As an IT professional, I have a different view on the matter.
1) We all know that the vast majority of security incidents occur from within an organization
2) If you're looking at data, and 1 sample is far beyond the standard deviation, you toss out that 1 data point.
3) If you don't follow manufacturers requirements, you do not get support, and YMMV with that equipment.
4) You use unique ID numbers to associate different data entries

What does this have to do with Landis?
1) The time stamps on the datafiles used in the case against him have differing creation dates and modification dates. This was after the initial red flag went off
2) The SOP for the T/E test specified that a sample outside the standard deviation of 4 was unacceptable. Landis sample B was 14:1... Had they been following their own SOP, they would have tossed this result as erroneous
3) The computer used to interface with the lab equipment ran OS/2, but the manufacturer of the lab equipment had only certified the machine against Windows.
4) The samples used in the case against Landis had a different ID number than Landis.

This whole case just illustrated the level of incompetence at the lab. To this day, I wholeheartedly believe that Floyd Landis won the 2006 TDF. Not because I'm naive, but because the case against him was so flawed. Had this been held in a US Criminal court, it would have been thrown out. But the WADA doesn't follow a typical court model, nor does the Court of Arbitration for Sport.

In regards to the stage in question. Landis was one of the first to race with a power meter, not just in training, but in the race. He had realtime numbers on his wattage output. He was able to determine that his effort was within his limits based off his training. Also, he used a huge quantity of water, but only drank a small portion of it (this is shown on the footage). He actively worked to keep his core temperature down by pouring something like 14 water bottles over himself.

They can't be done for cheaper than Flash, though, which is the only thing the bean counters that pay me care about.

To achieve similar levels of professionalism and quality in a training product (what I do for a living) using non-Flash tools would cost too much money and take too much time because it requires too much technical talent.

Flash looks good, accomplishes the goal in less time, and for less money. Doesn't mean it's technically better, only that it does what it's supposed to for less time and money.

V day would be the day you can get away with that kinda stuff =P

So until a realistic Sun/Earth computer model exists, a true "global warming experiment" can't be run.

The purpose of science is to create and confirm/falsify that model; it can't use that model as a basis for experiments. Nobody is ever going to confirm or falsify a global warming hypothesis using this approach, because if their simulation doesn't get the result they want, they'll just say the model wasn't realistic enough.

With weather science, we're at a point somewhere in between Copernicus and Kepler. We have a basic idea that appears consistent with the observations, so we're likely (but maybe not) on the right track, but we have lots of nagging details that keep us from having an accurate enough model to really propose a theory. I think it might end up being so complex that we never (even a thousand years from now) quite nail it down with enough precision that we can say, "The temperature will be n.i degrees at this time next month."

What we do have, though, are parts of the model. We can do an experiment in a flask and see how the gasses in the flask can influence its ability to absorb/reflect/etc energy. If someone thinks this isn't going to a factor in the ultimate (possibly unattainable) model for weather, I'd love to hear why.

Was Copernicus a scientist? I can't give a satisfying answer to that one, but I'm going to go out on a limb and say he was doing good work. ;)