Popular CS Books, or lack thereof

In an IM conversation, Ravi Kumar of Yahoo! Research points out that there aren't any "popular" computer science books -- there isn't a Timothy Gowers or E.T. Bell or Brian Greene or Richard Dawkins. (When I invited Ravi to write a guest blog article on this subject, he promptly backed out, though -- another reluctant writer within our ranks :-)

I wonder why. Oxford Press' "Very Short Introduction" series has nearly 30 titles in the "Science" section, including such topics as "Philosophy of Science" and "Jung" and "Consciousness", but none on the science of computing. There's one on cryptography, but I am quite certain no two scientists who consider themselves cryptographers will agree on what ought to be in a book with that title :-). Absolutely nothing on the "Internet" or "Algorithms" or even "Artificial Intelligence", a favorite among the non-scientific audience. John Battelle's "Search" has made it to under-200 on the Amazon sales rank, but that's more of a meta-CS book (historical anecdotes, biographical stories, business wisdom, etc.) than an actual CS book.

Even the mighty Christos Papadimitriou's "novel about computation" Turing doesn't seem to have made a dent (if it's any indication, his "Computational Complexity" book sells more on Amazon than does Turing). Noam Nisan has a new book "The Elements of Computing Systems: Building a Modern Computer from First Principles" (with Shimon Schocken), which I haven't seen yet (reviews welcome as comments!). From the blurb on Amazon, this seems to be more of a course textbook; don't expect it to become an NYT #1 seller in the nonfiction category.

Why isn't there a popular book talking about how some early computer scientists like Turing and von Neumann had tremendous foresight and got some basic "design decisions" right (like universal machine, stored program computer, etc.)? If there's a CS book for the masses, what, gentle reader, would you like to see in it?