Information Requirement elucidation has been intensively studied in MIS field. But yet, it had not gone far away. Why? Because people looked at it in an inappropriate way. Information Requirement had been considered from a positivism point a view, as a “positive data”, that is to say something that has a “real” “existence” in an user mind. We believe that the constructivism perspective explains better the nature of information requirement. Information Requirement is a construction, being constantly constructed and re-constructed by several actors being influenced by all the systems (organizations, society, etc…) in which that lived.

http://www.bim.edu

INTRODUCTION

Although the Information Requirement (IR) analysis is an old process, it is considered an important problem (Boar 1986, Telem 1988, Zmud et al. 1993) because it is one of the most difficult steps (Munro 1978), indeed the most difficult step in the life cycle of system development (Zmud et al. 1993). Why is that step acknowledged as the most difficult one? Because, according to many authors (McKeen et al. 1979, Davis et al. 1986, Leifer et al. 1994), a satisfactory IR analysis should record a comprehensive whole of the users' real requirements. However, the authors mention that it is difficult to achieve this exhaustivity and this exactness for many reasons:

- the users may be unwilling to express their requirements (Land 1982, Davis et al. 1986, Reix 1995), because they don't want the changes that the new system will unavoidably provoke;

- the users are not always « aware» of their needs (McKeen et al. 1979, Leifer et al. 1994);

- even if the users know their needs, they will not always be able to express them (McKeen et al. 1979);

- …

Reading these authors' articles, one understands that the underlying general idea is that users do indeed have a need. Often, they either cannot express or even realize what those needs are. Therefore, the analyst's task is to reveal them. This idea according to which the requirement is «real» is always found in literature, even in the articles dealing with the prototyping process. Indeed, according to the authors, this technique (like the others) allows requirements that were pre-existing to the application to be revealed.

It is clearly evident that the authors consider the IR as positive and real data which is important to analyze. It is a typical positivism approach. Believing that it is possible to take a census of all the real requirements is to consider IR as positive data. Now, IR can also be considered as a construction. So the constructivism approach could also be used. The IR, far from being real, imposed and determined data, is, hence, on the contrary, data which is built by the subject, according to the way he or she interprets his or her background, in other words, according to his or her own interpretations. Consequently, one must study a subject's representations before studying his or her IR. If one considers the IR as a construction, the objective of the requirement analysis phase (the one of taking a census of the comprehensive whole of real requirements) becomes null and void. Indeed, the notions of completion and reality do not correspond to the notion of construction. Thus, the inability to take a census of a « comprehensive » whole of « real » IR is no longer a problem[i]. Furthermore, an exhaustive census of real requirements is not an objective any more as well. However, everything must still be done in order to satisfy the users. The objective of an analysis of the users' IR is to produce a list of requirements to be satisfied which consequently result in the elaboration of an IS satisfying the users' needs.

Since the requirement is considered as a construction, the analyst becomes a fundamental actor in the process of requirement determination. His objective can no longer be to take a census of all the user's real requirements, but rather to propose a list of requirements which result in the elaboration of an IS that should satisfy the users. It is favorable to the process that the analyst be aware that, by taking part in the process of requirements determination, he has an effect on the expected results. Actually, if the individual is conscious of the ineluctability of his influence, he is more able to reduce it than if he were not conscious of it. As Watzlawick emphasizes:

«One cannot not influence. It is, therefore, absurd to ask how influence and manipulation can be avoided, and we are left with the inescapable responsibility of deciding for ourselves how this basic law of human communication may be obeyed in the most humane, ethical, and effective manner.» (Watzlawick 1978, p.11).

Therefore, the analyst's influence can no longer be considered as a problem that must be removed. It is an inevitable notion. The objective is no longer to remove the problem of influence because there is no sense in it when a constructivism point of view is applied. Then the objective is to reduce the influence to a minimum.

This approach, based on the consideration of  the IR as a construction (constructivism approach) is not fully opposed to the approach based on the consideration of the IR as positive data (positivist approach). The aim of the former is neither to « revolutionize» the analysis process of the IR, nor to change in any way the practices, but only to propose a different way to cope with a classical problem. Nobody in the literature surveyed seems to have explicitly considered the IR as a construction. On the contrary, the words the authors use show that the IR is always considered as a positive data (for example, when they use the word «real» and all its derivatives). But sometimes the authors seem not to consider it wholly as positive data, for example when they talk about the problem of the analysts' influence or about the intervention of their mental models.

Thus it is time to connect a constructivist epistemology to the IR because it fits with the current beliefs. In order to demonstrate it, we propose the following plan:

- firstly the problematic of the IR should be set out,

- secondly an explanation as to why it is possible to consider the IR as a construct must be given,

- and then the thirdly, a substitution of the word construction for the word analysis must be made in order to describe the process of determining the IR.