Minor issue: If there's a format stored in Calibre that Calibre doesn't know how to handle (DejaVu in this instance) the plugin throws an error and aborts processing.
Possible optimization: Abort searching through the book once a certain percentage/amount of text has been searched. This would help speed up the search for 95% of the books.
Building it into Calibre would be fantastic but since this is the major roadblock to me finishing my catalog, I'm going to continue to push it. <g> No worries, I'm looking at how to do all of my suggestions myself as possible improvements. I work in C#/C++ professionally, just not Python/Calibre. I'll just have to buckle down and do some (gasp!) reading.
As for jokes... ha! Trust me, I'm far from serious. One reason I don't participate in projects is because my joking attitude tends to grate on the more serious folks who tend to inhabit the programmer's world.