Alex wrote this article about "everything buckets" that got me thinking. And you know how dangerous this can be.

I've been thinking about "everything buckets," as "information buckets," and I think this term is a little more apt, given how these systems work. Basically the "everything" bucket, is a program that provides a database interface on top of our files in the effort of keeping us better organized. Features common to the information bucket, genre of programs are meta data and tagging interfaces, easier/integrated editing environments, more advanced search abilities, and easier input features. All wrapped up in one nifty package.

Alex's article makes a rather important critique of these programs, and one with which I mostly agree. He argues that the "buckets" themselves are non-free/closed source (bad for your data,) and that they abstract the organizational problems of organizing files rather than actually resolve those problems. Alex suggests that the best way to deal with this is to give up learn how to organize files on the file system itself and use better tools to store data.

Pragmatically I agree with him: I use structured formats to store most of my data, and have a simple but effective means for storing my data. It works, and I'm sure if you spend enough time talking about information management stuff with me, you'll hear more than you ever wanted to know about how I store and use information.

But the "information bucket" approach has a lot going in it's favor, and my approach isn't a really broad solution. Here's what's good about this approach:

1. If you store your data in one bucket then there's only one place that your data could be, which makes it much easier to find any given thing when you're trying to look for something that you just know you saved somewhere.

2. If we're left to our own devices to develop structured data a couple of things happen. First is that we don't give it enough structure, so we end up with a few hundred files in a directory and no good way to make heads or tails of them. Then, to compensate for this, we create folders has a way of bookmarking specific bits of data, and end up with (potentially) too much structure that doesn't convey productive information.

I agree that the OS X information bucket isn't an ideal solution to this problem, but I think there is a substantive problem here that need some sort of unified solution. I'm not sure what that unified solution is, or even if there's going to be a one-size-fits-all response for all users. But I think it's a question that we need to begin to think about.