Abstract computing machines

Abstract

During last five years I have felt the need to possess powerful theoretical means of improving my designing capability by automating the production of operative implementations. This implies on one hand the identification and avoidance of the causes of errors, and on the other the rigorous definition of all the implementative phases in order to make unnecessary the use of ingenuity and creativity for the product implementation. Most computer-users consider the problems (i.e. errors, lack of documentation and a generallly low level of service ) in s/w and h/w as unavoidable, inherent “disasters”. Unlike them, I believe that these problems can be solved. In [I] we show that most designers and purchasers accept having to spend as much as 100% or 120% more than the time predicted and the cost of the definition, design and initial manifacturing, stages, merely in order to achieve an acceptable product. Delays, efforts, and energy spent on meeting the requirements and the product during the test and the error-avoidance stage, yield a new product. This product conforms to specifications, requirements and financial evaluations that are completely different from the original and, in general, expensive compared to the environment from which the design originated. The errors due to indeterminate specifications, to the imprecisions inherent in working practice, and in basic tools such as programming language, operating systems, or electronic components, al,d the impossibility of exhaustive testing for purposes of experimental validation and fault avoidance are the cause of growing costs. Indeed, there is an unpredictable interval between identification of the needs of a computing system and its actually becoming available.