A great example is the benchmarking fiasco of 1999. Microsoft paid for the best benchmarking statistics that money could buy. At first, it tried to deny that it had paid for them. The truth came out rather quickly.

Then it tried to say that the tests were fair and independently run, but there were many objections to this point of view. The Mindcraft benchmarks were done with special hardware and special configurations that the average public user would not have access to, and on hardware that was in the most basic sense, extreme. And they were performed in a lab somewhere out of community sight, with experts who knew Windows NT and oddly enough, no Linux experts to tweak the Linux configurations at all.

All to prove that "Windows NT was faster than Linux."

What a mistake. To be honest, initially there were things about Mindcraft that were painfully right, and exposed weaknesses in some of the subsystems that needed tuning under Linux. The long-term effects were extremely positive for Linux, as we will see.

The benchmark was examined, Linux was scrutinized and overall, the product was improved, as were various other open source systems, such as SAMBA and Apache. In the mean time, Windows 2000 is an enormous product that appears to be less efficient than Windows NT on the same hardware. The most recent benchmarks of Linux versus NT show it smashing the competition in ways that Microsoft would rather not publicize.

Odd, the story hasn't received the same amount of media attention that the first benchmarks created.

There was a lot of honesty in the Linux community surrounding the first benchmarks, and everyone pretty much faced the real facts. Very little obfuscation of the end results occurred. The payoff is that today the equation is very much tipped in favor of Open Source. I believe wholeheartedly that you're just not hearing about it because it doesn't make that good a story for those people who have big advertising budgets paid by enormous corporate interests.

Next: The Golden Rule »