Analysis And Research On Quality Metrics Of All-electronic Computer Interlocking Software

The reliability and security of railway signaling equipment or system is of great significance to train protection of the normal operation. We must follow the principle of “fail-safe” in the process of designing the system. The computer interlocking system has been widely used in the railway signaling system, of which a core topic is safety. The safety integrity level of the interlocking software must achieve SIL4. In the development of the software, we can not ignore the reliability and the maintainability of the software, it will suffer the quality of the whole software. Therefore, it is critical to research on the interlocking software quality characteristics.This paper studies the all-electric computer interlocking software. First, analyses the architecture and functions of the interlocking system, which is a safety-related system so that its software must be reliability and safety, and proposes the concept of software quality and the metrics model of software quality, explains many techniques for guaranteeing the software quality such as software test, software metrics, code restructuring. Second, introduces all kinds of the representative software metrics methods, for example, software metrics for lines of code, McCabe metrics allowing for the program complexity, Halstead software scientific metrics for the program text complexity. We find these methods can not measure all aspects of the quality of the whole software through analyzing these methods.In order to improve the problem of partial metrics, this paper proposes the method that can measure more aspects relatively, which is combined with the software quality evaluation, and modeling the interlocking software quality metrics. The method allows for many aspects of software quality such as software length, structure, information flow etc, achieving the target of measuring all aspects of software quality. We refer to software quality metrics standards, select clarity, maintainability, testability as the interlocking software quality characteristics, and work out the value of each metrics element about quality characteristics, and based on the comprehensive metrics method forwardly. Then we build the static metrics of the interlocking software quality evaluation, and work out the interlocking software metrics in several cases and calculate internal quality of the software in the way of quantization. Through analysis and calculation, some metrics below its minimum, while others overrun its maximum. This will suffer the software overall quality.In order to improve the software quality, this paper proposes a method of Quality-Metrics-Directed Optimization (i.e., QMDO). We come up with the following key idea of QMDO: we modify the program by using code optimization techniques, while referring to its metrics results, and then measure the quality of the program modified. Contrasting with the first quality metrics results, we find each metrics items can reach their scope required. The QMDO obviously improves the clarity, maintainability, testability of program codes, at the same time makes the quality of codes better, and provides an effective way to improve the software quality through software metrics.