Once upon a time, a computer science PhD student (I: H_I) went on IRC, where he learned what a bot was: a program which was also connected and could interact with users. But what poor interactions! Only "physical" ones (kick, ban, topic changes, ...). The student thought: "Could I make a bot talk?"
He finds a bot's source code (a vladbot2.1h), modifies it a bit, and makes it react to a particular sentence: when someone says "hello", the bot responds "hello". "A bit limitated," the student says. "When someone says hello to anyone else, the bot reacts. This should be modified." As the bot has a name (Achille), it can react when its name is in the sentence (or when it is concerned). And so it is: the bot says "hello" only when someone says "Hello Achille", or "Hello folks".
The student makes Achille reacting to some other stimuli ("bye", "how are you", and so on). To make it more realistic, the bot gives differents answers to the same stimulus ("Bye", "It was nice to meet you", etc.).
As the student discovers IRC, he realizes that some humans are antipathetic with Achille. So the bot should not answer the same way to all IRC users. It may now like or dislike users according to what they already said to it (and also to what they say and do on IRC channels where the bot is).
But Achille, can't converse in a really intelligent manner. Alone, it can't sustain a conversation with a human. However, it is adapted to IRC: on a channel where many humans talk, it only reacts when it knows answers.
As the student pass its PhD on an Artificial Intelligence architecture, he thinks that it could be applied to a more intelligent bot than Achille. It would be named ECTOR (an acronym for the French: Entité Cybernétique Totalement Obligée de Réfléchir). However, it is likely that it will not be as impressive as Achille (in the beginning): it has to learn many things before!
The next pages show how I (H_I, the ex-PhD student, ie the doctor) intend to give it such an intelligence.