PIP: Progressive Implementation Pattern

When developing a persistent, distributed, and concurrent system, implementation and tests are usually hard. During tests, database, distribution, concurrency, and functional errors might appear at the same time, increasing debugging complexity. When using EJB [10] as the persistence and distribution technology, the deployment time might be very high. To fix errors — including functional, persistence, and concurrency control errors — we might waste a lot of time by compiling the code and them deploying the system into the application server. Another problem happens when using a database to persist data. We might have to write specific programs to check if the data stored into the database conforms to the expected results. Similarly, if the system can be concurrently accessed, programmers should worry about concurrent executions when implementing functional requirements, increasing programming complexity.