Following is my view on how it should be done. I'm young, fresh out of school, just taking part to a debate, and the answer to this question may not be found in my participation, only elements of it if anything.
The fairy tale of post...
I see two disciplines in post, editing (and sound design) and mixing. Ideally, the mixer will be given as dry a material as possible, implying that the editor/designer deals with problematic sounds cues like the one depicted in the first post. There are solutions to the problems we may find - in this particular case I'd think of multiband compression - but this belongs to the preparation of the material, i.e. the editorial. Some sounds however do come to the final mix with their own reverb, and that's part of the designer's job, but if it doesn't work for you, you should send it back for re-editing (that's what all the books I read say). Unless you think you know of a solution, I'd say "leave it to when you have extra time to spend on mixing."
When mixing, the further the sound source, the wetter it'll get. If I don't use reverb, I find it sticks out and just brings it really close to you. Well, I say "I find" but that's the definition of a dry sound: if it's dry it's in your head, then the wetter the further...
On the very few mixes I've done, I like to start cutting up my space into three slices; close, medium-range and far, and still the further the wetter, so in Pro Tools 3 aux channels with the reverb on it receiving spot effects tracks "by slice". I sort of grind it coarse first and then refine.
I hope this serves the discussion and I'm not off-topic.
And they live happily ever after...