meeting -*- Outline -*- * Inception (Larman, Ch. 4) Q: What are the goals of the inception phase? ------------------------------------------ INCEPTION (LARMAN, CH. 4) The best is the enemy of the good. -- Voltaire Goal is to explore: - Vision and business case - Is the project feasible? - Buy or build it? - Roughly estimate cost (order of mag.) - Proceed or stop? Get agreement from stakeholders on these ------------------------------------------ Have to explore requirements for this, but don't want to define all requirements, or even generate believable estimate Idea: it costs time and money to estimate a project seriously, this step is to see if that should be done Q: How long should inception take? A week? A month? A year? not much more than a week it can be very brief or omitted if this kind of project has been done before ** an analogy (4.1) Q: Is this like anything in real life? consulting with a doctor to do exploratory surgery or deciding to do exploratory drilling for oil ** artifacts (4.3) Q: What are the artifacts for inception? ------------------------------------------ INCEPTION ARTIFACTS (4.3) Key idea: - only partially completed in inception, - only create if will add value, - content should be light Possible Artifacts: - Vision and Business case - Use Case Model (names of cases) - Supplementary specification - Glossary - Risk List and Risk Management Plan - Prototypes - Iteration Plan (for first elab. iter.) - Phase Plan and Software Devel. Plan - Developement Case ------------------------------------------ The programming of prototypes should look to clarify a few key requirements (typically in UI) or to look for techical "show stoppers" The point is not the artifacts, but thinking, analysis; the artifacts should just record these Record objects digitally and on-line rather than on paper ** questions (4.4) Q: Should you define the software architecture in inception? Q: Should there definitely be a Business case or Vision artifact? Larman seems to contradict himself here, saying no above, yes here Q: Should the names of most of the use cases and actors be identified? yes Q: Should all of the use cases be written in detail? Some of them? Larman suggests 10-20% of them