Silver bullets

Do-it-alls: Tuomo Valkonen (ion et al), Charles Moore (Forth), Bob Armstrong (CoSY), Ian Piumarta (fonc), Niall Douglas (Tornado et al), the Smalltalk community (possibly the only entire community with this mentality?), Niklaus Wirth (the Oberon system).

What is the allure of neat, compact, universal answers (sometimes called, especially in extreme cases, “silver bullets”)? Is it a longing for the simple old days when one really could know everything that was known? Is it not in any case necessary for some people to build their universal systems, because only by over-emphasis of new principles can one elucidate them sufficiently to convince others of their utility in even appopriate, restricted domains, at least, when they are organising principles rather than theories of specific interactions? For example, without the Smalltalk people would we have Java? Similarly, Charles Moore, even if we don’t agree with him, poses hard questions about what is really useful, and what is just “customary cushions” (isn’t, but should be, a quotation from somewhere). Niall Douglas reminds us that there are clear benefits to a relatively modest throwing out and reworking of our existing systems; CoSY (and Emacs) that an apparently over-specialised system can serve well as an operating environment for at least some users (rather less specialised and rather more users in Emacs’s case). Tuomo Valkonen is like Niall Douglas but more hard-line, but he too reminds us of some uncomfortable decisions (e.g. the vast amount of effort put into anti-aliasing rather than into improving displays).

Late 2008–mid 2009


Last updated 2009/02/01