Theoretical foundation for incremental development

zhaozj2021-02-08  501

The theoretical basis of incremental development is not allowed. It is strictly forbidden to reprint this section This paper is reproduced from the Software Engineering Expert Network www.21cmm.com. Do not reprint without CBS license, thank you for the increase in practice in cooperation Massive development techniques are based on the principle of reference transparency. Quote Transparency means that the only thing that an expression is concerned is its value, which can be replaced with a sub-expression equal to its value. Reference transparency suggests that the low-level details of an entity are abstract, rather than ignoring the specific system described above. Therefore, when it is in large context, each of the high-level descriptions must understand the expression entity. The concept of citing transparency has been applied in a wide range, including: computer science, linguistics, mathematics and logic. The reference transparency cited transparency in the arithmetic, ensuring that the estimation of lengthy arithmetic expressions will produce the correct answer. For example, in the expression (6 2) * (5-3), since 6 2 = 8, the expression is equivalent to 8 * (5-3). In addition, because 5-3 = 2, it is also equivalent to 8 * 2. Finally, because 8 * 2 = 16, it is equivalent to 16. Reference transparency guarantees no conditional use 8 instead; if you do not have to worry about whether or not the result will continue to be multiplied (5-3), or subtimate from 17, and so on. The reference transparency is only established in form arithmetic logic. Note that because of the reference transparency, it is possible to further answer each step. Therefore, we are full of confidence that in three steps, the first step is completed, the problem is 1/3 has been completed. This is, completing the follow-up step, the previous steps do not have to be re-accessed. In addition, since each step is not negligible to other steps, the panel can be evaluated in any order. Therefore, this characteristic is similar to important software concepts in abstraction, specification, solutions, and no negative effects. The basis of reference transparency software incremental development in the software is to develop math function rules (linger, mills, and witt1977) for programs and program parts. This view considers that the program is developed as a function improvement (specification) process as a self-turned control structure or subunies (sub-specification), which will result in object or function decomposition, or binding. For example, a given function (specification) f can be improved to any of the form: Do F1; F2 enddo order IF P THEN F1 ELSE ENDIF Select the While P Do Endo loop in which F1 and F2 indicate a more further improved subunies (Sub-specification). Subsequent function improvements must maintain the function equivalent of each step correctness. For example, in the above order improvement, the merger of the subunies f1 and f2 must be equivalent to the original function F. The reference transparency requires an arbitrary function (F, F1 or F2) to completely define the effects required in this point in the design, and there is no additional improvement to other design portions. Because of the reference transparency, the verification of the refining step can be performed independently of the other refining steps, which means that in the early increment before the system is written, the system can be verified. In the later increment, the architecture does not have to be re-verified. However, it should be noted that the norms of entering the system certification system components have actually provided accurate interface documentation, when coding is completed, ensuring that the system achieves the required performance as a whole. Reference transparency is refined in the program design grade function, such as a graph. The two-step refinement on the right side of the figure maintains the equivalent between each function. Preferences, refine the initial specification for F cycle initial code G and sub-specification K, where K specifies the processing effect required for an instance design of interface and graph refining reference transparency. Second, the second step refined sub-specification K is a cycle. These expansion steps are referenced by transparency and represent possible incremental definitions.

转载请注明原文地址:https://www.9cbs.com/read-571.html

New Post(0)