There's been a lot of "fail" talk about this article and it's parts. I personally found it interesting but the test procedure seemed a little tedious. If I were running the tests, as an engineer, my test procedure would look like this and save the authors a lot of time.
1st off, get 2 processors from each brand, a popular low power consumption and a popular high. For right now I would go with Intel's e2___ and the q6600. These will be used in the final tests of the best, and worst (passable!) coolers.
The setup would be a temperature adjustable hot plate, a copper block contraption that holds it's own thermalcouple, a thermalcouple datalogger (can be had/designed for cheap), and a pwm fan controller (any motherboard with speedfan program control of the fan header would work). Now power up everything and start testing on your own terms. A logical test would be 45w, 65w, 100w, 200w. These are typical thermal design limits specified by Intel and you can do more or less, I don't really care so long as there is some logic behind it.
Using this equipment you could easily, and speedily, test each cooler to their limits. Then, having all the data you can get an idea of which is the best and which is the worst. Test these on the 4 processors and report the temps stock speeds, overclocked, idle, and full load. A min of 16 cpu tests would be performed. Probably more because you would have failing coolers like the stock box e6300 that wouldn't be worthy of a q6600 oc'd.
Now you would have some hard core facts to satisfy even the minority (or majority, I don't know) of readers that like to come to their own conclusions. Tom's becomes the unbiased real deal for those of us smart enough to look at the numbers and still, with fewer cpu tests, less of a challenge for the non-technical to understand what they're looking at. All this could be accomplished with a bit more setup time but considerably less testing time. With a rig like I explained, you could test every cooler out there in a matter of minutes as apposed to what I can only imagine was a grueling task of changing the coolers on the motherboard and having to wait while things started then got stable.
Just my 2 cents.
Completion of the cpu charts is all i want. get them all in there on the same test rig. If its consistently tested across all coolers, i just want to see the final temp result chart. forget the installation rating, etc. a synthetic heat block at 133 watts would be fine too, just give us something thats complete so we can say "this cooler is better than that cooler" and end the threads about randomness.