Apple makes a more complicated interface. On purpose.

I've got a new piece up on Slate today, about Apple's iApps, Microsoft's Longhorn (the rumored next generation OS), and the issue of interface uniformity. It's basically making the argument that by bundling applications like iPhoto and iTunes with the Mac OS, Apple is implicitly making the argument that the "one interface fits all" model doesn't work for organizing some types of data. In other words, once you've organized your music collection with iTunes, or your photos with iPhoto, you'll never go back to organizing them via the Finder again. So it becomes a swiss-army-knife approach to file management, rather than a one-tool-fits-all. Interestingly, Microsoft seems to be going in the opposite direction, at least with its desktop interfaces. (The Media Center interface, designed to be manipulated by a remote control, is another matter.)

One interesting analogy here, which I'd originally included in the piece, is the way this struggle mirrors debates about the brain: is the brain a swiss-army-knife of specialized modules, or is it more of a general intelligence? The evidence right now certainly makes it seem like evolution selected the modular approach. It will be interesting to see if our interfaces follow a comparable path.

Comments

[Eric Hancock]

Hi Steven -- very interesting article.

What I find most interesting are the points you make in the last few paragraphs of the Slate piece: that the divergent shifts in the MacOS and Windows interfaces represent 180 degree reversals for both companies.

The Apple Finder was a revolution, and the simple organization of the MacOS made the computer seem more understandable (even if the finder was just a reflection of the users' mental model, rather than the software's physical model).

Windows started life by hiding more information than it displayed; the early file mangers were baffling and quite frightening for novices.

Perhaps the first sign of the philosophical change, Apple attempted an even-more abstract implementation of the "one tool to do everything" strategy with its implementation of OpenDoc (something that was a work of art, in terms of software design, but a confusing mess in terms of users). OpenDoc never took off, for many reasons.

It would be interesting to develop the comparison of the two (Windows and MacOS) philosophies; what is the motivation, what prior art are they referencing, and what is the final goal?

[Aaron Brethorst]

I'd like to make a point about the zooming effects you describe in this article. Interestingly, SGI's OS, Irix, has had some of this technology for years. Also, Jef Raskin has been flogging that particular horse for at least the last 15 years (with the Canon Cat, in particular).

[Alex Soojung-Kim Pang]

Steven-- An excellent piece. Two comments.

First, within Apple (and the user interface world more generally) there have long been arguments over what constitutes "simplicity." One might think that the definition of simple would be, well, simple; but it turns out to be like one of those optical illusions that makes your eye water if you get too close to it. It's like figuring out what's "intuitive" in interface design: it may rest partly on how our brains and bodies are wired, but it also depends a lot on culture and experience. (The desktop metaphor worked because it drew on all kinds of habits and metaphors that were already deeply familiar to people.)

This means that interfaces can evolve with users in some interesting ways. The mouse, for example, started out with between one and three buttons. Today, we've got scroll wheels, pressure-sensitive pads, and even buttons under the thumb. The last are interesting because even though it's technically very easy to put a button under the thumb, no mouse maker did so until recently: but cell phones, Blackberry devices, and game controllers have given a whole generation of users a level of thumb dexterity that people didn't have in the 1980s, when the mouse first made it into the mass market.

Second, Apple is inching into territory that lots of interface developers are wrestling with: namely, what kind of interface makes sense in a world in which the personal computer is no longer the only, or even dominant, tool for dealing with the electronic world-- or as you point out, the PC itself begins to take on an ever-wider variety of roles. We already have a proliferation of PDAs, cell phones, digital cameras, and other devices that have their own interface designs. Even if we don't reach a world of "ubiquitous computing," in which the PC all but disappears, we're facing a situation in which the circumstances in which we interact with devices and the content on them are multiplied radically. The desktop interface certainly doesn't make much sense in all these varied contexts. The big question is whether a single competitor, like a zooming user interface (ZUI), can work comfortably in most niches of this richer ecology, or whether we're in for a world of varied interfaces and mixed metaphors.

[Anil]

I'm not sure that these are different software architectures so much as different mental models that are being suggested to users.

The eventual form of Windows Longhorn will probably be a logical extension of the "My Pictures"/"My Music" way of working on things. Go into Windows Explorer (your file database) and the side pane will present you with tasks you can perform on that type of media. In OS X, go into an iApp (which has your media file database for that type) and the tool bars will present you with tasks you can perform on that type of media.

The only initial difference, really, is whether you're clicking into a folder or launching an iApp, and those distinctions are subtle right now, and will become almost purely academic in the future. Could I describe the picture-management functions of today's Explorer as an app? Sure. Could you describe the OS X Dock as a menu with iApps as menu items forming one big integrated environment? people already have.

Seems to me that the macro-trend is task-based applications. I think that the biggest problem right now is that we're still using half a dozen different applications just for our text-based search, retrieval, and communications tasks. We could easily fit all of those functions into one context-aware program. iWrite, anyone?

[Steven Johnson]

Folks, really interesting comments here. It's late and it has been a long day, so a few quick thoughts...

First, it's interesting to have OpenDoc mentioned, both implicitly and explicitly. (In Eric's post, and Anil's reference to task-based interfaces.) Haven't thought about OpenDoc in a long time...

Secondly, I was really trying to focus as much as possible on the question of file management in the piece, because once you get into document manipulation it's a more complicated question -- that's why I focused mostly on the iPhoto zoom metaphor (and in an earlier draft I talked about the ways that iTunes handled playlists.) Actually editing files is a separate matter. The weird thing about the iApps is that they're half file organizers and half document editors, and so part of them is very close to what the Finder does, and part is closer to to the role of traditional applications.

Finally, I only had time to skim Anil's microcontent essay, but it looked brilliant, connected to lots of things I've been thinking about, and I will try to revisit it tomorrow, after I file yet another piece. (!)

thanks for the great comments,

sbj

[Robert Daeley]

Hi, Steven. Very nice article that I'd like to think about more before commenting on -- however I wanted to let you know we picked up the story over at MacSlash and posted a link to it. We'll see what sort of discussion gets generated. :)

[Anil]

fwiw, mcwetboy, Windows Explorer (in XP, at least) shows ID3 tags from MP3s as columns of information, just like it shows dates and names.

[Jonathan Crowe (mcwetboy)]

Organizing photos and music through the file-browser interface will always be less than ideal so long as the user cannot access the metadata embedded into or associated with the files (such as EXIF data or MP3 tags). Thumbnail data in My Pictures is an improvement, but being able, in iPhoto, to associate keywords with each photo or assign photos to albums makes it that much easier for me to get at them.

[K. M. Peterson]

Hi Steve,

Very interesting article. I got caught on one thing, though...

Although iPhoto is cool, and the ability to dynamically change the size of thumbnails is great fun, it does suffer from being "closed" in that if you're already a Photoshop user it is not all that useful.

But, in the Finder now in OS X, one can have 128x128 pixel "thumbnails" which in fact _do_ make it pretty easy (ok, ok, "much easier") to manage image files as files. And one can also set the view to include other information as well - for MP3s the Finder will include the running time of the file, for some image files the dimensions of the files are listed; it seems likely that Apple will take this level of integration further.

The first (?) iApp, iTunes bothered me because I was used to organizing my music in the Finder (the previous MP3 players like Sound Jam did this). The jury is still out, as you indicated, but there's no good reason not to allow both, although Apple's file systems are now storing less interesting metadata than they used do.

Anyway, again, thanks for an interesting article.

[Jef Raskin]

I've watched the Mac get more and more complicated ever since I first started the project. It's quite a mess, and 10.2 doesn't help. Now it sounds as if Mr. Gates has been reading my articles and my recent book ("The Humane Interface", Addison-Wesley, 2000).

I've known how to vastly simplify interfaces for years, and am now doing it (see my web site). But most people can't see beyond a GUI-style interface.

The writer who commented on the difficulty of saying what is "intuitive" in interface design is wrong. It is very easy to define the term when it comes to interfaces. References to what I've written on the meaning of intuition in interfaces are on my site (jefraskin.com) and on pp 149 ff in "The Humane Interface".

[kenny]

the register has an article up on "spring!"

---
Expect to see Clay Shirky dust down his "P2P/blog/blank as revolution" essay once again. Expect to see lots of self-serving posts mentioning "Cluetrain", "weblogs" and "revolution". But the software is strong enough to survive being lauded by the bullshit circuit.

It's got style and imagination, makes great use of Cocoa, and the worst you can say about it is that it's fun. :D