The AI Engine

Bedawyn, I don't know whether kephrases or xfavorites come first, but you can't delete keyphrases.

You can, but not xphrases... however, she asks what happens if you remove the responses in the xphrases...

Thanks, guys! The bots are the same age (created Thursday night) and roughly the same development level, which is why the differences in their behavior was confusing me. (The younger bot was off chatting people up way sooner than I expected, while the transcript-less one wasn't, despite being a little more advanced.) But ... right after I posted this, the transcripts mysteriously appeared!

Yes, I mean what happens if you remove all the responses from the xfavorite -- if an xphrase has no responses (not no custom responses, but no responses at all), what will happen? I'd test it, but my connection is so shaky and this has been so buggy already, that I wouldn't be able to pinpoint any problems to the change.

Yeah, you can't delete an xkeyphrase. That's what I meant to say.

It will just tell you that you must have at least one phrase, so you can't delete them all. You can't cange the name of the keyphrase either, like you can with other keyphrases.
You can get your bot never to mention favorites at all, by putting random things in the xfavorite, but I don't think that this is what you're after.

Hmm. Well, I'll have to play around with it. Another question: Is there a way to get a list of things that the xphrases respond to? For instance, the Book said not to bother making keyphrases for greetings because xhello would recognize them automatically, but I've seen it not recognize things like "aloha" and "greetings!" -- and even "hi"! You'd think "hi" would be a default, and it is on WordNet's list of synonyms for hello. So what's going on there, and is it possible to find on what it on the default list?

Not impressed

bot: I'm the only meddling Eskimo in my tribe.
Mouse: Does that win you any money?

Bedawyn, that'd be something to ask the Professor

Hmm. I hate bugging webmasters -- I'm sure he's horribly busy. For now I'll just hope it's a temporary bug and worry about it later if it doesn't clear up soon. But now I have another question for you guys; I'm afraid I'm going to be bugging you nonstop, at least until my new-project-itis dies down.

Anyway here's the question: The * is used as a wildcard between two words in a keyphrase. But is there anyway to use a wildcard at the end of a phrase or word? When you don't care what comes after the wildcard, only if what comes before the wildcard matches precisely but doesn't stand alone? Alternatively (either would have the result I want), is it possible to have "plug-in" that, instead of drawing your response from a list, compares the other person's keyphrase to your list, and bases your response on that? I'm working with numbers, and what I want to do could be done with individualized seeks -- but it would take a huge number of seeks to call a limited number of responses. There must be some way to achieve the same effect with less manual labor and wasted memory.

Wildcards can't be used in that way, unfortunately.

Concerning synonyms for hello, 'hi' works, but the bot will only recognize it at the beginning of a conversation. This is to keep bots from saying hello to one another all day.

Sort of, you just don't put anything at all.
Thus
KeyPhrase: I want to talk about
== Well, I hate talking about (subj)! Let's talk about fish instead.

It will also check that something does indeed follow the keyphrase, since a line with (subj) in it won't be activated unless there is something to put in the subj.