اولویت انتخاب نواحی فرآیند بر اساس نمایندگی مداوم CMMI

An essential decision that must be made by software organizations that adopt the continuous representation of the capability maturity model integration (CMMI) for software process improvement concerns a suitable path that best meets their business objectives and mitigates the organization's risk. However, the CMMI models released by the Software Engineering Institute do not give their adopters any guidance on how to make such a decision. Thus, managers often make subjective selections of the areas in which to implement process improvement. Our study presents a decision support model that assists managers in determining the priorities of the CMMI process areas based on the characteristics of the is being developed. The proposed model was validated by using the ISBSG repository, and an example is presented to demonstrate the application of the model. Given the fact that hardly any research has yet been done on how to select the CMMI process areas to initialize process improvement, this study provides a starting point for the community in considering this important issue.

A well-defined software development process can improve software quality and reduce the risks of software development projects. The Software Engineering Institute (SEI) at Carnegie Mellon University firstly released the capability maturity model for software (SW-CMM) Version 1.0 in 1991 to evaluate the maturity of software development of US Department of Defense contractors and provide a roadmap for software process improvement (SPI). The SEI solicited feedback from its adopters and then released SW-CMM Version 1.1 in 1993 [23]. Thereafter, SW-CMM Version 1.1 became a widely accepted benchmark for software organizations in both the initiation of their SPI efforts and evaluation of the maturity of their software development processes [1], [8] and [16]. According to the SEI maturity profile in August 2004 [25], software organizations spent an average of 2 years to raise the level of their process maturity if they choose the staged representation of a CMMI model. Since SEI released SW-CMM Version 1.1, it has been applied to different areas; hence, many capability maturity models have been announced. These included the software acquisition CMM (SA-CMM) [6], system engineering CMM (SE-CMM) [11], integrated product development CMM (IPD-CMM) [26] and people CMM (P-CMM) [2]. These models were developed by various organizations, and so they had overlapping scopes of applications and lacked consistency in architecture, terminology, and assessment methodology. These problems have increased the cost and time to implement multiple model-based process improvement. Therefore, SEI released capability maturity model integration (CMMI) in 2001 to integrate existing capability maturity models. Its primary advantages are: in the elimination of the inconsistencies and duplication and thus streamline the enterprise-wide process improvement, in the reduction of the cost and time associated with model-based process improvement, and thus in the increase of return on investment of organizational SPI efforts [3] and [4]. To accommodate different process-improvement needs for software organizations, the CMMI product team provided them with two choices (staged or continuous representation) to increase the maturity of their processes. The CMM staged representation provided a framework for organizing the evolutionary steps into five levels of maturity (initial, managed, defined, quantitatively managed, and optimizing). These are ordinal scales for measuring the maturity of an organization's software process that can also be used for its internal process improvement. SEI then added the continuous representation to the CMMI for providing flexibility to enable software organizations to choose their improvement paths. The continuous representation allows comparisons of a specific process area across software organizations. This representation therefore allowed process improvement to be compared with that of the ISO/IEC 15504 standard. When a software organization adopts the continuous representation of a CMMI model, the foremost decision of the project manager is to select the order of implementing the process areas that best meet the organizational business objective [4]. An unsuitable path can negatively affect the results of the activities, thus reducing enthusiasm for software process improvement. However, the CMMI models currently released do not provide any guidance to software organizations adopting the CMMI continuous representation on the priorities of developing the process areas. Recent reviews of empirical studies of software project management have indicated that a software framework must be configured according to the specific characteristics of the project [14], [22], [24] and [27]. Also an improvement path must be configured according to the specific software development environment. The characteristics of the software being developed crucially influence the choice of the path. We therefore designed a decision support model based on information system characteristics (ISC) to help adopters choose a suitable improvement path for their SPI efforts.

Pursuing efficient and effective software process improvement is a major goal of software organizations committed to producing high-quality software within a reasonable budget and time. Many software organiza- tions worldwide are adopting the continuous represen- tation of the CMMI model. However, they frequently encounter difficulty of selecting suitable process areas in which to initiate their effort. The current CMMI models do not, however, provide any means of determining the priorities of process areas; thus managers often make a subjective decision when selecting the CMMI process area. Many factors, such as business objectives, project risks, and available resources, can influence an organization’s decision of an appropriate path that is suited to its environment. Thus, building a formal model that captures the factors that affect the selection of the CMMI process areas is difficult. Our model is intended to aid in this process. It was validated by using the ISBSG Repository, and an example provided. The validation revealed that significant rank correlations exist among process areas across eight types of information systems in the ISBSG, indicating the soundness of our model