CSE is an approach that stresses he need to build correctness into the software as it moves through product and project life cycle. It rejects the classical model of development, i.e. analysis, design, code, test and debug. The philosophy of CSE is: write the code first time so perfect and verify its correctness before testing. It does not advocate debugging as a method for improving the correctness of software. The principle is “Do it right the first time” so that no mistakes are left in specifications, design and code. It does recommend testing for discovering the mistakes/errors/but wants to use it to verify the correctness.

CSE is supported by two main strategies, use of the incremental software development model and statistical test planning and use testing. It recommends the incremental model, as it is best suited to implement the CSE philosophy “Do in right the first time.” The CSE process can be described as under:

•    Break RDD into small components that can be built in a logical sequence and complete the software product by integrating the components.

•    An increment, a component complete in itself, can be treated by white box and black box methods for development and variation of its correctness.

•    An increment goes through a formal development process, design, design correctness verification, code generation, code review and inspection, followed by statistical testing for certification of correctness.
It is necessary to explain statistical test planning and statistical use testing, which are the key features of CSE. For statistical test planning, software usage is forecasted gives a probability distribution of usages of the software. using this, probability distribution test cases are planned and designed to use in each incremental phase. In view of this approach, each increment is correct in design and application given valid results.

Statistical use testing is resorted to select those usages that are required by most users in the organization. The basis for statistical usage testing is usage and their users. The increments are used for test case design based on the probability distribution of incidence of software component usage. In other words, in the CSE test plan is included test cases of usages that are used by most users.

After resorting to statistical test planning and statistical use testing, each increment is verified for correctness. This eliminates the risk of discovering he errors in the real time usage. It also means that the increment is ready for integration.

The CSE approach is different from the conventional SSAD and OOSAD approaches for the following reasons:
•    It uses principles of statistical quality control.
•    It verifies design specifications before coding with mathematical accuracy.
•    It uses incremental development model, where increment is a box abstracting a system or sub system.

CSE uses the box structure specifications similar to the white box and black box used in software testing. A box representing an increment is an abstraction of complete module/sub system/application lending itself for full verification to test the correctness. The box cannot be termed as small or big. What it required is that it models the complete behavior of the increment of ease of integration into the next increment. A box has three sub-boxes, namely, black box, state box and clear box. These sub-boxes have a specific role in development:

Black Box specifies the behavior of the system. It specifies that responses to each input stimuli. The input stimuli could be events or triggers and its behavior will be governed by a set of rules applicable to process the input. The Input – Process – Response (ouput) is mathematically correct, as each response is deterministic and tangible to ascertain the correctness.

State Box models each of the state of the black box, where each state leads to another state through transition. For each transition, there is an input, rules to process, and response indicating the new state.

Clear Box models the procedural steps required to executer each state to transit to a new state.

In short, black box defines the module/subsystem/application. For each such black box, several state boxes are defined, showing state transitions based on input and rules of processing. Further, for each state box, there is a corresponding clear box that defines the procedure of state transition processing.

CSE is a formal mathematically oriented approach to software development. It uses the box structure, based on a hierarchy of three box type, black, state and clear. This approach is used for analysis, design and modeling of a system component. This component is developed where correctness is verified at the design and coding stage before testing. In this approach, we resort to statistical methods of testing and quality control, the using probability distribution of usages and users of the software. The key driving principle of CSE is correctness verification at each stage of development and not keeping it handing till testing is undertaken. In other words, software developed using the CSE approach will have very few or no errors at the time of delivery to the customer.

For more help in Clean Room Software Engineering click the button below to submit your homework assignment