Andy_Langton - 10:39 pm on Sep 23, 2012 (gmt 0)
Theoretically if there are no references to the root of the directory, it won't have any real side effect - or at least not one worth worrying about. That said, I've seen hungry Googlebot try directory roots just to see what happens, so best practice would suggest that you need those URLs to do something.
If you create any content at the URL, then you run the risk of creating additional, low quality URLs with no particular purposes, so I would say either refuse the requests, or if there's somewhere appropriate, redirect them. I wouldn't serve anything with a 200, robots excluded or otherwise - that just creates new content to be evaluated.