System Design Methods

Rationale:

System design methods constitute the framework that enables the building of information systems to be treated as a disciplined engineering activity rather than as a craft. The range of methods available varies from very informal sets of guidelines to highly formal and mathematical approaches. While claims that some method is universally applicable are often made, all methods are, in practice, applicable only to certain classes of application.

All information systems engineers should be familiar with at least one system design method. This option is intended for the large numbers of candidates who need to go beyond a simple familiarity with the method used in their own organisation because they have to:

• choose a method appropriate for a new class of application
• monitor the implementation of a new method
• create a method appropriate to a special set of circumstances

Aim:

• To develop a wide understanding of system design methods,together with a critical practitioner's knowledge of at least one such method and a general familiarity with a number of others, not restricted to a single application area
• Objectives:
• Identify the weaknesses and limitations of proposed design methods
• Develop expertise in selecting a system design method (or combination of methods) appropriate to a given environment, identifying and taking into account all relevant factors
• Be able to assist in planning and managing the introduction of a system design method into an existing development environment
• Be able to assist in planning and implementing monitoring procedures to evaluate the effectiveness of a method in practice
• Identify areas in which changes to the method might usefully be introduced

Content:

1 BASIC ELEMENTS OF SYSTEM DESIGN METHODS

• Waterfall, V-model, spiral model, prototyping, incremental, reuse-oriented. The relationship between activities and life cycle phases; deliverables associated with each phase. The relationship between life cycle models and system design methods.
• Graphical notations including inter-alia dataflow diagrams, entity life history diagram, entity relationship diagrams, state transition diagrams, state charts and the UML.
• Formal notations based on mathematical logic and algebra.
• Techniques for validation and verification: reviews, inspections, walkthroughs, etc; automatic techniques.

2 CONSTRUCTION OF A METHOD

• The idea of the virtual machine underlying a design method and the way in which this affects the applicability of the method. Examples of methods illustrating the use of the techniques and notations listed in section 1 and also categories of methods such as structured, object-oriented, agile, RAD.

3 SELECTING A METHOD

• Technical factors: matching the method to the type of application; suitability of the method for use with existing software development environment; life cycle coverage; interfacing with other methods; tool support; comparison frameworks e.g. NIMSAD.
• Non-technical factors: how widely used is the method; documentation and training; availability of staff; how is the method supported; standardisation; track record.

4 INTRODUCING A METHOD

• Piloting and evaluating the pilot. Motivating staff who will be using the method. Role of consultants. Education and training. Reverse engineering of existing systems to fit in with the new method. Pitfalls.

5 EVALUATION AND TUNING

• Statistical process control as applied to the software development process. Appropriate software metrics: strengths and the dangers inherent in their use.
• Use of metrics to improve the software development process.
• The relationship between structured methods and software quality assurance.
• Assessing the benefits obtained through the introduction of a new method.