Prototyping

Cost over-runs and failed projects are often the result of building the wrong solution, not validating required features, and/or not evaluating usability. A computing solution, regardless of how technically elegant and acknowledging the best intentions of the developers, is still a huge assumption until it is verified. Organizational factors and other technical implementation risks, if large enough, can sink a project. Major risks must be assessed early and throughout the project and dealt with promptly so alternatives, if required, can be developed. Sometimes major changes to the project are required. Then the earlier the need for such changes is recognized, the greater is the likelihood of project success. Prototyping of solutions is a useful means of testing design assumptions, eliciting requirements, evaluating usability, and assessing alternative solutions to major risks. Web sites can be prototyped with site structure maps, storyboards, and other graphics. Languages such as Visual Basic or Delphi permit rapid development of user interfaces that can be linked together to simulate look-and-feel of the end product without writing the actual implementation code. Application users are almost always better at identifying what is wrong or missing in a program than they are at specifying all features and functions up-front. Simple user interface facades are excellent at refining and validating application requirements. Data warehouses make quite different projects, yet prototyping options exist. Dimensional designs can be tested and the concepts of multi-dimensional thinking taught by building scaled-down "cubes" in a commercial multi-dimensional online analysis processing (MOLAP) tool for demonstration to key business users and sponsors. A star schema data warehouse design, however elegant and well-thought-out, may not work in practice due to issues with the operational data to be used to populate it, and/or business rules that govern how data is generated and entered. All operational data sources must be profiled and summarized early in the project. Business rules must also be documented as early as possible. Then, prior to developing the routines to populate and maintain the warehouse, simulations and walk-throughs are required to validate that the routines to populate the warehouse can be written on-time and within budget. The old cartoon of the firing squad being commanded "Ready! Fire! Aim!" must be avoided. A computing solution must be based on a design, and a design that has been tested to catch major errors, omissions, and false assumptions. Prototyping accomplishes this goal.

Arynth, Inc., 101 Route 130 South, Grant Building, Suite 460, Cinnaminson, NJ 08077. Tel: 856-829-7222; Fax: 856-829-7618.