Literate modeling and neatness
That's what we were all told growing up, I imagine, and it is a pleasure to work in a neat environment.
Then why do (I suspect) so many of us who model have directories with files named model-2008-10-15.model, model-2008-10-15A.model, model-2008-10-16.model, etc.?
It's because we were also told that modeling is an iterative process, which it very much is (one student in one of my classes thought "Iterate" must be my name). If we iterate on models (writers do the same), and if we want to keep older versions around just in case we need to go back to one of them, we end up with heaps of models to keep track of.
Literate (and text-based) modeling can help. Instead of filling directories with files, I now have one model file per model, and I let bazaar, my revision control system, track the history of that file. If I want to see what it looked like last week (or last month), I simply check out the old version or compare it to the current version. Bazaar tracks comments by revision, too, so it's easy to find the model I created to address a particular issue (assuming I noted that issue in the comment for that revision).
Why do you care? Have you ever worked with someone who couldn't find an aspect of the work you needed to see? Have you ever "improved" something you're working on, only to find a prior version was better? Perhaps literate modeling, text-based modeling, and revision control are in your future!