Hello,
The "Mystery-Interview Demo Brain" posted here was inspired by the work and experience of Dr. Thomas Whalen, who won second place in 1995 at the Loebner Prize / Turing Test competition.
Dr. Whalen's concept was to create a chatterbot with a specific persona and backstory, and then engage the judge into helping the chatterbot solve its "problem."
The backstory was this:
Joe the Janitor is being fired from his job cleaning a college laboratory. He was accused by a researcher of stealing a small transistor radio that looked like a soda can. Joe the Janitor is innocent, but Joe isn't smart enough to realize that he's being framed. Why was he framed? Joe witnessed the researcher using an electric shaver to shave hair off the bodies of mice. The shaving was evidently part of a scheme to falsify data on a hair-growth medicine. By getting rid of Joe, the crooked researcher gets rid of a witness to the fraud. If only Joe could recognize the motive of the researcher, he could tell his story to higher management and get his job back!
Therefore, the "goal" of the user is to get the facts out of Joe, make Joe understand, and get Joe to agree to take his story to upper management.
Here are the transcripts of what happened at the Loebner/Turing event:
http://hps.elte.hu/~kampis/Loebner/Whalen1995.html Dr. Whalen observed that the judges absolutely didn't want to "suspend disbelief" and let Joe tell his story. The more that Joe tried to tell his story, the more impatient the judges became, and the more they tried to "fight" Joe. The transcripts are fascinating (to see Dr. Whalen's excellent work), but also funny and frustrating, to observe how the judges just refuse to "get it"!
At any rate, I wanted to show that Ultra Hal and the VBS language were extremely well suited to such a "mystery interview." The mystery I created was much simpler, to make the programming shorter and simpler, and useful as a tutorial for other programmers.
Enjoy, and have a great day!
Sincerely,
Don
