RSS AI Zone
Hi was wondering for those who do much AIML (looks like mostly Steve but all responses welcome!) - some suggestions for organizing AIML and debugging it?
Right now, I design for a “happy path” and then just through trial and error start to come up with a way to see what happens if I go off a pre-planned path.
I have seen some illustrations of AIML paths and was wondering if those were used to debug or design?
This is more of a higher-level “best practices” question since I find I am writing against a happy path, maybe using wild cards to help capture some minor devitations.
I also try to have the path go through a loop but keep the response interesting or “cascade” through different type of questions depending on whether a predicate gets filled out or not by using <conditional>
So no specifics, but maybe some design principles.
