Some complexity is necessary. If you're
serving an inherently complex business process, then failing to reflect this
complexity in a systems design will lead to poor functioning and error. But
much of the complexity we find in information systems isn't this kind of good
complexity. It's complexity grown out of systems development, maintenance and
integration. It's the kind of complexity that grows like topsy.
There's one kind of needless complexity that can be tackled relatively easily. This is the complexity created by failures to use relevant data standards. Generally speaking, the best way to reduce complexity of any kind is to restrict choice. This is what regulations do. For example, you don't get to be creative about which side of the road you drive on.
Standards work in a similar way. Standards reduce choice where choice is at best unnecessary, at worst downright dangerous. A data standard restricts the rights of participants to deviate from agreed norms of behavior and flows of communication.
Good standards are non-arbitrary mechanisms for restricting choice. The form of such a standard is derived from the natural characteristics of the business being served. (Which side of the road we drive on can be based on a coin toss, but the structure of an insurance policy record had better be based on insurance policies.)
Great standards go beyond good standards by exploiting choice restriction so as to enable and encourage growth and innovation. Such standards model the business they serve with a whole-industry coherence, rather than focusing only on islands of interest.
Decision theory has a concept known as bounded rationality. This is the idea that anyone's ability to make a decision is limited by the information they have, their ability to process it, and the time available. It's a major corrective to the basic idea in economics that people make optimal choices based on perfect information. I bring this up because it seems to me that an industry data standard is a map of the bounded rationality of the community that creates and exploits it. Data standards are significant artefacts which express the information capabilities of a group of participants.
So, when you deploy industry standards you anchor your organization's information activities in explicit statements of competence. By adhering to standards, you are using as much or as little complexity as is really needed to get the job done. Not too much, not too little, but just right.
Comments