Amazon.com lets people access their "Properties" via SOAP, and several publishers now like to mine that data for all sorts of reasons. Amazon.com could have a small market information business if they wanted it---maybe they do, but people seem to be doing it themselves.
This puts me in a new situation when I do this sort of work. Each person (or organization) that wants to access Amazon.com (which they call Properties) needs to have their own access token. I have an access token, and I can access data but I cannot redistribute it. That seems fair enough. If the person I work for has their own token, and they contract me to do some work for them, I can use their token and give them the data they asked me to get for them. They cannot redistribute it either. Nothing changes in the technology, and only a little bit changes in what actually happens, but it is completely different legally.
As I was collecting sales ranks the other day, I wondered how various Perl books were doing. I grabbed all of the data for those books, processed it, and output a nice chart. Then I realized I probably should not post it to my journal (since that would be redistribution of their data), at least not without Amazon's approval. I would like to add this sort of data to Perl at a Glance so we can see how Perl is doing compared to other technologies.
Instead of posting the data, here are some interesting conclusions.
What does this mean? Write Learning Perl PHP Cookbook if you want to have a high sales rank, but leave off the Win32 bits.