Ross Jeffery When, as a result of pressure from the CEO, the Chief Information Officer poses the question aJust what is this information system worth to the organization?a the IT staff members are typically at a loss. aThatas a difficult question, a they might say; or awell it really dependsa is another answer. Clearly, neither of these is very satisfactory and yet both are correct. The IT community has struggled with qu- tions concerning the value of an organizationas investment in software and ha- ware ever since it became a significant item in organizational budgets. And like all questions concerning value, the first step is the precise determination of the object being assessed and the second step is the identification of the entity to which the value is beneficial. In software engineering both of these can be difficult. The p- cise determination of the object can be complex. If it is an entire information s- tem in an organizational context that is the object of interest, then boundary defi- tion becomes an issue. Is the hardware and middleware to be included? Can the application exist without any other applications? If however the object of interest is, say, a software engineering activity such as testing within a particular project, then the boundary definition becomes a little easier. But the measure of benefit may become a little harder.Such a valueneutral investment may reduce the ability of testing to efficiently adapt to changes in the design of the system under test (Kaner et al., 1999). Furthermore, the initial effort for automating tests may be better invested in manually running different tests, e.g., ... important requirements first, value-based test planning focuses tests based on desired system parts, quality attributes, and project risks.
|Title||:||Value-Based Software Engineering|
|Author||:||Stefan Biffl, Aybuke Aurum, Barry Boehm, Hakan Erdogmus, Paul Grünbacher|
|Publisher||:||Springer Science & Business Media - 2006-02-23|