“Clean-up” implies removing redundant stuff and complexity. Large enterprises with a history have at least some 5-10 times more applications and data bases than they actually need. Of course, they are all in use, and you cannot remove any without hurting the business operation. But there is a lot of redundancy.

State of the art

You must understand why and how this redundancy and complexity were allowed to build up. Remember, this has been “state of the art” for many years. Actually, from the early start of computing and system development. You cannot expect another result if you continue with more of the same. You need to break out and take another route.


Enterprises are different when it comes to how they manage change and business development. Some let change be managed locally, while others coordinate centrally. Investment philosophy differs accordingly. This affects the process for acquiring business applications. We have for a long time said that building systems based on the local need from organisation or process will lead to redundancy. This is correct under one specific condition: applications own their data. And that condition has been the state of the art so far.

Loose coupling

If we break data loose from applications (programs) and manage data as a common resource for all (processes and organisational units), then you will have much more freedom to build small applications that serves local needs. It also gives you more freedom to remove applications. When the applications do not own its data, they will not create the same redundancy and complexity.

Trained incompetence

In the book Software Wasteland, Dave McComb claims that the biggest obstacle for change to happen is in our mental models. We are so used to build systems the way we do, and have done, that it is hard to change. You may say that the profession has “a trained incompetence”. If that is the case, then you are more or less blind for other ways of doing things.

Do not expect cannibalism

Another obstacle for change is the IT industry. While IT is investment and cost for enterprises, it is at the same time income and profit for the IT industry. Let us say that enterprises spend twice of what is needed (due to redundancy and complexity) – then we can assume that the IT industry actually should be reduced to half of what it is today. There is no reason why the industry itself should look for solutions like that. Do not expect that to happen.

No silver bullet is needed

But there is a good chance for this development to happen. First, there is no silver bullet of technology you can buy. That is good. Because it is not such a bullet you need. Implementation of a new and better architecture is only dependent on existing standards and technologies. But it must be architected. And architects and developers must work together.