杭州交通信息网冷冻食品针式温度计:谁能帮忙翻译一下这篇计算机文章?
来源：百度文库 编辑：杭州交通信息网 时间：2024/04/29 16:17:58
In a production environment, a set of test data should be created, preserved and maintained for each program. Any time a modification is made to a program, however minor the patch, the program must be retested on the entire test data set to ensure that the correct results are still produced.
This use of specially designed test data is usually preferable to the testing of programs on ‘live’ data. The use of live data, that is, data which has been taken from the actual system in which the program is to operate has a certain attraction in the process of program testing. Live data contains real instances of the types of input which the program must process rather than data merely contrived for the testing procedure alone. However, it is highly likely that whatever subset of live data is used in the test, it will not contain all of the combinations of conditions which the programmer needs to test. It is also unlikely that boundary, or threshold situations will be present in a random sample of live data. For example, if tax is to be calculated on gross pays over, say, $200, then the program should test that specific $200 value to make sure that the program has dealt with it correctly. The program may have been written to tax incomes of ‘$200 and over’ instead of ‘over $200’. If a very large volume of live data is selected as test data in the hope that it will include as many variations of conditions as possible, the volume of the output from the test will be so large as to preclude the programmer from examining it in sufficient detail to detect subtle errors.
This use of specially designed test data is usually preferable to the testing of programs on ‘live’ data. The use of live data, that is, data which has been taken from the actual system in which the program is to operate has a certain attraction in the process of program testing. Live data contains real instances of the types of input which the program must process rather than data merely contrived for the testing procedure alone. However, it is highly likely that whatever subset of live data is used in the test, it will not contain all of the combinations of conditions which the programmer needs to test. It is also unlikely that boundary, or threshold situations will be present in a random sample of live data. For example, if tax is to be calculated on gross pays over, say, $200, then the program should test that specific $200 value to make sure that the program has dealt with it correctly. The program may have been written to tax incomes of ‘$200 and over’ instead of ‘over $200’. If a very large volume of live data is selected as test data in the hope that it will include as many variations of conditions as possible, the volume of the output from the test will be so large as to preclude the programmer from examining it in sufficient detail to detect subtle errors.
在生产环境中,应建立一套测试数据,保持和维护每一计画. 任何时候,修改了一个节目,但在小片,计划要在整个测试数据测试,以确保正确的结果,仍制作. 利用特别设计的试验数据,这通常是最好的测试程式「活」资料. 使用活数据,数据已经从实际的计划体制运作有一定的吸引力正在测试计划. 生活中真实情况的各种数据资料,计划必须是过程,而不是人为的实验数据仅程序. 但是,很有可能什么活子测试用的数据,但不包括所有的组合情况,需要测试程式. 这也是不可能的边界或界限的情况将在现场随机抽样数据. 例如,如果税率计算总值超过自付,譬如$200,那么测试计划应具体价值$200,以确保该计划已处理正确. 该计划可能已写信给税务收入的$200,在'不是'超过$200'. 如果大量生活数据选择测试数据,希望这将包括许多不同的条件越好,商品生产的考验将非常严重妨碍程序员检查它足够详细探测细微的错误.