On iTunes and Data

by F.

There’s a remarkable change taking place in the way we conceptualize information. In the old days, we thought in terms of containers. Something either goes into this box or that one. Now we think in terms of tags. We can put something in this box, and that one. I think some of the explanation for the transformation is iTunes.

Time was, no one manipulated large amounts of data. We didn’t have “personal databases.” At work, we might hook into a SQL database using some shitty client that would talk to the database and then barf back at us some long, seemingly random pile of information. We might have formed some basic queries using SQL or boolean operators. But most folks found that stuff too annoying to learn.

But then we started downloading music. Your iTunes collection is just a big database, which you slice and dice in all sorts of ways. Playlists and Smart Playlists are the two primary ones. Smart Playlists teach you that a container is equivalent to an attribute. That is, suppose you group songs into a folder called Recently Added. The contents of that folder are a function of an attribute (and a “collection operation”): “songs that have been recently added.” Call that attribute A. The property of having A defines the members of the folder. Everything A is in the folder; things without A are outside it. But—critically—the folder can contain things with other attributes B, C, D, E…, too.

Let’s make that more concrete. You might have another container (Smart Playlist) that holds Highest Rated songs. Some of the songs that are highest rated are also Recently Added. They go in both containers, yet no “duplication” is required. It’s just one song in two places. (The old solution to this was creating an “alias”–but that kept the container metaphor in place, beacuse we had to “clone” things to put them in multiple boxes.) This breaks the physical container metaphor, because in the physical world, things either go here or there. (Think of a bookshelf.) Not both. But information is abstract. It frees us from some of our physicality.

The liberating thing about this new metaphor is flexibility. You realize quickly that there are an arbitrarily large and equally valid number of ways to organize data. What you think is “valid” depends on your goal (highest rated? recently added?) You can slice and dice data (for instance, songs) all you want, seeing new patterns, connections, similarities, differences.

It’s really beautiful.