Abstract: Software construction, yet to establish its role in the world, inhabits an invidious position between mathematics, engineering and craft — losing out by the comparison to each. I draw a distinction between different levels at which the term ‘craft’ could be applied and list four key virtues desirable for crafted items at a higher level. The current products of computer scientists or software engineers are deficient in these four virtues, but not necessarily so. I propose a shift in the value system underlying our work—recognising that the function of material is to bring communities into contact. I will argue that software is material to the extent it promotes an ecology of artefacts which cooperate harmoniously in the lives of their owners. Rather than the workmanship of certainty, we should promote the workmanship of risk, turn from correctness to truthfulness, and consider time horizons of generations rather than at most a couple of years.
PPIG 2016 - 27th Annual Workshop
Building Software is Not [yet] a Craft