Clearly the original web poll was flawed. Precautions should have been taken to prevent repeat voting. Personally I would have stuck to last year's method, which restricted voting to registered subscribers. I'd think it would be more reliable, and it never stopped OS/2 from winning before.
The new poll does have meaning, but it certainly doesn't qualify as a "Reader's Choice Award", at least not in the way most people understand the term. I would expect a Reader's Choice award to be self-selecting, since it's basically a popularity contest. If some products inspire users to vote out of proportion to marketshare, then that simply indicates stronger feelings on behalf of those products. Come to think of it, it's a good way to distinguish between products people use because they're forced to, and products people use because they want to.
Even if a phone poll was necessary, I think it was a bad idea to eliminate a large segment of the readership from consideration. That only makes the results seem less trustworthy.
Also, I think Sandy should just stand clear of the argument. This little "controversy" may have helped stir up some interest in the magazine, but it has done some lasting damage to its reputation, even among readers who wouldn't want OS/2 anywhere near their networks.
For questions or comments, email the Forums Editor