The old school way was basically try a reaction, see if it matches a group of chemicals. Try another, see what you get. You can only close in on a reasonably ac

The old school way was basically try a reaction, see if it matches a group of chemicals. Try another, see what you get. You can only close in on a reasonably accurate guess if what you're working on is fairly simple. Otherwise you'll be stuck in a local maxima of has this group, doesn't have that group, without the ability to get any better resolution.