The Systemic Evaluating Method About The Software Process Ability

At the present, a major problem is how to control the complexity of the software development. Now the the method of controlling the process is mainly used to guarante the software quality. The Capacity Maturity Model (CMM) is a process model of the software development. To achieve the quality of the software development control, CMM advocates a strict process . CMM evaluates the software developing process through the criterion extent of software developing process documents. There is no unified data measure to judge whether some process is good or not。 This paper attempts from the system angle to discuss the software developing process. Here a method of evaluating the software developing process with data is offered.About the difficulties of software development, there are four essential question : the software complexity, uniformity, variability and invisibility. By analyzing the software crisis characteristics we find that the essential reason of software crisis lies in the software producing process complexity. In fact, this kind of complexity also exists in other research fields For example: large-scale system of industrial production, social system, economical system, education system ,etc. How do we discuss and research them?I find the industrial production process is a fixed logical repetition work with comparing the production process of both industry and information . The main process of the information production is the software production, however, the software production is a logical creation process. It is this process that that makes the software production complex. In the discussion process of the large-scale system of industry, social system, economical system and so on process, I gained significative things from the discussion in the point of system with systemic science methods.This paper also discusses software production from the point of system, then proposes the relation between software process and software production system. In the actual software production, each software production system has only one production process. Because each software project group is produced by different organizations. As a result, the discussion about software production system becomes the discussion about software process ability. Therefore, this paper concerns thermodynamics entropy, the statistical physics entropy and the information entropy concept, and makes the entropy concept further promoted. I think the entropy is a state variable that describes each subsystem condition in big system. So believe the software process state variable is made up of the quality entropy, the time entropy and the resource entropy .Meanwhile ,I say the software process entropy is the linear combination of the three entropies above.This paper further discussed the method about evaluating a process ability with the state variable offered above. That is we compare each software process entropywith the standard process entropy, if its compatative dispersion exceeds the allowable scope, it shows that this process is different from our standard production process. This difference is possibly the performance optimization or possibly is the performance deterioration. And described the gathering method in detail with which the data was measured in a project as well as the selecting method of standard production process entropy SO . Thus in an organization with multitudinous software projects, each project will have a value appraisal to all the projects process ability after completing each project. Through comparing this value with the standard So, we can see whether this process and the standard process have a big difference . If that case, the upper level organizer have to go to all troubles to analyze the reasons.The computational method proposed in this paper is a very formulated method. With this method we may compile an automated software to achieve the appraisal to automatedly realize the process ability and intercalate the auto-alarm function if it surpasses the standard. We can estimate the process ability by a data measure.