Dharmesh Shah has written two good articles considering the view that software should be as simple as possible. The articles try to clarify what sorts of simplicity are good and what kinds aren't. (The articles are a response to the 37signals book Getting Real, which is big on the simple-as-possible thing).
The first article argues that software, being focused for users, has to meet their needs, and simply providing the minimal set of features can bring them up short.
The second article is a bit of an elaboration on that. It argues that while software should be made simpler by being 'opinionated' and deciding 'the little details so your customers don't have to' it shouldn't go too far and become 'stubborn' -- stubborn software is not only opinionated but it is inflexible and doesn't provide users the ability to configure the software to their preferences.
This stuff is related to what I've written on the nature of 'simplicity' and 'complexity':
(1) Sketching on Simplicity As Qualitative Perceptual Concept
(2) Notes on What Qualitative Perceptual Concepts Are
(3) Complexity as Qualitative Perceptual Concept
(5) Quick Drafting on Tradeoffs Between Local and Global Simplicity/Complexity
(6) Factors in Tool Complexity/Simplicity: Viewing Value as Additive
(7) Factors in Tool Complexity/Simplicity: Vertical- and Horizontal- Features
(8) Initial- and Standing- Simplicity/Complexity
I also noted down this this
comment by Linus Torvalds on this matter of oversimplification.