On 09/14/2011 12:48 PM, Steven Schveighoffer wrote:
But what makes your statistics more "valid" than Tiobe's? As Walter said, there are multiple ways to measure popularity, and every one is guaranteed to dissatisfy some people.
I demonstrated the fundamental problem with their methodology. They use only the first 100 results to get statistics for a distribution that will have the vast majority of the false positives past the initial results. It's one thing to say "number of web pages about X" is an arguable measure of popularity, it's quite another to use a profoundly flawed way to get that number.
The statistics I quoted were varied, simple, and easy to verify. Come up with alternative evidence instead of just using a broad brush to make all measures equal. Clearly some languages are more popular than others.
