Software durability : Is it important? Can it be achieved?
Susan G. Archer, Laurel E. Allender, Celine Richer · International Conference on Human-Computer Interaction · 1997
Software products are always undergoing revision. Successful products are expanded to incorporate new functionality, adapt to new platforms, and appear to new markets. Less successful products are modified to fix bugs and to make the product more acceptable to the current users. In either case some software changes are inevitable. When it is necessary to change the user interface of a software product; however, the costs of traversing the typical iterative software development model must be repeated for user and technical documentation, marketing and promotional materials, tutorials, and training materials. For these reasons, it is desirable and cost effective to make a durable software interface. That is, to design the interface right in the beginning so it can be used effectively for the initial functionality and extended easily to incorporate new functionality. In our work, we noticed that some products were generating many more user interface changes than others. Many reasons contributed to the need for changes: some reasons were positive (e.g., users were so excited about the product and its potential applications that they wanted to make it fit new requirements) and some were negative (E.g., buttons or commands didn't work as expected). In this paper, the distinction between functionality that doesn't work as expected versus functionality that doesn't work correctly is deliberate and important. The quality of software implementation is an important topic that should be addressed separately from the idea of software design quality. The focus of this examination is on design issues. We hypothesized that a thorough software usability effort would improve software durability, and that therefore, it might he possible to justify such an effort based on cost effectiveness. This study was not scientific. Rather, it was relatively informal and relied heavily on anecdotal data. This was necessary because we wanted to select two products that were comparable in terms of complexity, user population, and time since product introduction. Since the products were selected quite some time after they were introduced. historical data to support this study had not been formally collected. However, we are confident that a more formal study would find similar results.