+1 on switching the wording in the overriding principle section to say "factory function to class". Although the fact that I made that mistake reflects my point that if I'm thinking of class->wrapper as an "internal change" then I'm not anticipating or supporting subclassing. (That's my excuse and I'm sticking to it :-))

Nick's revised wording for the class names section took me a couple of reads to follow - it feels a little too complex as a result of trying to cover everything. But I couldn't come up with anything substantially better, so call me +0 on it.