Stemming certainly must be fitted in at some point. It would be part of above-mentioned streamlining while gathering regexes from real-life, published sentences.

Thanks for those links. I still have to study on several fronts, I'm afraid...