I'm not sure about how many people know how a poll is supposed to be conducted, but it is much closer to how IW did the second one. IMHO, it should have been with a much larger portion (3,000 readers instead of 300), and with less bias. Rachel has mentioned that letting all of MS vote would allow the vote to get skewed. The way you conduct a survey is to limit those that can skew the voting to an off amount.
As for the "OS/2 zealots", I don't blame you for getting upset at the way Sandy phrased things; but I think that with the limited use of OS/2, the results *were* skewed. And perhaps in the second one, the results were skewed the other way. Asking people to call up or log in and answer questions is *not* a scientifically valid method, as should be well known by any journalist, and using too small a pool of representatives is invalid in sampling surveys. So we need a larger pool, consisting of both "IT pros" and non-"IT pros", and have it done that way. The main problem comes in that everyone wants to participate, but at the same time, everyone wants a scientifically valid poll. You have to pick one or the other; it's nearly impossible to have both.
Finally, this should not be redone until next year, if at all. I'd like to see another, but, considering all the flaming that goes on either way......
For questions or comments, email the Forums Editor