Wednesday, October 31, 2012

Architecture and the Imagination

"Thinking about the future is a form of unreality." Leif Frenzel, Lost time, sedimentation, and the future as a form of unreality (March 2011)

An architect looks at a valley and imagines a viaduct. She then describes this imaginary viaduct in great detail. As a result of her imagination, and the efforts of many engineers and other workers, when we visit the valley ten years later we too can see the viaduct, now fully realized in graffiti-daubed concrete.

Similarly, much of the work of enterprise and solution architects refers to things that don't exist yet. Most obviously, this applies to systems that haven't been built yet. But it can also apply to business concepts that haven't been "realized" yet.

For example, before Apple launched the iPhone, it must already have had a reasonably well-elaborated concept of IPHONE-USER, and it would had ensured that this concept was adequately supported by a combination of existing and new systems. (Of course, the concept of IPHONE-USER may well be a specialization of the concept of CUSTOMER, but there is a lot of new conceptual matter to accommodate.)

The concept of IPHONE-USER is essentially an exercise in imagination. However this imagination can be grounded by various practical experiments - for example, test engineers creating artificial or proxy instances of IPHONE-USER to make sure everything works properly.

Some 1980s methodologies, including Information Engineering, preached that a conceptual or business information model was in some sense timeless, and that Information Strategy could essentially be reduced to Information Systems Strategy. (This agenda was of course promoted by companies who wanted to sell information systems.)

I now think it is perfectly reasonable for the conceptual model of the enterprise to evolve over time, as the business starts to conceive of previously unconceivable things. Here are some well-worn examples.

  • The introduction of loyalty cards into retail, allowing retailers to recognize their customers as "the same again", and therefore replacing the concept of a customer-per-visit with the concept of a customer with continuity over time.  
  • The ability to track activity at ever-finer levels of granularity. For example, monitoring every click on your website, or watching your customers navigating the store. (Did she pick up the lemons from the display next to the fish or from the display next to the gin?)
  • Exposing a wealth of associations between products and customers. For example, Amazon's development of the "people-who-bought-this-also-bought-that" pattern.

So as I see it, Information Strategy includes imagining new things for the business to pay attention to. This is a lot broader and more interesting than Information Systems Strategy, and is just one of the areas where the architect needs to use some imagination.

See also

Updated 30 March 2013


Rob Vens said...

Imagination is a very powerful human faculty, in fact I would argue the distinguishing faculty of being human. So your article struck a chord with me, Richard, thank you.
Architecture, when dealing with the "real" world, uses "models" which are replica's, so to speak, of this "real" world. These internal models are what imagination is all about! There is an important and usually overlooked relation between language and writing (in which we build and enhance these imaginary worlds) and architectural models. I have quite a few problems with how models in business architecture are handled, because they lack what is called loose coupling and often try too much to be the same as the things they model.
Do you have ideas about characteristics of "good" (and thus, scalable) models?

Richard Veryard said...

Thanks Rob. What you mean by "real world" - if the architect imagines a bridge that hasn't yet been built, which "real world" contains this bridge? Because of this difficulty, I don't find talking about "real worlds" to be very useful in understanding how architects work. See my post on Architecture and Reality.

On your second point, see my posts on The Value of Models, The Value of Models 2.