Hans:
Aside from the OCR algorithm that I told you about, do you think that similar algorithms could detect 2D or 3D shapes?
I could see another application in data storage and compression if the first, last and number of letters are used as only variables saved/stored. Three bytes to recognize/save words.
Of course, additional bytes would be needed at times eg to differentiate grease from grouse. Thus the maximun number of bytes needed would be four. Do you know any word that requires five bytes?
Your comments please.