Well, technically, you are paying a small price in performance for the restriction that the smaller barbs create. I can't say for sure since I haven't done any experiments along those lines, but I doubt it's that big of a factor.
The reason 1/2" ID lines work better than 3/8" ID lines is that the ratio between cross sectional area and perimeter is much higher. This means that proportionally less water is interacting with the surface of the tube, which means less friction for a given quantity of water. Less friction means less head loss, which means greater flow for a given pump. When the water runs into a short region with a smaller cross section, like one of your barbs, for the most part it just speeds up. Some turbulence is created as well, which leads to some head loss, but shouldn't be a deal-breaker. For waterblocks with impingement mechanisms, like an RBX or WW, a larger restriction occurs anyway farther downstream, so the smaller barb is irrelevant. For high-flow blocks, like a Maze 4, the higher flow velocity can actually improve performance by causing a thinner thermal boundary layer (for a good example of this kind of difference, see the difference in performance between a Koolance CPU300-H10 and a CPU300-H13). I'm not as certain about radiators, but I think that they are similar to impingement blocks in that they introduce a large amount of resistance already, so the barbs aren't a big deal.
That's my take on things, based on what I know of fluid dynamics. I'd be really interested if anyone has done any experiments along these lines. Another possible lead would be to look up minor losses in a fluid dynamics textbook or reference. My books are in storage now, or else I'd definitely have the figures for you.
Any other mechanical engineers out there?
