It would be impossible get a "hard count" of VFP developers/users. Therefore, statistical sampling is the only way to get this information. But you are correct, the number could be low or even worse, high. However, any sampling taken will have an associated margin of error, so the number will be "statistically valid".



>
>There is no need for snottiness, Craig. Actually I took a couple of statistics classes in college. My point here is not that you can't make estimates based on sample data. My point is that these are not to be confused with hard data. Anyone stating that there are X number of FoxPro users is making an estimate which may or may not be valid. That was my only point. Clearer now?
>
>Mike