DEFINING QUALITY CRITERIA FOR SITUATIONAL METHOD ENGINEERING

Method engineering is a field of engineering that has been created a need expressed by engineers. These engineers faced problems in using conventional methods such as SSADM, DSDM, OOAD...In this way, Method engineering was created to intend to develop methods to meet these needs. Therefore, these methods have been employed in a wide range of projects led by these architects. In other contexts, we are led to customize a method for a particular project. The method obtained is called Situational Method (SM).


Introduction
In software engineering, quality is a concept based on several criteria such as completeness, functionality, reliability, and flexibility of use or otherwise. The concept of quality is pushed to its maximum in the finality of satisfying the users of software. In consequence if two software,that address the same problem, are going to be evaluated, the best, conceptually, is the one that has complied with the various quality rules applied. In the world of method engineering (ME), particularly SM engineering (SME), a profusion of methods can result from user's requirements without being able to say which is better. Indeed, in these methods, some parts may meet with some requirements but occasionally an overall view is missing.
To achieve the creation of a method, the experience of the Method Engineering Architect has a major role in the choice of components. From the selection of components to their compositions, MEarchitect is confronting a multitude of challenges to ensure their proper assembly. For this purpose, we have attempted to provide an efficient solution in [1] ensuring the smooth running of these tasks.
Our solution is based on the concept of quality criteria, which was taken from the world of quality in software engineering.
In this paper, we present the different criteria that we found useful to allow the architect to properly manage the quality of the method. We also classified these criteria according to the ways of Seligman. Moreover, we tried to map the relationship between these criteria and these ways.
Our approach is presented in this paper as follows: In section 2, we present work related to the concept methods quality. In section 3, we give an analysis on the world of methods quality. Section 4 describes the different criteria identified. In the section 5, we discuss our contributions to improve quality in SME. Finally, section 6 presents conclusion and future works. w w w . i j c t o n l i n e . c o m Some criticisms can be stated: the analysis of the tactics they are using is made of informal factors and thus should be considered as general analysis; -the list of tactics may seem arbitrary in terms of their orthogonality and level of abstraction; and; -the list may omit some important types of tactics, especially those used in other areas of methods.
Given that the concept of quality is treated exhaustively in the world of software engineering, we will build on work such as McCall [13], Boehm [2,3] or ISO 9126 [8,9,10,11] to determine the criteria that can help us in the selection and assembling method.

Method Quality Analysis
The ways of Seligmann Different definitions of a method have been proposed. According to [4,14,19,12] the main ideas converge on the principle that a method is based on a set of models and consist of a number of steps that must be fulfilled in an ordered sequence.
As defined by Grady Booch, a method is "a rigorous process to generate a set of models that describe various aspects of software being built using a certain well-defined notation."

Fig 1 The ways of Seligman
As shown in figure 1, Seligmann in [17] specified that a method is defined according to different ways: • The description of the visions of a methodology define the way of Thinking. This way is also known for defining the paradigm of a method; • The description of the models used throughout the development process define the way of Modeling. This way is also known for defining the different models used in a method; • The description of the techniques of support that are able to represent the different models issued from theprevious way define the way of Supporting. This way is also known for defining the different support tools; The way of Control (how): describes the management of the information system development process and its products. Therefore, we can see that a method is characterizedby: a process that describes the procedure (the approach). -a set of templates that defines the product that we want to achieve.

Our Approach
The major problem we may face is to build a method that responds to the expectations of the team, which will use it.
According to all these aspects, we may say that the notion of quality is very complex to implement. It is the convergence of three axes: (1) the product axis defines the result of applying a method;(2) the process axis provides process models or approaches. It is expressed as a set of interconnected and activities carried out in order to define a product and (3) the toolsaxis which defines the support by software tools.
This complexity results from a multitude of factors and actors involved in the method. For example, during the setting up the structure of the method, the architect must make choices to select the fragments of the methods. These choices are ultimately very important to users. If these users do not find interest in using a component of the methods, or worse, if they do not understand the interaction and integration of components together, they can move away from this method or one of these from components and to look for other alternatives. The architect must remain responsive to these end users to be sure to achieve a high level of assimilation of all components of the method and respond to feedback from these customers if there are any complaints.
It is important to notice that the notion of quality is associated to the methods applied and it depends on the context in which the method is applied. It is also in inference with the terms of use of the method; and how a user drives method to achieve his goal.
In addition, this increases the complexity of identifying elements that help to set up quality in methods.
Accordingly, we can give two definitions to method quality: • It can be defined as a satisfaction contract from the use of the method resulting; and; • It can be defined as a ratio between a set of criteria to establish, their appropriateness and the expectations expressed by the method designers and the end-users of the method.
As defined in [1] we have opted for the enrichment of the map proposed by Rolland / Ralyté / Deneckere [15] on the model definition the process of assembling of situational methods (figure 2).
The original map provides different ways of selecting fragments of methods that correspond to initial requirements as well as the strategies for their assembly.
Our extension of the model provides mainly two intentions in addition to the initial process map:

-
The first step comes right after the definition of the objectives of the construction method in "Specify method requirements" intention to enrich its objectives by providing the definition required for the quality characterization of the different methodological fragments of the final method. In order to ensure that valuation of the choices, we have defined a set of criteria for qualification and validation cited in the next section.

-
The second step is located at the construction of the method and the end of the process. This is due to a concern with evaluation and validation of the initial decisions.
Having said that, we should not forget the changes made to the construction phase of the method to enrich this construction by the rules defined in the previous steps.

Quality criteria
In this section, we will specify the different criteria related to the notion of quality methods. We have selected to use the ways of Seligmann to classify these different criteria from those ways. For each criterion, we will give a definition, followed by the ways of Seligman that are affected and we will end up by giving the different possible values as determined in [1]. These criteria are summarized in table 1 at the end of this section.

Discussion
The work that we have begun in this article is based on the concept of criteria. This is due to the lack of works covering this subject. At the same time, the question of quality was well fixed in software engineering world and this concept was defined and controlled at all levels.
Given that, methods are divided into several ways, for a certain way the quality is not assured optimally. For example, the working way that defines the process of operating a method, these mechanisms do not cover and we must ensure by other means.

Fig 3 Relation between Component criterion and method criterion
Other point to raise is that we must ensure the kind of relationship that exist between a method criterion and a criterion fragment ( fig. 3).

Conclusion and future work
As we have seen all along this paper, the concept of quality is a concept very coveted in the field of ME and particularly in SME. Having methods of quality, will guarantee a degree of satisfaction attained from the beginning of the use of this method and this satisfaction by maximizing gradually and as the method mastered.
Just like software, a method has to be designed to satisfy situational requirements including both NFR and FR.
In this paper, we propose the concept of quality of the method and we have defined with the way we understand the meaning of quality methods in the world. We intend to improve the way in which the operation is performed in decision making for the selection of fragments. We also plan to define an incidence matrix between the different criteria to determine if there is any relationship or influence between these criteria.