Sorry Neuro but focal length does not affect DoF it's yet another myth which has been perpetuated, nor does subject distance.
http://www.luminous-landscape.com/tutorials/dof2.shtml
Sorry, Flake, but it's not a myth - it's fact. In the article you link, the author is changing the both focal length
AND subject distance simultaneously (decreasing both focal length and subject distance), in order to keep subject size equivalent. This was, in fact, the point of my statement that aperture is the primary determinant,
in practice, as in the example of the head shot.
Pick the DoF calculator of your choice, e.g.
DOFMaster, change only the focal length, and DoF changes. So yes, focal length affects DoF - but only if you keep other factors constant. To reiterate, in practice we
care about those other factors. If you just go by the numbers, you can get more OOF blur with the 24-105/4 than the 24-70/2.8. For example, I can frame a head shot with my 70-200/2.8 IS II at 70mm f/2.8, and then zoom in to 105mm and stop down to f/4 - and I'll get
more OOF blur at 105mm f/4. But I won't have a head shot anymore, I'll have an eyes-and-nose shot, and I don't want that. So I'll have to walk backwards to frame the whole face at 105mm, and have less OOF blur (but if I stop down to f/2.8, I'll have the same OOF blur as at 70mm f/2.8, with increasing focal length and increasing distance effectively canceling each other out, again reducing it to a matter of aperture).
Magnification affects DoF which is why Macro has such a thin DoF & aperture, circles of confusion are not really anything to do with the sensor size, again that comes down to magnification, you need more to fill the frame on a FF camera.
In optics, a circle of confusion is an optical spot caused by a cone of light rays from a lens not coming to a perfect focus when imaging a point source. It is also known as disk of confusion, circle of indistinctness, blur circle, or blur spot.
Yes, I can read wikipedia, too, and even copy and paste from it. Understanding what you read or copy/paste is not the same thing. If you go one line up from where you copy/pasted on wikipedia, you'll see: "For the closely related topic in microscopy, see Point spread function," - PSFs are something I work with routinely in an experimental setting. Circles of confusion are, arguably, the most important determinant of the DoF, in the strictest sense. By definition, 'depth of field' is that part of an image which is acceptably sharp, as opposed to that which is blurred. CoC is what determines what is sharp and what is not, and thereby determines what is within the DoF and what is not. What makes CoC confusing (pun intended) is that real CoC depends not only on the way in which the image is captured, but also the way in which the image is viewed. For example, an image that appears acceptably sharp when viewed at 600 pixels wide on an LCD display may be a blurry mess when printed at 20x30", even if the resolution supports a large print.
However, when discussing CoC in terms of cameras, we have no control over viewing/printing - therefore, 'standardized' values are used based on an assumed print size (A5, if memory serves). Those standard values are applied to different sensor sizes, and APS-C has a smaller CoC than FF. The take home is that the standard values for sensor formats mean smaller sensors will have deeper DoF for the same subject framing (and MF/view cameras have larger CoCs, meaning they have shallower DoF for the same framing compared to FF/35mm. So yes, CoC values are relevant for both sensor (or film) sizes and for determining DoF.
Does Macro have a thin aperture? That's a new on on me. I didn't even know apertures had measurable thickness. But yes, DoF is thinner with macro distances. In fact, most of the approximations that underlie typical optical formulae do not hold at near 1:1 magnification or greater (and when you get significantly higher magnifications, i.e. microscopy - there's a whole new set of formulae).