg, the head of the Music Genome Project when he gets to Columbus in a couple of weeks. I can only see a project like this becoming more and more refined and interesting. It is, in many ways, one of the best examples of how music will be distributed through distributive networks.

Oh, and Pandora has a good beat and you can dance to it... I give it a 98.

No comments: