I saw an interesting article on the pro’s together with con’s of the waterfall model for software evolution at Builder AU. On the testing stage it states:
"Testing: In this stage, both private components together with the integrated whole are methodically verified to ensure that they are error-free together with fully run into the requirements outlined inward the origin step. An independent character assurance squad defines “test cases” to evaluate whether the production fully or partially satisfies the requirements outlined inward the origin step. Three types of testing typically convey place: unit of measurement testing of private code modules; organization testing of the integrated product; together with credence testing, formally conducted past times or on behalf of the customer. Defects, if found, are logged together with feedback provided to the implementation squad to enable correction. This is likewise the stage at which production documentation, such every bit a user manual, is prepared, reviewed together with published."
It desribes the Pro’s as:
"First, the staged evolution cycle enforces discipline: every stage has a defined start together with destination point, together with progress tin reach notice live on conclusively identified (through the role of milestones) past times both vendor together with client. The emphasis on requirements together with blueprint earlier writing a unmarried business of code ensures minimal wastage of fourth dimension together with endeavour together with reduces the opportunity of schedule slippage, or of client expectations non beingness met.
Getting the requirements together with blueprint out of the agency origin likewise improves quality; it’s much easier to grab together with right possible flaws at the blueprint stage than at the testing stage, afterwards all the components conduct maintain been integrated together with tracking downward specific errors is to a greater extent than complex. Finally, because the origin 2 phases destination inward the production of a formal specification, the waterfall model tin reach notice assist efficient cognition transfer when squad members are dispersed inward unlike locations."
And it describes the Con's as:
"Despite the seemingly obvious advantages, the waterfall model has come upwards inward for a fair part of criticism inward recent times. The around prominent criticism revolves around the fact that real often, customers don't genuinely know what they desire up-front; rather, what they desire emerges out of repeated two-way interactions over the class of the project. In this situation, the waterfall model, alongside its emphasis on up-front requirements capture together with design, is seen every bit somewhat unrealistic together with unsuitable for the vagaries of the existent world. Further, given the uncertain nature of client needs, estimating fourth dimension together with costs alongside whatsoever marking of accuracy (as the model suggests) is oft extremely difficult. In general, therefore, the model is recommended for role solely inward projects which are relatively stable together with where client needs tin reach notice live on clearly identified at an early on stage.
Another criticism revolves around the model's implicit supposition that designs tin reach notice live on feasibly translated into existent products; this sometimes runs into roadblocks when developers genuinely laid about implementation. Often, designs that facial expression viable on newspaper plough out to live on expensive or hard inward practice, requiring a re-design together with hence destroying the clear distinctions betwixt phases of the traditional waterfall model. Some criticisms likewise pump on the fact that the waterfall model implies a clear segmentation of labor between, say, "designers", "programmers" together with "testers"; inward reality, such a segmentation of labor inward around software firms is neither realistic nor efficient."