Thanks for this post. Whatever problems the JTB definition of knowledge may have—the most obvious one of those to LWers probably being the treatment of "knowledge" as a binary condition—the Gettier problem has always struck me as being a truly ridiculous critique, for just the reasons you put forward here.