針對信息遞送程式 做可及性測試的研究

Abstract

軟體測試一直是我們最重要的課題，因為軟體出錯的比率不只是我們人類所撰寫的程式語言會出錯，就連系統執行所有程式時所不可能預期的錯誤也會發生。在現實的情況中，通常我們所執行的程式都是concurrent program，很少程式是單一直行的程序，所以軟體出錯的機會很大。 黃冠寰博士提出了reachability testing[1]，此篇主要的研究是根據reachability testing 做一個核心的延伸，所用的架構為 message passing 的模型，依照reachability resting 的精神下去做研究，並且提出一個解法，利用unique message id和filter 增進reachability testing 的功能， 專門針對multiple receiver 的 synchronization sequence 產生race variants 做過濾，進而得出符合synchronization model 的race variants，此篇論文主要針對兩種message passing 模型進行分析，分別為synchronous message passing 和asynchronous message passing，並且針對Richard H. Carver 和Yu Lei 的work[5][17]提出一些不同的看法和質疑，另外我們發現改良Richard H. Carver 提出的race table 的方式去解決multiple receiver 的問題。

Software testing is always an important problem during our life, because the big chance of software which has errors is not only in our writing programs, but also in the achine system programs which execute all the times, and then all the nondeterministic errors occur. In the reality world, we usually execute concurrent programs, the very few times to execute a sequential program, It is the big chance that errors will occur. Dr. Gwan-Hwan Hwang has proposed “Reachability testing”, this paper is follow this work and view this major study is based on a core of reachability testing an extension of the framework used by the message passing model, reachability testing in accordance with the spirit of this research, and propose a solution, using unique message id and filter to enhance the function of reachability testing ,we focus on the race variants of synchronization sequences of multiple receiver, this paper mainly two kinds of message passing model for analysis of synchronous message passing and asynchronous message passing, and for Richard H. Carver and Yu Lei’s work[5][17] to make some different views and queries. In addition, we found that improvement of Richard H. Carver made race table approach to deal with multiple receiver problems.