"Dream Break Code" read Note 03

"Dream Break Code" reading notes

The book says the most common problem with software development is that "the project is progressing far behind schedule". The Chandler program was released in 3-4 months, but each version took more than 6 months, and there were many reasons. The first appropriate measure of development progress itself is a very difficult thing, which means the plan itself is too harsh. Even testing the progress of software development is a very painful thing, with the number of code or the number of defects to measure biased, the text mentions the Mbwa method, but this method is difficult to get a general development progress. Next is the software development plan often goes beyond the foreseeable scope, causes the software development one to stay in the design stage, quoted the text the sentence, "With today's tool and the process, plus yesterday's memory limit, we really can do better". Another is the software flaw, Chandler in the development process seems to fall into the mire of defects, they spend a lot of time to repair software defects, how to reduce the software development process defects is a big thing.

In the chapter on design, I've seen the edge cases mentioned in the book, which is what we do when we're doing a pair of sub-array maximums in the array in development that exceed the Int32 representation range. Programmers are trained to be exhaustive, they are too obsessed with being wary of problems, and the results will inevitably be dull. They racked their brains on the fringe case, so much that they strayed from the center store. And I also feel that, our current study, and do software still have very big difference. We first consider how to achieve the teacher requirements of these functions, and did not think of users as some people do not understand the program to see, we do the program, the only user is the classroom teacher.

We always want our projects to be the best, but in fact, as long as the parts are good enough to make the whole project the best. The project cannot be completed if the local is also optimal. Of course, another interesting detail of the book is the scale of the software engineering methodology, a total of 12 items, and the book explicitly mentions Microsoft's full score. The book is an amusing metaphor for the way it is in the army. Although cumbersome but ensures that programmers are always at their best, entering this state fully illustrates the enormity of the current Microsoft Empire and how cumbersome it is to sustain such a huge need.

"Dream Break Code" read Note 03