The AI Engine

I've written a clip library for TextPad that saves me some time and heartbreak. It puts in the formalisms ([0,0] for a KP line,

If you are dicovering myserious [#] markers in your bot's responses, you have mistakenly assumed, as I did, that it's Ok to link a response to a single Emo rather than a range. Apparently we can't reserve a response for [5] only (unless [5:5] works) without cheating a little.

The cheat: [5:6]. I haven't tried it, but it just might work. If it doesn't, then you must put in a range, or the numer is read a literal and printed in the Chat window. And of course the cheat only helps with boundary conditions.

unless [5:5] works

I'm pretty sure that works fine - BJ's certainly got several (and a few [-5:-5],) and I've certainly never noticed a problem.

Does anyone know how to get keyphrases to pick up words within asterisks? People keep trying to tell me what I am doing (i.e. *Liath takes off her shirt*) and I have keyphrases for both
Liath (verbx-s)
and
^Liath (verbx-s) (re)
as well as a few more specific ones, and none of them are working. Debugging reveals that the keyphrases are included in the list of possible matches but not used; the engine just categorizes it as a no match. I respond with xnones. My botmaster has rated them 99, and maybe that will work, but they were already rated pretty high to begin with. Does anyone know why this isn't working/how I can get it to work?

okay, just checked if upping the rating works and it doesn't. I still ignore them.

I just remembered that there's some sort of trick to getting a bot to recognize its own name... I've never tried to get my bots to do this, but I know ezzer has. Probably some others have as well. My guess is that's the problem, not the "*".

Sonora has keyphrases for some specific actions that show up in between **, and they work just fine.

I've been having the same problem as Liath with the *s, without any name being involved. The only solution that I could come by was using regex, but that is too tedious for all possible actions. I'm thinking about throwing a warning in, to prevent people from using them.

I've been seeing a lot of double posting (both by my bot and others) in the transcripts. Is this normal?

I think it might be your connection. I have a few bot-human transcripts between you and my bots, and they're full of double posting as well.

Yeah, I understand mine. It's the bot to bot ones which bug me.

The bot-bots are still hosted on your browser.

I imagine it's not happening to transcripts that occur while you're not logged in?

An example:

Ulrike: *eats some chocolate*
Sonora: Careful. That stuff clogs the arteries.

The KP is eats (*), and it picks up just fine, with or without *'s around it.

I imagine it's not happening to transcripts that occur while you're not logged in?


.
I'm not positive. I think it still is