Modelling in Information Systems Engineering When It Works and When It Doesn’t

Abstract

This talk will reveal two secrets.

The first secret is that it is not due to lack of formal methods or inappropriate formalisms that roughly 50% of all implemented system functionality is thrown away even before the roll out of the systems concerned – and 70% of the rest is unused after some two years in production.

The second secret is that today, if you make a small effort, you will learn how to beat 80% of the consultants on the market when it comes to modelling processes – just by a simple twist of perspective.

The aim of this talk is to make you less easy to fool and to give some hints concerning potentially profitable research directions as well as some inspiration for further thinking. A basic assumption is that businesses as well as academia exist to create value. After some 25 years of experience in commercial modelling, this leads to some questions concerning general aims and basic paradigms in research and practice.

In my opinion, the academician has an extremely important role to play, but something seems to have gone wrong. The basic reason might be that, paradoxically, the concept of information as well as the consequences of its proper definition seems to be more or less completely disregarded in information science. Naturally, this leads to severe misconceptions concerning the relevant problems to attack. As a further consequence, academic structures often represent a real mismatch with respect to necessary interdisciplinary cooperation.

(Yes, yes, of course, there is a plethora of definitions of information.)

In my opinion, the practitioner in modelling has avoided responsibility and real influence by being too narrowly focused. Modelling as an art has often deteriorated to simple description. While simple design instruments are used analytical instruments, if they are at all known, are not. Moreover, in some cases, commercially developed methods are sometimes downright dangerous to put in the hands of the normal analyst. Finally, inviting disaster, the modeller at large has limited abilities and instruments for the necessary cooperation with management and business development representatives.

There are, however, some remedies available...