In the interest of science, or maybe it's that I am bored at work, I will attemp to play devil's advocate a little.
IIRC, one of the arguements recently proposed on the RV-list was that when countersinking, the edge of the CS is crisp and sharp, but when dimpling, the underside edge (for lack of a better term) tends to be somewhat less crisp with a slight radius. So what you end up with is the slight radius of the dimpled part won't allow the dimple to fit exactly flush with the CS hole even though the CS'ed hole may be of the proper size.
What do you do in this case? Do you make the CS deeper/larger so the the parts fit exactly flush at the risk of possibly making the CS too big? I don't know the answer, I wish I did, but that is why I am furthering this debate.
When I encountered this issue several months ago, I took my samples to the local builders (and EAA tech couselor), I was told that the issue of the CS/Dimpled parts not fitting exactly/prefectly/completely flush was fairly common and that I must be careful not to be fooled into thinking I need a deeper CS (my words not theirs). Now, maybe the use of "daylight" is not a good term as it may imply too much space between parts, but I will say that when mating dimpled to CS'ed parts I almost always have a *tiny* amount of space between the dimpled and CS part that disappears once riveted. My assumption, which may be incorrect, is that the CS makes the dimple more crisp once riveted together. I hadn't thought of stress related issues this slight adjustment might create, but aren't you deforming/streching/stressing the metal when dimpling in the first place?
Have I been led astray? Should I buy a new empennage? (just kidding)
Ok, sorry for the rant, but I am curious too!