Here is a small diagram that summarizes the (analytical) approach allowing the organization of the primary data of a computer system.

​

But despite the obvious satisfaction you feel ... you know that, unfortunately, the all powerful machine does not exist and that, in spite of the phenomenal technological advances that the man realizes continuously, the processing of the data also evolves at least as quickly, if not more. Consequently, we must (although in principle all processing must be able to be executed from the primary data) store secondary data (calculated by our programs) so as not to be obliged to redo all calculations to from the primary data.

As a result, you will no doubt have to decide to keep secondary logical files in the computer system. We will not treat the subject here because it is first and foremost a matter of computer scientists, everyone will have understood the why of the thing.

​

I only have one extremely important point to remind about secondary data, remember that these data correspond to an image in time, you will have to keep all the necessary data that will allow to characterize these data at the moment when they were Created. I give you a small example, let's take the turnover realized by a customer at a time T, the latter belonging to a category of customer X at this time, the company evolving, this customer changes category of customer at a moment T + 1, if one wishes to do statistics by client category, it is perhaps the category to which the client referred that must be taken into account, and not the news to which it belongs now.

​

We must at all times store the data (images of the real world) in time !

​

Now, I have to tell you that we still have a step to be taken, I let you to think two minutes ........ 1 ......... 2

And yes you have found, we have to study the data redundancies both at the level of the records of the FLP's and at the level of the data fields, because as you have seen in the LCS study, all the sets ( FLP) defined are disjoint, therefore we will study the meeting of all files to discover the existing redundancies (intersections) and make some decisions. This last study will lead to define the Logical Operational Files.

​