Further Thoughts on Filesystems

A while ago I put down some random thoughts on computer file systems. Not the most coherent thing I ever wrote, but still — it’s on my mind.

The basic problem is this: Computer files are organized the way paper files are — folders within folders within folders. But computers are different, and it seems to me there ought to be a better way to store data.

My thought for today is a simple one.

Imagine if all your data files were stored in one place on your computer — not nested folders, but a single space. Heck, you wouldn’t even have to know that it was called “My Documents” or whatever. Your data was just ‘on the computer.’

See, today we organize our files by the folders or directories they’re stored in (e.g., “2004 Taxes”) and their file names (”SalvationArmyReceipt.jpg”). That could be a problem. What if you had a bunch of files named, say, “Receipt.doc” and the only way you could tell them apart was by the directory they were in. If you lose the directory structure, you’re faced with a bunch of Receipt.docs.

Instead, every file should include a bunch of identifying information: Keywords, a description, etc. Some kinds of files already do this. MP3 files include “tags” that identify among other things the artist, song, album, etc. You don’t need to rely on the file name. Ditto for some image files; a standard called EXIF is a way of including information about images within the image file — creation date, artist, camera settings, etc.

But these are separate standards, and they don’t require much user intervention. When you rip a music file to an MP3 or snap a photo with a digital camera, this information is included automatically.

Maybe what we need is a standard, probably XML-based, data format for all files. Then they could be lumped together on your hard drive — it wouldn’t matter where — and you would find what you needed by browsing or searching on the information in those built-in identification files.

Just a thought.