UX nerd

This, together with a copycat review of Apple’s latest gadget (whatever that is at the moment you’re reading this) is the kind of post that I never though I would write here: “10 worst usability mistakes”, “10 best website designs”, etc. But I would like to think that I’m vindicated by originality and Paul Graham… These are my 4 bits of design wisdom that may come in handy some day, you can judge for yourself :)

1. A thorough gathering of requirements should include the following questions: What should the system do? But also… What could it do? What must it do? What must it not do? Nice piece of advice borrowed from my dark (hard-core engineering) past, who would have thought it’d turn out to be so useful in design.
2. Any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes (Goodhart’s law). What does this mean? That relationships between two variables in historical data (as for example between speed and user satisfaction) will not necessarily hold after you start acting on one of them to influence the other one (like in: increasing speed past a certain limit will no longer produce an effect in user satisfaction).
3. Conway’s law ([...] organizations which design systems [...] are constrained to produce designs which are copies of the communication structures of these organizations) is not meant to be humorous! Evidence courtesy of Microsoft Research.
4. Good design needs not be perfect. At one time, Westinghouse began to work on heat pumps that could provide heating and cooling, and believed that he might be able to extract enough power in the process for the system to run itself. Lord Kelvin told him that he would be violating the laws of thermodynamics. Westinghouse replied that might be the case, but it made no difference. If he couldn’t build a perpetual-motion machine, he would still have a heat pump system that he could patent and sell. The rest is history.