- Feasibility study
This is where we carry out and gain an understanding of the customers current system and document using DFD’s, entity modelling and normalisation.
- Requirements analysis
This is where we break down what are our new system requirements are e.g. software and hardware needs.
- Design
At this stage we would make an overall design of the system architecture and any faults should definatley be identified before moving onto the next stage. Build project dictionary.
- Coding and testing
The programmers now begin programming the code and beta testing it. We would also train the people who will be using the new system.
- Conversion
This is also sometimes called integration, where we assemble all the pieces of the new system together and see if it matches up to the original system requirements.
- Post implementation review
This is where the software would be delivered to the customer and is ready for acceptance testing. Operating documentation will be generated and the procedures for the system will also be documented.
The main benefits of using an SDLC such as the waterfall model (some people prefer to call it a method) are it provides a structured approach to design, from the initial need for a new system to be recognised to the detailed lyses of the business and practises of the business a systems analysis can take an in depth look at how things relate to each other, remove any duplicity with in current systems and the employer can (for many for the first time) see how their business or computer system works. It is therefore also an ideal time to restructure or reorganise a business or system. This system can also be incorporated and is known as PISO (process improvement for strategic objectives).
The main downfalls of the waterfall model are that the waterfall model assumes from the outset that you know exactly what kind of system you would like and by doing this it leaves no scope whatsoever for making late changes or implementing new ideas at a later stage. The standard methods for calculating cost of the system are calculated using COCOMO developed by Dr. Barry Boehm while he was at TRW [Boehm, B., 1981:] Software Engineering Economics, Prentice Hall]. Generally this states that by every stage the cost will multiply by a factor of 10, so the outcome of this is that the Waterfall model is an expensive method to use.
To not use an SDLC would be suicidal, as many people have found out to their own cost in the past, to simply implement ideas and software without using a solid, tried and tested method is I suppose like jumping into a car without an idea how to drive. To design and implement a new system you must use a life cycle model whether or not the Waterfall model must be determined, there are many others.
To compare the Waterfall model against two other models I have drawn up a table, on this table I’ll compare the Waterfall model to the Incremental and Prototype models.