I just can't get it up for user experience design. I'm sorry, I've tried. It just doesn't turn me on. I mean, I get it—I get why people are into it—but it just doesn't do it for me.

At least, as a first-order thing.

User experience, as I can't help but read it, is about feelings. Feelings are great and all, but what they aren't is objective. I mean, you can't really be sure that two people will perceive something the same way, and less so that they will perceive something the same way you do. As such, how can you really confidently design a user experience?

Perhaps it has something to do with the fact that I have a few solid years of software engineering sandwiched between two design careers, but I need something a little more solid than that. It helped to have somebody much more famous than I broach the heresy. As such, I'd like to publicly thank Don Norman for writing the piece that reconciled my experience by setting it apart from the confusing messaging of others. That said, my perspective warrants its own treatment.

Flickr user Thaya Nanth

Consider the humble toy called the Rubik's Cube. I choose it because it is an extreme example of what I intend to demonstrate. A Rubik's Cube is an implementation of a symmetry group in a colourful package with a pleasant size, weight and feel.

What makes the cube special is that it is an interactive system, much more tightly constrained—by its fidelity to the underlying mathematics—than the interactive systems with which we are typically concerned, but interactive nonetheless. From a representational perspective, the cube is also digital, since it is a group after all: a set of meaningful symbols connected by an operation—twist—which enables the set to permute into a number of representational states. 43,252,003,274,489,856,000 states to be exact. But the person using a cube doesn't have to remember anywhere close to 43-quintillion-and-change things. She has to remember exactly that the object of the game is to align all the colours, and twist. No more.

I submit therefore that the true power of any interactive system is its ability to create a consistent and lasting mental model in the user, and likewise one that can be appropriated as an analogy for solving other problems in life. If you play with a Rubik's Cube enough, you will get a glimpse of the properties of symmetry groups, maybe even an inclination to look deeper into them. If you go and research group theory, you gain a solid theoretical understanding of the mechanics of a Rubik's Cube. The bulky-yet-dynamic process and the compact-yet-static model reinforce one another. The idea that so much complexity can be collapsed into such a simple object is one that makes you smarter for learning it.

Conversely, if I enter the problem of creating a cube focusing on how users feel about it rather than how it works, I might pore over different materials, sizes, tensions and colour palettes. I might fashion a raft of prototypes and double-blind test them against an army of participants. I might generate a mountain of very compelling data about the look, feel and interaction with the cube.

But what if the damn thing doesn't work?

What if there is no kernel of truth in the device? What if it encodes no knowledge that if-I-do-this-then-I-get-that? What if it hides its state and demands that its users try to keep track of it? What if it exports no valuable rule or lesson that can be applied to other aspects of an individual's life? What kind of an experience is that?

This, my friends, is why I don't design user experiences. I design structures and methods. Models and processes. From these the user experience is derived, at which point the details can be refined. But first the system must be consistent, with the rigour of the cube in mind.