In the mid-1960s, great progress was made by those who envisaged taking the problem from the top to the retail level. to a reasonable approach, consisting of going from the beginning to the end, begins a new approach going from the general to the particular. This approach is referred to as "Top down programming".

​

Professors Boehm and Jacopini in Turin, Djikstra in Eindhoven, Wirth in Switzerland, and many others contributed to this development.

​

However, it was not until the 1970s that the theories of structured programming influenced the behavior of computer scientists. The list of those who contributed to it would be too long. Let us mention only: H. Mills, Baker, Yourdon, Constantine, Jackson and K.T Orr.

​

In France, different programming methods were proposed, of which the most widespread and most seductive was the CORIG method.

​

All these efforts had beneficial effects: Programs were clearer, faster written, more reliable, and above all easier to modify.

​

Users of these methods seek to list in an orderly manner the functions that the program must perform. They study independently the conditioning of each function and the programming of the instructions that make it possible to perform it.

​

These programs, which are clearer than empirical programs, are more easily modified, it is a fact. However, there are some commonalities to these approaches, which are still obstacles to the evolution of programs. We can summarize them as follows:

   - We take more account of programming languages ​​and techniques used than logic.
   - Consideration of the organization of data at the output and at the input of the program is mostly         non-existent, one is concerned only with the physical files.
   - We neglect to define the input and output data as well as the program data as mathematical sets.

​

Consequently, these methods do not provide programmers with the means to approach the problem through a rigorous hierarchical approach, since they have neither a definition of sets nor precise subdivision criteria. It seems, however, that such means are necessary to make the requested changes later under the best conditions. The approach advocated by J. D WARNIER consists in organizing the data, and more particularly the data of a program into hierarchical mathematical sets.

​

It is obvious that the teaching of such concepts implies that computer science must be considered as a science and not as an art, and only on this condition the training can be disseminated and received by the greatest number.

​

You could tell me, all this is outdated ! Today we talk about AGL, SOFTWARE, object oriented programs, WEB, and I forget surely ... but should we not always write Program instruction lines as powerful as they are?

​

So if curiosity animates you, follow me ...

​