The concept of a simplified, abstract filing system sounds great at first, until you actually try to get some serious work done with one. Forcing all files to be managed in a proprietary database, grouped according to their respective applications and barring the use of nested folder structures doesn't work so well in real life. A few common scenarios come to mind...
What happens when you have more than one app capable of viewing and editing the same file type? Obvious example: image files. Imagine your iPad is loaded with iPhoto, Brushes, and a couple of other apps - each capable of viewing and editing image files. Theoretically, each app's files will be stored separately. So that means that when you want to browse for an image, you have to browse across multiple file stores in search of just the right one. Not so simple and elegant anymore.
These demos always look great with modest file libraries that have been automatically organized for the user. But this is not how real life works. For example, when I'm working on a big Keynote presentation, I need to collect, create, and edit a variety of files from various sources. These will include photos, screenshots, tables, graphs, video clips, etc. At this point, I couldn't care less which application was used to create or edit each file. The only thing I care about is the subject matter. With the large number and variety of files involved, it makes sense to group those files together in a working project folder, away from unrelated files. But with a "simplified" file system, I'd be forced to continuously hop around various application file stores and wade through thousands of unrelated files in search of what I need. Needless to say, this is completely unworkable.
Data Migration (A horror story)
Let's step away from the iPad for a moment, and examine how well managed file systems work on a real computer. I am reminded of the time when I bought my first Mac, and the first thing I wanted to do was to browse my photo collection with iPhoto. This collection consists of tens of thousands of photos, all nicely organized into file folders according to date and event.
My first shock came when I discovered that iPhoto was incapable of simply browsing my existing image collection. My photos first had to be "imported". My second shock came when I realized that importing all my photos would take many hours. Third shock was the realization that iPhoto didn't do a very good job of recognizing and preserving my simple and straightforward folder system. iPhoto organizes images by "Events", and it treats each imported batch of files as a single event. This means that in order to keep my existing photo collection intact, I had to import them one event at a time. Nice. Additionally, iPhoto only displays events as a collection of thumbnail tiles, without a "list view" option. If you want to view your events as a list - you guessed it - you now have to create an "album" for each event and view & sort the albums separately.
At this point I am so overwhelmed by iPhoto's "simplicity" that I'm tempted to hurl my Macbook Pro towards a fast moving freight train.
Now some of you might think that these hassles are a worthwhile, one-time task justified by the advantages of using iPhoto. I thought so too. And I kinda liked how iPhoto allowed me to import and browse my camera's pictures and video clips together. That is, until the release of Aperture - a better, more powerful image management and editing application.
And then I found myself right back in square one. Aperture, like iPhoto, manages images in a proprietary database structure. This means that in order to reap its benefits I needed to (you guessed it) import my iPhoto library into Aperture. But wait, it gets better. Turns out that Aperture, unlike iPhoto, does NOT support viewing of video clips. So if I migrate my photos to Aperture, I lose the ability to group related pictures and video clips together. Oh joy!
The moral of the story is that so-called simplified file systems are anything but simple. They're inflexible, impractical, and in the long run will cause far more aggravation and wasted time then can be justified by the initial illusion of simplicity.