Thought I’d engage in a little dancing about architecture, a pursuit that has been compared by some to writing about music. But to get to architecture, and here I’m really referring to networked computational communications systems on whatever technical stack, I’ll make an initial move toward the user. And in particular, some thoughts about the practice of user-centered design.
Just as with the concept of ‘usability,’ the words ‘user-centered design’ now simply mean ‘good.’ As in, ‘For this project, I’m looking for a usable web site created through a user-centered design process.’ The user is the customer and the customer is always right. You might be given to think that the user is a person, a human being—someone like you and me. But you’d be wrong. Users are constructs of the system of use, they have no existence outside of the system.
The user experience (UX) world is beginning to realize that while it may seem like they’re crafting experience for humans, networked business systems don’t actually care about humans. Frankly, they don’t know what a human is. On the other hand, they have well-defined formulas to compute return on investment. If there’s ever a question between achieving a business goal and a human goal, UX designers are learning the issue will always be decided in favor the the business. In a sense, there’s not even a decision to be made.
Why then, do we hear so much about user-centered design in the world of corporate web site construction? Putting customers first seems like the right thing to do. And, of course, they do it because they care. The question is, what do they care about?
When a system refers to ‘user-centered’ design, it’s really asking for an optimization of what the system defines as a user. On its surface it sounds like a transfer of authority from the system to the user, but ‘user-centered’ simply means that friction in the transaction interface should be reduced to the point that the user’s inputs are within the range of responses the system can accept as parsable. The system isn’t actually able to respond to the what the user, as a human, wants.
In some sense, the goal of user experience (UX) design is to limit the incidents of users speaking nonsense to the system. In the old days, users could simply be rounded up and sent to re-education camps where they would study thick manuals that would instruct them on how to stop speaking nonsense to computer systems. These days the system must provide immediate feedback and a short learning curve to move the user from spouting nonsense to crafting inputs that are parsable by the system. These small corrections to the user’s behavior makes the user a more efficient gadget, as Jaron Lanier might say.
If enough users speak the same nonsense to the system, a pattern is recognized and the system is moved to assign this new nonsense to a well-defined function of the system. But, in general, it’s the system that will train the users to utter the appropriate nonsense. As David Gelernter notes in an interview with Der Spiegel about the Watson system, all human input into computerized systems is nonsense. These patterns of nonsense are assigned meanings within the system of relations of the machine. The system doesn’t know who you are, doesn’t know what words are and doesn’t know what you mean by them.
SPIEGEL: But let’s assume that we start feeding Watson with poetry instead of encyclopedias. In a few years time it might even be able to talk about emotions. Wouldn’t that be a step on the way to at least showing human-like behavior?
Gelernter: Yes. However, the gulf between human-like behavior and human behavior is gigantic. Feeding poetry into Watson as opposed to encyclopedias is not going to do any good. Feed him Keats, and he will read “My heart aches, and a drowsing numbness pains my senses.” What the hell is that supposed to mean? When a poet writes “my heart aches” it’s an image, but it originates in an actual physical feeling. You feel something in the center of your chest. Or take “a drowsing numbness pains my senses”: Watson can’t know what drowsy means because he’s never fallen asleep. He doesn’t know what pain is. He has no purchase on poetry at all. Still, he could win at Jeopardy if the category were English Romantic poets. He would probably even do much better than most human contestants at not only saying Keats wrote this but explaining the references. There’s a lot of data involved in any kind of scholarship or assertion, which a machine can do very well. But it’s a fake.
If computer systems don’t understand humans, how do humans have an influence on systems? The humans who program the systems have a big influence prior to the point where the system is embedded in a business model. The other point of influence is via the system of laws in which the computer system is embedded. For instance, there are laws about security breaches, the use of social security numbers and zip codes.
And so we come to the dancing about systems architecture. The big corporate backend systems that have been exposed to the Network weren’t conceived as occupying a connected space. It was the rise of Java, XML and web services that created the connectors to put the big iron on the Network. The fact of connection changes the system at the margins, but not in its core.
The big web systems like Google, Twitter and Facebook have built big data repositories that allow them to rent out the correlation data. Google and Twitter in particular have simplified user interaction to the point that there’s basically one action—type and submit. But the center of power remains with the data correlation store. That’s what makes the train go. Doctors are beginning to look at the big data available about their patients and wondering whether they’re treating the data or the patient. Of course, the data will survive regardless of the outcome with the patient.
Changing the balance of power may be a long time coming, and as some have noted, it will need to be baked into the architecture from the start. There are a few new approaches that begin to move in a new direction. Jeff Jonas’s G2 rig combines elements of John Poindexter’s original design for Total Information Awareness, the Privacy by Design principles and Jonas’s own previous systems that do sensemaking on big data in real time. Particularly notable is the system’s ability to course correct based on every new piece of data and to hide the human-readable facet of data through anonymizing and encryption. Other architectures move toward establishing the user as a peer (P2P), in particular Searls’s VRM, Windley’s KRL, Bit Torrent and the recently departed Selector.
A true user-centered design practice will probably have to start on the user’s side of the glass, establish the user as a peer, and not be architectural in the way we’re used to. It’s only in this environment that a possible economics will take root. It’s also here that a developer and designer would finally have standing to do user-centered design. We might hope that such a move would happen because it was right, true and good, but this kind of dance may require a platform that isn’t a platform.