Our Broken User Interface

Imagine this: you’re in the process of putting together a powerpoint presentation on your company’s procurement practices when you remember an article you saved off from a web page a couple of days ago. It had a very cool quote on ferret mating rituals that perfectly illustrates your point about widget acquisition protocols, but you can’t quite remember what it was. So you minimize powerpoint to get to your desktop, only to find your email client lurking behind it, completely concealing the desktop. So you minimize that, but find your web browser staring back at you; and behind the web browser, there’s a promising game of solitaire you didn’t quite get a chance to finish, and behind that is an angry letter you were writing to your feckless jackass brother in Cleveland who still owes you $300.

And then, finally, you’re at your desktop. But, alas, the document isn’t there. You bring up your file browser, poke around half-heartedly for a while, give up, then decide you need to go back to the web to find the original article. You squint along the bottom of your screen until you find Internet Explorer, bring it up, and try desperately to remember the URL of the ferret site. After a couple of desultory, unsuccessful google searches, you give up and finish the solitaire game.

Does this sound familiar at all? It probably should. Although there aren’t any hard data to support this, I think it’s probably safe to say that computer clutter is reaching epidemic proportions on most of our desktops. There’s a lot of information out there, and it’s coming at us from all sides. The bad news is that it’s only going to get worse. The worse news is that our computers, in their current state, are not only ill-suited to deal with this infocrunch, they’re unlikely to get better at it any time soon, at least not the way things are going now. This is because the course that interface design is currently on is a dead end, and not the kind that stops at a blank wall: it’s the kind that leads over the edge of a cliff, into a bottomless chasm that we’re all pretty much doomed to fall into, unless things change radically, and soon.

Take Apple’s Mac OS X operating system, for example. By all accounts, it’s a stellar feat of engineering, much slavered-upon by technocenti and end-user alike. Case in point: the latest version, called Panther, introduces a new feature called Exposé, a wickedly cool interface trick wherein, with the press of a button, all of the windows on your cluttered, crowded display shrink and draw apart from each other until they no longer overlap, allowing you to see everything you’ve got open in a single glance. Click on the window you’re looking for, and everything goes back to the way it was, with your chosen window in the foreground. A different button does the same thing, but only with windows in a single application. A third just clears everything off the desktop, so that you can see any icons you’ve got lurking back there. The effect is absolutely stunning, and it works great.

But, beyond the coolness factor, I think this feature has a pretty fatal flaw — not in its implementation, which is first-rate, but rather in its entire concept. What Apple has done is find a good way to treat the symptoms of a serious problem in the current state of interface design, rather than offer a cure. And the irony here is that the problem is one that the Macintosh, in many ways, helped to create.

Rewind to 1983. The personal computer market, such as it was, was dominated by machines that could only speak to the humans they were made to serve through the crudest of interfaces — written commands (in a tower-of-babble profusion of languages) typed on a command line. Start up an IBM PC XT in the early eighties, and you were greeted with a little blinking white cursor on a black background, and not much else. It wasn’t an interface so much as a challenge. “Can you figure me out? No? Well then get lost, moron. You’re wasting my time.”

And then came the Macintosh, an unassuming little beige box that introduced the world to the desktop/windows/mouse paradigm that has become the overwhelming standard everywhere. It allowed people to understand their machine in terms that were already familiar from the physical world: folders, pointers, documents. It hid the incredible, intimidating complexity of an entire operating system behind a set of familiar metaphors. It was a marvel of simplicity and ingenuity.

But the solution, cool as it was, contained the seeds of its own demise — because it failed to account for a reality that had not (in all fairness) really manifested itself yet, and wouldn’t for many years. The picture on your display may look like the top of your desk, and may even act like it; but the reality is that it lets you do many more things than a physical desktop ever could, and that, underneath the pleasant lie it represents, is a massive, inchoate, undifferentiated, sucking datahole capable of storing as much information as you care to throw at it.

Our computers weren’t always such insatiable infowhores. The first IBM PC shipped with a 10 megabyte hard drive, which was a mountain of storage at the time, but doesn’t even rise to the level of molehill in this age of cheap 200 gigabyte behemoths. Our capacity has increased geometrically over time, and into that unslakeable sinkhole flows rivers of raw data, not only from the applications we’ve begun to base our lives around, but also from our overloaded, incredibly verbose pipelines to the networked world: email, web pages, news readers, RSS feeds, and more. We store photos on our computers now, and movies, and music; not to mention our email calendars, our letters, term papers, bake sale fliers, tax returns, and more. We’re besieged by data, and our computers are more than happy to swallow all of it up for us. It can handle it. We can’t.

These days, your average desktop is an unholy mess, a post-apocalyptic proliferation of documents and applications scattered like storm leavings across a blank, undifferentiated landscape. But, under the covers, things are even scarier: towering directory structures ramifying like cancerous fractals, filled not only with the documents that you filed away and forgot, or filed away and remembered but don’t know where you put them, but also with strange, oddly-named entities like “salkj34343.dll” or “/dev/disk01s1″ that you must not touch under any circumstances lest the whole rickety edifice come crashing down around you.

And, to make matters worse, the desktop metaphor gives us the ability to do many things with that data at the same time. Each task gets its own window, and each window its own share of our increasingly fractured mindspace. You can check email while you work on your report while you listen to the latest Fountains of Wayne album while you balance your checkbook while you compile your new kernel. Which, don’t get me wrong, is all very cool. But it’s too much for the metaphor to handle: the insane proliferation of windows that all of this furious multitasking scatters across your display strains the whole notion of a desktop — not to mention our ability to keep it all straight in our heads — to the breaking point.

The OS people haven’t been ignoring these problems, but they’ve been paying the wrong kind of attention to them. Exposé doesn’t make the clutter go away, it just spreads the haystack apart to make it easier for you to find the needle; Windows XP’s ability to collapse multiple windows from the same application into a single item on the startbar is nice, but it requires you to keep a mental model of what’s available on your desktop, and then hunt through a profusion of hidden apps for the document you seek; Unix’s virtual desktops can be an effective tool, but ultimately, they fail for the same reason: they shift the burden of organization onto the human, when really, it should be the responsibility of the computer itself.

And none of the major OS vendors have even tried to address the problem of data overload, beyond some seriously hyped up search facilities. The next version of Windows will purportedly augment the existing filesystem strategy with a database paradigm that eschews hierarchy in favor of association — which is to say, instead of storing your macaroni and cheese recipe in the “/user/stuff/recipes/carbs/pasta” directory, you’d simply associate it with a couple of keywords (”pasta” and “recipe”, for example) and then call it up by asking for a pasta recipe.

That’s a step in the right direction, but it’s not quite enough. Simply put, the current desktop / directory hierarchy paradigm doesn’t scale. We need to chuck the whole thing, and come up with something brand new. I don’t know what that something is. But I have some ideas.

First: I don’t think we can address the issues of application and data overload separately. They’re really two manifestations of the same problem, which is the current metaphors’ tendency to give us the tools to organize our computer experience, but leave the bulk of the actual organization to us. Whatever solution we come up with has to deal with both, simultaneously.

Second: Whatever we come up with has to be smart. Not human smart, and not even artificial intelligence smart, but contextually and organizationally brilliant. It has to divine our intentions based on a minimal amount of input, and then do the right thing.

Third: It has to work with current technology. No, I don’t mean it has to be able to run well on a 1.5 GHz desktop with 512MB of memory (although it does), I mean we have be able to use it with keyboards, mice, and monitors; although it would be nice to come up with some three-dimensional virtual reality avatar-based control mechanism for all this stuff, the truth is that’s way down the line, and, as I said earlier, we need results now; or as close to now as possible.

So what does this mean? What I envision is a single, monolithic application, an alter-ego that handles everything for us, and does it with a minimum of fuss. You sit down at your computer, turn it on, and just start typing; an editor comes up, records what you’ve written, and when you press “Done”, tries to figure out what it’s looking at. Is this an email? Then scan the text for some indication of who it’s for. Ah, the first line is “Dear Aunt Emmy”. Go to the address book and pull up Aunt Emmy’s email address, and sent it to her. No email address? Then print the document out, along with an envelope addressed to her snail mail address, with the appropriate postage barcoded in its upper right hand corner.

Or maybe you sit down and type (or say) “Washington Post”. Your pegboard (let’s call it a pegboard; I like the sound of that) shows you the Post’s web page. You click around until you come across a grilled ferret recipe that looks yummy. You select it, and drag it to a corner of the pegboard. A prompt comes up, you type “ferret” and “recipe”, and hit enter, and it’s stored on your computer, keyed under the terms you entered, but also the date on which your saved it, the web page it came from, and maybe some other stuff that the computer figures out on its own while scanning the text of the recipe. Next time you get a hunger for grilled ferret, you sit down at your computer, type in “grilled ferret”, and there it is.

Something like that. The idea here is that you should spend no time at all thinking about how to do what you want to get done. Instead of starting a word processor, bringing up a file dialog to find your recipe, loading it and printing it out, you should let the computer handle as much of the work for you as possible. It won’t get things right all the time, but it will never get things terribly wrong, and, over time, it’ll learn your tendencies, and become that much smarter.

That’s the holy grail, I think: an interface that’s so intuitive that it’s no longer even really recognizable as an interface. It’s just an extension of you. This will happen eventually. It almost has to. The only question is when, and how much longer we’ll have to endure the broken system we’re stuck with today.