rob.rho.org.uk

Dan Bricklin (of VisiCalc fame) has posted When The Long Tail Wags The Dog, an interesting essay on the appeal of general purpose applications (e.g., spreadsheets) compared to specialised ones for popular cases (e.g., tax calculation). His argument is that, given that you can only learn a finite number of tools, you're going to pick at least some general purpose ones, as those are more likely to fulfill your specific needs when said needs diverge from the popular. My view is that he doesn't properly consider hard problems, where solving the problem, in your way or any other, is too difficult or costly for the user, but nevertheless he makes a good point.

A good book on the subject is A Small Matter of Programming, by Bonnie Nardi. The copy I borrowed proved invaluable during my PhD, and I've just picked up a copy to reread (under a tenner on Amazon Marketplace - thanks, Bournemouth University). I'd thoroughly recommend it to anyone interested in end-user programming.