Research On Structure Characteristics And Information Metabolism Of Software Network

With the development of the speculation and the complexity of software systems, the interaction of the constructive parts makes the software system as a whole appear a new feature which contains in the construction and metabolism of the system. Software system is an effective means for the expression of human wisdom through computers. As the product of the thinking process of the brains, the network of software is a side-product of the construction of the brains. Physiologist Donte Chalf holds the idea that human brains are kind of complicated network presenting very distinct features of a small world which enables an effective connection. Thus the software system clearly presents complicated features as the network. The degree of the complexity of the construction is one of the most important factors which decide the quality of the software. Through the description of the complicated relations among the elements in the system, the software system has already presented a constructive method of multi-levels, different particles sizes, multi mode of collection. Through the research on the software network, it’s beneficial to know more about the basic characters and regulations of the system and it’s of great significance to realize the quality control and the generation development and reconstruction of the system.First, basing on the combination of the complicated network theory and the traditional measuring methods of the software project, using the original basic parameter of the complicated network measure, the paper introduces three measuring parameters and makes macroscopic and regular analysis on them, finally the paper brings the measuring method on the basis of statistics which use major measuring parameters to measure the important features and constructive characteristics from the three aspects of the individual quality, the relativity among the identities and the integral measure of the software and brings important proof for qualification of the software qualities.Second, the paper makes research on software network from the aspect of inheritance tree, and finds that the nodal points in the software network make the inheritance tree as the aggregation unit, and the number of the nodal points of the nuclear inheritance tree goes beyond20percent of the whole network. The paper uses the measuring methods of Fetion and Melton to measure the coupling of the inheritance tree and finds the measuring quotient of coupling of the whole system is comparatively high which goes beyond5, showing the fact of tight coupling. This shows that the coupling among the inheritance trees is in the form of tight coupling. The paper also makes statistics and analysis on the distribution of the characteristic quantity of the internal node inside the inheritance trees, and finds the linear relation between the degree of the nuclear inheritance tree and the audit, which means that the function and the reusability of the software are of very intimate relationships. The distribution of collecting quotient power rate of the degree and nucleus show that there is Hierarchical structure of modularization inside the inheritance trees.Third, using modules as the estimation standard of community division, the paper designs and realizes the searching methods for local community to make community division and analysis, and finds in the massive software the linear relations between the application times of the community constructions and the speculation of the software. Making analysis on the development of the multi version of the software, the paper finds that if the inner nuclear of the software has changed a lot, the application of the community construction will present a sharp change tendency. The software system is a special and complicated network and has the feature of community. Software design holds the principle of "High cohesion and Low Coupling" and at the same time packs the information, thus generally speaking, the object-directing software presents fine module constructions.Finally, the paper makes analysis on the information metabolism from the aspect of life features, and makes the comments that the software network are destined to have the information metabolism with the outside world to keep the stability. The paper picks up the model ware to testify that the software network continuously keeps the feature of free scale. Subsequently, the paper gives the definition to the standard structure entropy of software network and makes the counting and analysis on it, and finds the phenomenon of the minus quotient in speculation variation and version development, which is totally consistent with the information metabolism as one of the life features. In order to make further description on the ordering of the software, the paper defines the deviation rate of standard structure entropy (D (E)) and draws the conclusion through counting:when D (E)>0.05, the ordering of the software system is considered to be comparatively bad.