Chapter 7: Models Introduction 7.1 Model progression 7.2 How to model 7.3 Why use models? 7.4 Conclusion
Introduction Captures the essential aspects of a system or process Has to be clear and concise Using mathematical notation is ideal but not essential Model may Provide different views on the same system Describe a system as consisting of various components Examples ISO OSI Mathematical model for access rights: Tuple (s. o. m)
A model for the interaction between computer attitude and computer anxiety
7.1 Model progression Tentative model (Clarification stage) Differentiated model Address specific form of the problem General model Models that cater for most of the assumptions made by earlier models
7.1.1 Clarification A new problem arises that begs for a model to clarify it. For a new paradigm New solutions for problems that have already been solved New solutions for existing problems
7.1.2 Differentiation Are the proposed models adequate? Investigate the assumptions, starting points and restrictions of proposed models Try to propose a model that differs from existing models Motivate why it is better Why is your model needed?
7.1.3 Generalisation Develop a model that can do everything that all the others can do Sometimes surprisingly simple, e.g. OSI
Work, Insight, Iteration, and luck 7.2 How to model Work, Insight, Iteration, and luck By design Use common design methods and principles Identify major components of a software system Use dataflow techniques By metaphor Analogy to the real world: Actors, brokers, clients, even zombies, files, windows, editors, etc. If car is a metaphor for something in the computer world, then driving, parking, servicing, crashing, washing or towing will be applicable as well. Sometimes the real world example resembles the computing concept so closely that the metaphor is obvious By formalisation Choose tool carefully Alternatives: Set theory, logic, algebra, formal languages, automata theory By genius / luck Think of your problem before going to bed and sleep on it
7.3 Why use models? Simplicity Comprehensiveness Generality Exactness Miller’s rule: A human can only have 72 concepts in short term memory at a time Comprehensiveness Models can systematically address all aspects of a problem Generality Models can address numerous variations of problem Exactness If the model fits the problem closely, it is likely to be accepted Clarity Purpose of all components and operation of every component, interaction between components should be evident