It would have been useful if it had "English words starting with" and "English words ending with". Or a simple way of filtering sets and using them as subsets.