Innovation dynamics across theory, technology and tool

Pieter Wisse

Some time ago, I gained better insight into why a particular subject my independent design company Information Dynamics is working on is so extremely difficult. Originally, I believed my goal was just a new tool, or application for processing information. Regretfully, however, no technology was available to develop the tool as I saw fit. The next discovery was that the required technology — which of course is also a tool of some sorts — couldn't be properly developed yet for lack of an appropriate theory. It meant, as I came to understand along the way, that my company ought to be busy innovating three t's in conjunction. Awareness grew in the order that principles are more important: tool, technology and theory. At the company, however, the work only started to progress when the problems were tackled in the reverse order. That is, with first principles first, and so on.

My own contribution has mainly been to develop theory, i.e., a new approach for conceptual information modeling. It will undoubtedly undergo further changes. But it is now in a stable version, about which I published the book Metapattern: context and time in information models (Addison-Wesley, 2001). Forthcoming is Semiosis & Sign Exchange: conceptual grounds of business information modeling, a text dealing outright with ontology. Martijn Houtman, chief developer at Information Dynamics, has turned it into an innovative technology. It is called KnitbITs® and is now in prototype stage. Finally, then, new tools can be developed. Everything included, it is a long and difficult process. The reward is that the ultimate tools are really better.

Many people are sceptical about innovation. And the more fundamental the change, the more sceptical they get. Is my claim with respect to innovative theory credible? Indeed, what I propose with the so-called metapattern is a new way of organizing knowledge. Actually, theory and a way of organizing knowledge are synonymous. The metapattern amounts to denying absolute existence for objects. The first principle of the metapattern's epistemology is context, rather than object. The number of contexts is variable, as is what counts as context. Object, and with it relationship, are now only secondary knowledge principles. It follows that an object always belongs inside a context. The next logical step is that object behavior is essentially context-dependent.

Now, taken one by one, many of these particular ideas are admittedly not novel at all. As a general concept, for example the contextual nature of behavior is already well-established in — the context of — social psychology. However, its formalization suited for application for digital information systems, it seems to me, is an innovative step.

The traditional technology supporting object-oriented development is not grounded on the concept of context. It also cannot be easily adapted to it on account of the fundamental position required for contexts. So, it's quite logical that only new technology can adequately support development of context-oriented information systems and services. It is interesting to note that, in fact, traditional object orientation neatly fits inside the new context orientation. For what has been surpassed by shifting the first principle to context can now be considered the simple boundary case where an object implicitly is its own, single context.

With the technology developed to a stage where it actually works, it is possible to produce new tools. The elegant mechanism for supporting multiple behavior replaces the awkward practice of subtyping. The solution space for tools has grown significantly with the combination of new theory and technology. Its potential has increased additionally because support of change has been a prevailing requirement right from the point of theory. At the level of a specific tool this means that any information, regardless of level of detail or aggregation, may be kept available as requirements demand.

In hindsight, it is obvious that context and time should be core concepts for information modeling. Isn't it also the influence of — applications of — digital information and communication technology in general which cause people to experience their world as increasingly pluriform and changeable? A uniform, static theory of knowledge nowadays fails in utility. Context and time fuel an innovation of knowing with the result that a person sees the world differently. It has been this observation which calls for fitting control measures. There originated the requirement for essentially novel tools. And bridging theory and tools, a new technology is developed, too. Gladly I would have saved my company the extended efforts, but the dynamics of theory, technology, and tool are simply unavoidable.


© February, 1999; web edition 2001.

Pieter Wisse is president of Information Dynamics.