The complexity of satisfaction
We're in the final throes of getting Mathematica 8 out.
It's a massive task. As with each release, it's a far more complex piece of engineering than its predecessor. But so too is our build process, to handle that. Which complexity wins? Is it a more or less fraught process each release? Don't ask me in the next week(!) but I think the release process satisfaction index is improving.
Of course it's not just for Mathematica builds that there are rapidly escalating complexity or expectation competitions. It's a common facet of modern life. Cameras, computer games, smartphones and many other gadgets run this gauntlet. Beating expectations is key to individual satisfaction, though the absolute level of achievement is important to mass adoption of the genre in question.
Key to my satisfaction with products is depth of design: as I discover more, do I like it more or less? Or to put it another way, is the product's satisfaction derivative positive with time? Because I tend persevere with products, I get to the point of sampling this eventual rather than just initial satisfaction; many don't; and to figure it out upfront with complex, modern products is pretty hard.