Indeed, that is a very justified question to ask. Also very much connected to XP’s rule “Always do the simplest thing that could possibly work.”

And for your last question, for most cases I would say no. In some special cases it might be valuable (e.g. when desiging something like the .NET Framework). But in a general I think ultra general designs tend to be harder to understand.

The only one thing that I would like to add, as a comment, to your text is the following idea, or suggestion, (or mantra, if you will ).

Everytime one feels tempted to extend that simple required feature that she is implementing, this person should ask herself a question:

“what is the trade-off between spending time on implementing a ultra general design and successfully implementing the required feature with a simple, yet less general, design?”

In other words, do those hypothetical extensions that would justify the general design really matter now?

Yet another question, does a design need to be ultra general to be a very good design? I do not think so.