ANALYSIS
Interface Burdens and Mac Usability

It's generally clear that the tools you use influence the way you think. To a two-year-old with a hammer, everything looks like a nail, just as PowerPoint's indented bullet lists both structure and limit much business thinking about more complex issues.

We often think we know how to do something because we know how to use the tool that does it, but don't stop to think that other tools might produce better results or to question whether the tool we know might not be dictating the results we expect. In part, that's what those jokes about the Microsoft (Nasdaq: MSFT) car or airplane are about, not the products, but the habituation that leads us to accept them: a generalization to human behavior of the story about frogs not jumping from boiling water if you bring the heat up slowly enough.

It seems reasonable to believe that the extent to which the tools we use affect how we conceptualize the work we do should vary with the degree to which the tools themselves require our attention. In other words, the less attention the tool requires in use, the more clearly we can focus on the job we're trying to do with it. And, conversely, the more the tool itself intrudes on our awareness, the more it will influence how we use it.

Client, Interface and Application

In the Windows world, for example, the typical PC client imposes a control structure, or conceptual space, that differs from that of the Windows host OS. As a result, the user perception of the business process managed through this type of client is influenced first by the client, then by the Windows interface under which it runs, and finally by the realities of the application itself.

In total contrast, Apple (Nasdaq: AAPL) has always made the Mac OS interface transparent, thus allowing it to disappear from the user's consciousness by requiring that applications use the same interface standards as the OS. Notice that this isn't just a matter of familiarity: The Mac's commitment to standardization doesn't just have the effect of knowing what the main controls do in a rental car; it removes the need to focus on the interface from the process of learning to use the application. In other words, that Windows interface barrier is completely missing in the Apple world because Mac OS transparently hosts the applications without imposing itself on the user's attention.

One immediate corollary to the hypothesis that intrusive interfaces affect how we understand the work to be done is that the design assumptions in a tool that dominates a profession should eventually be reflected in nearly all the work done by the people using it. We could speculate, for example, that long-term use of Windows among the architects, engineers and planners working on a building project should eventually produce buildings that feel like physical implementations of the Microsoft Windows interface.

Toronto's New Air Canada Terminal

I got to thinking about all this the other day because I had the misfortune of passing through Toronto's new Air Canada terminal. To set the stage for this, bear in mind that the feds have been converting airports from cost sinks to revenue sources by handing them over to local control -- meaning that most are now run by independent airport authorities with taxation powers over passengers and airlines but no responsibility to answer to anyone.

As a result, almost all of them are building lavish memorials to themselves with Toronto currently in the fourth year of a ten-year rebuilding process, but already a year behind schedule and considerably over the total budget. Air Canada, meanwhile, has been operating under bankruptcy protection for some time, having tried every known strategy short of serving the customer to avoid having to recognize that combining the cost structure of a high-quality airline with the operational behavior of a third-rate discounter doesn't work.

Between the two groups, they've built Air Canada's new Terminal One into the perfect image of themselves, and their design tools: bloated, facile, dysfunctional and expensive.

In theory, the thing was to be a connected series of alternating embarkation and arrival halls with each pair laid out in a crescent to interface parking space for super jumbos on the inside to passenger ground transportation on the outside. In practice, the halls look huge because they contain something like five stories of empty airspace, but the floor area won't support even an A380 load without crowding, and the automated baggage handling isn't. Worse, there's some shock-absorption capability built into the steel frame but garbage cans aren't positioned in blast director cells; airflow can't be controlled between halls; the walkways from the parkade debouch on partially open islands overlooking pairs of halls; and, there are cameras everywhere but no staff anywhere.

Parallels to the Windows Interface

Problems with the parking operation reflect the deep thinking that went into overall operational design - and the parallels to the Windows interface are striking. For example, you can't pay on the way out; you only can pay in the lobby, but you don't find that out until you try to leave. When you do find that out, it turns out that there are three separate steps, no obvious links between them and no failsafes that I could see. What you do is this: You take the entry time card issued when you get there to a machine in the lobby, which takes CDN$5.50 an hour and returns a receipt card.

Put that card in the machine blocking your exit from short-term parking and it too returns your receipt, which the third machine, this time at the actual exit, retains whether you need a record or not. Of course there's nobody there to tell you any of this up front, nor could I get any information on what happens if a machine fails or more than a few minutes elapses between accesses by the three card readers. Because there's no one to pay at the exit, I assume you can either just get the first card punched on the way in, thereby stealing the stall rental, or just block exiting traffic until they either reboot the parkade or send someone to help.

In total contrast, Delta's hub at Salt Lake City is a model of both clarity and functionality despite being both larger and busier. Get off the airplane from Sacramento and it's obvious where the flight to Edmonton should board because the layout follows the airline's own hub and spoke flight design and there are directional signs -- as well as staff -- everywhere. As a result, Salt Lake International, although big, busy and noisy, is a lot like the Mac interface in that it imposes few burdens of its own on people using it.

Cool, Unreliable Features

Unfortunately, Toronto's new Air Canada terminal does make a lot of demands of its own on its users. In fact, the whole thing feels like something out of a bad Blade Runner remake: a physical realization of what working with Windows means -- right down to the existence of some really cool but unreliable features. For example, each parking stall has what looks like a motion detector with little red (occupied) and green (empty) lights, while aisles have electronic signs showing how many spots are open.

That'd be great, if it worked, but it doesn't. In two out of three aisles I checked, the counts shown were wrong. The bottom line is that the place is just like Windows XP -- navigation through this interface between airplanes and ground transportation isn't intuitive because much of it doesn't make any sense, but frequent users can memorize routes and procedures well enough that the whole thing would seem really slick if everything actually worked and there was no terrorist threat to be concerned about.

Now, obviously, Microsoft didn't design this terminal, nor can we reasonably ascribe all the design problems here to the use of Windows by those responsible; and yet calling it Microsoft Airport 0.9 feels right. The reason, I think, is that many of the factors driving this interface between planes and people so far over budget, and making it so obstrusive and dysfunctional to the uninitiated, are the same factors that people who aren't committed to Windows see in the Microsoft interface: design bloat, superficiality and the total absence of functional discipline.