The maintenance is the implementation OR Why people misunderstand IT systems

This short paper has been written more as a position paper intended to stimulate discussion than as a research or experience description paper. 1. The traditional scenario of mechanization Since the 18th century humankind has reassigned to machines jobs previously performed by human beings and animals, and this process is well understood. Assume you have to harvest a huge field or to assemble thousands of pieces or to execute billion of calculations. In the pre-industrial era you needed a lot of people and possibly the help of several animals. In the industrial era you buy a machine and use just a few specialized persons. The machine does the physical work, while the people drive it and supply the required cognitive processing. If there is a big stone in the field, a broken piece, a meaningless number, the person sees it, recognizes it, and avoids it. This applies to almost all industry mechanization. Machines are a substitute for physical effort by humans or animals. They need no capability to adapt to a changing environment, since persons driving them have this capability. They need no intelligence (in the human sense), since persons guiding them can supply theirs. Once you have mechanized and automated a task, more or less 80% of the job is done. It just remains a 20% maintenance job. This is the current view. 2. Computer-based mechanization and flexible/adaptive mechanization Then computers arrived, and humankind started to automatically solve a different kind of tasks by means of this new kind of machines. A computer is a machine automatically executing a "problem solving" task working on a suitable abstract representation of problem data and their relations. However, this time the problems to be solved, i.e. the tasks to be automated, are the ones where mainly human intelligence is required, and not so much strength or manual ability. But, on the other side, a computer just " shuffles symbols around " : each step in the execution of its problem solving 76 ECSS 2010