The FPS project can be said to be a new attempt I started, developed with the MVC three-layer architecture, the design of the backend BO layer, everything is so beautiful. However, once it is really entering the project, I found out that the initial idea is too simple, and the problem taken into account is too small, starting to enter the struggle.
Question List:
1. The originally envisioned interface cannot completely have existing requirements, it is necessary to change the way to implement this interface. In fact, it is only destroyed the pureness of the interface, but it has set an obstacle to the future maintenance, unclear, difficult to track .
2. A simple Java bean is assigned to the BO, a carrier of a data, a carrier of the background, and then passed to the TAG display to the front desk. The initial imagination seems to be good. However, when the code of Coding is increasing, it will find that the Process class is difficult to manage, and it is a static handler. When you encounter a public bo process, you must first see the Function you want. Does it exist, no you can add it. This is also difficult to avoid Function of different names of the same function.
Solution:
1. The overall planning of the project is more important, and now there should be some current development models, and the real use of UML has a detailed description of the system in the previous period, rather than thinking that it can be realized.
2. The entire process of operations uses BO to transmit the coupling of the system is too strong, the flexibility is not high, and it should be considered to isolate the BO object of the persistence layer and the display layer, and there is a beanutil tool. Complete the exchange of data.
3. Only when doing XML data import, it discovers the importance of the original interface function, I think of the original project, first define all interface functions first, then Coding, this is not a way to consider a value.