Do Mac and iPhone users really need a file system?

Many Mac power users and developers are concerned about the iOS-ification of the Mac OS and how much more of that process will be found in sessions at the Apple Worldwide Developers Conference (WWDC) next week in San Francisco. A recent developer blog post recalls talks by Steve Jobs from a 6 years ago, where he speculated on whether computer users really needed to interact with a file system. But to me, it also recalls a mostly forgotten Apple OS from around 20 years ago.

Many Mac power users and developers are concerned about the iOS-ification of the Mac OS and how much more of that process will be found in sessions at the Apple Worldwide Developers Conference (WWDC) next week in San Francisco.  A recent developer blog post recalls talks by Steve Jobs from a 6 years ago, where he speculated on whether computer users really needed to interact with a file system. But to me, it also recalls a mostly forgotten Apple OS from around 20 years ago.

In his blog, iOS developer Ole Begemann, tells of watching a 2005 video of Jobs speaking at the D: All Things Digital conference. (The videos were recently posted to iTunes by Walt Mossberg and Kara Swisher.) Part of this talk was about the then imminent release of Mac OS X 10.4 Tiger. Jobs discussed Spotlight and e-mail, and the use of search to uncover the files we need instead of folders and directories. I recall that Jobs gave this same spiel elsewhere — I heard him go on about this at WWDC (or was it Macworld?).

Begemann transcribed some of Jobs' talk:

In every user interface study we’ve ever done […], [we found] it’s pretty easy to learn how to use these things ‘til you hit the file system and then the learning curve goes vertical. So you ask yourself, why is the file system the face of the OS? Wouldn’t it be better if there was a better way to find stuff?

Now, e-mail, there’s always been a better way to find stuff. You don’t keep your e-mail on your file system, right? The app manages it. And that was the breakthrough, as an example, in iTunes. You don’t keep your music in the file system, that would be crazy. You keep it in this app that knows about music and knows how to find things in lots of different ways. Same with photos: we’ve got an app that knows all about photos. And these apps manage their own file storage. […]

And eventually, the file system management is just gonna be an app for pros and consumers aren’t gonna need to use it.

Begemann said this sounds in many ways like iOS and its relation to users and data. The app manages it. "Anything else would be crazy," he suggests. (Check out Begemann's full post, which has the time mark for the quotes.)

Check out: Will Apple's Success Kill the Mac As We Know It?

However, to my ears, it also sounds like the object-soup file approach that Apple used in the Newton OS, which was first released with the Newton MessagePad in 1993, almost 20 years ago.

Here's a selection from Newton — Something New Under the Sun, an article by Andy Gore from the August 2, 1993 issue of MacWEEK. See if it sounds familiar:

At the heart of the technology is what Apple engineers call an "object soup" - a massive collection of data "frames" containing information, such as names, addresses, an appointment or a sketch.

Within this pervasive object system, all data is created equal. Information within one frame can be related to any number of other frames, each of which in turn can be related to even more frames. This way, data is inherited, avoiding duplication, and each object can be referenced in any manner the user or developer wants. The specific data contained within a Newton's soup is called a data set.

Newton applications are basically forms that give the user controlled access to a data set. They can add, modify and delete data within frames, and because all Newton applications share the same data set, they can do the same to references between frames.

There are no data-access boundaries between applications, and even Newton's built-in applications are built on top of this framework. That means developers can reuse code that already exists within Newton's ROM and can readily add to or even modify the behavior of that code.

Newton doesn't care about what a data object is or where it is located in memory. For example, there are no memory pointers in NewtonScript, the programming language developers use to build Newton applications and Apple's own engineers used to create the Newton interface.

One way of looking at how Newton manages information is to compare it with someone who has a sloppy office but always knows where everything is. When you look into such a person's work space, all you see is chaos. But if you ask for a memo written several weeks ago, that person can all but instantly dig it out from among the piles of paper. Newton manages information in a similar "controlled chaos" manner.

Sound familiar? Yep.

In an interesting way, iOS is similar and very, very different than the Newton approach. Yes, apps give users access to data and files without giving them control over where that data is stored. But under iOS, the sandboxed apps don't have access to anything but their own data.

Note: For the Newton and OOP freaks out there, I am aware that the Newton was really a pre-emptive multitasking information environment, and not strictly speaking, an operating system. Excuse me!