As many have said, Tim's comments are from the book perspective (which he is perhaps most qualified to comment on).

Exploring the reasons for the ratio of book sales among different languages (and indeed, among older and newer titles, and later editions, within a language) can inform us of the dynamics of language adoption, use, development, and (shudder) disfavor and abandonment. It is a complex relationship.

To me the more important question is:

Why did the original questioner think the question was interesting?

While I'm waiting for an answer to that, I'll take this opportunity to ask a few questions on the dynamics of languages in relation to book sales.

Is Perl, being perhaps in middle age as far as languages go, sufficiently simple, yet powerful and well documented, with both the interpreter and documentation freely available on the Net, limiting its own book sales?

Are Java, C/C++, and Visual Basic so cumbersome, arcane, or difficult to use that they require a small library self-help tomes to compensate?

Is Ruby so easy to use that only a few books will ever be sold?

Is Perl still relevant? Book sales are only a small piece of that puzzle.