Microsoft Research Institute also launched a new model verification software, ZING. Sketching about Zing, as if you sell some models that can handle larger models, there is a universal framework. In other words, there is a highlight. But did not see what specialized in theory.
Simply simply science what is software model test. Simply put, it is to automatically detect a software, or a software model (such as a UML model), is it correct. The advantage of automatic detection is exhaustive, not tested, always missing (don't forget EWDIJKSTRA big Niu name: Test can only prove the existence of errors, but not to prove the wrong goodness); Automatic, no need Intervention; and strict, less blurred and secondary. It seems that it is not realistic? Oh, there is a ready-made inspection software, such as SMV, NUMSV, and Spin. Microsoft even automatically analyzes the Windows driver with SLAM and discovers a very concealed error. The research and development of model inspection software is quite different from ordinary software development. The model testing research mode is from bottom to: first strict definition, then theorem, and finally launched a strict algorithm according to the theorem. Reflective system research is often heuristic. As for which one is more effective, I will see the benevolence.
In fact, hardware companies have long been examined in the model. After all, it is difficult to ensure that the key system is not lost. I spent a lot of money on various model checking software like Motorola, IBM, Intel, Nasa, Lockheed, etc.
It seems that the software is now more and more attention to the model test. After all, in theory, if you can find errors in the modeling phase (demand analysis, function planning), it is much more economical than in the test phase. Or if you can automatically check the code, find an error, and generate an example of generating an error (how many programmers dream! :-). Including Tony Hoare? (Quick Sort's inventor, Algo60 author, Hoare Logic author) also said in Microsoft research. In 1996, Israeli cow A.Pnueli got the Tu Wan, which was because of the basic theory, timing logic, and aspects of the model detection. The academician of Tang Tang is also a master in this area, it is said to be good than A.Pnueli's private hand.
Last to the theoretical basis of model inspection. Mainly timing logic. Think about it, you can use strict logic to derive things related to time, or cool. Secondly, the Boolean, the grid algebra, and the automaton theory (not the finite automatic machine for compiling theory, but infinite automaton). // Sigh, regret it doesn't work hard to learn mathematics.
