Skip to Main Content
Software standards. Now there's a subject that brooks no "loyal opposition," right? Standards are material provided by some software god, biblically significant, subject to no doubt? In recent years I've come to question all that.In 2006, my colleague Johann Rost wrote a guest Loyal Opposition column on the standard for requirements documents. He explained that at the conceptual level he understood the standard quite well, but at the implementation level he found it impossible to follow. I offered him sympathy, hosted his column here, and sort of forgot about it. Time passed. Another colleague, Barbara Kitchenham, was beginning to struggle with certain software standards. One standard gave "inappropriate advice for measuring software engineering processes." Another standard was "not suitable for measuring the design quality of a software product." Echoing Johann's concerns, Barbara commented that "experienced designers will be able to construct a number of different interpretations of the standard, implying that [it] is not a standard at all." Barbara pointed me at yet another critic of software standards, Magne Jorgensen. He noted that one standard for software quality "requires quality measurement and at the same time admits that there are no (universally) accepted quality measures".