Also there is no way for author to know that his optimizations wouldn't work without trying them. We are not even sure that his analysis of why they work is correct (though it sounds reasonable). There is also the possibility that author could use the new found information to fix his optimizations so they are better.