Author: Anders James
Currently, the object-oriented design method is used. The design granularity is divided into two levels: class and method (attribute), similar to the table and field of the database design;
In the existing implementation system, a method can accommodate multiple use cases. At the same time, because of the horizontal expansion of use cases, a use case will be associated with multiple methods; this is a many-to-many relationship that brings huge costs to our development and management.
To effectively manage use cases and their implementation ing, AOP technology becomes a good choice. AOP allows us to establish an independent and manageable design granularity for each use case: upgrade a code segment from a method to an independent method and class, and allow these use cases to be organized reasonably and orderly.
The existing technical system has established feasible solutions for us. How to organize logical operations between use cases: And, or non-operations becomes the key to implementation.
In existing practices, the logic operation of non-business use cases is clear: the execution sequence is also clear with operations (the change sequence hardly affects the business correctness ), AOP has been widely used. Due to the unclear logic of business operations, there are also uncertainties in the execution sequence. At present, it lacks proper practice management;