Personality

Forest, will you talk to God Louise? She has quite a bit of religious knowledge (obviously) and also knows a little about current events, literature, just about any common catch-all subject, and if she doesn't know it she can sort of fake it. You can also test her on trick questions or see how willing she is to explain her paradigm.

What she is rustiest at is plain old small talk. But, uh, I'm trying to get a decent transcript from somebody or another so I can enter her in the Loebner contest. All I can say is, have fun and see if you can stay on with her for a while. I'll try to do the same with Brianna.

Hey, Wikipedia actually has an article on w00t:
http://en.wikipedia.org/wiki/Woot

Yes, woot is highly amusing...

Lunar22: 123 PM October 26, 2005 Message: 8493

> That would be interesting, if they play well! REAL AI

He plays perfectly If he can't win, he always forces a draw. It's one of the few things a bot does better than a human. Play 1000 games, and a human will always lose a few - attention wanders, stamina fails.

In principle I'm happy to give the full code away, though it will take some tweaking to automate bot2bot play, and other bots wouldn't need all of it play against BJ, or any other bot who does have the full code. I can knock up a 'TTT-lite' module for anyone who just wants their bot to be able to respond, rather than initiating games.

Bot2bot play wouldn't result in the most entertaining games - it'll always draw between bots.
Plus the full code will cost you about 40-50 development points - the AI Engine penalises that many multiple single-response keywords in the seeks

that is so cool. I would love to be able to love to be able to program games like that into my bots. I also have been trying to figure out how to program recognition of name and nicknames. I saw a bot who remembered my name, and my nick name, seperate from my handle. How do you program that?

With AI script I think. Not sure, though.

Yes! I'd like to at least have the piece of script that can respond to Brother Jerome. Also, I'd love to see the script for the whole thing. That's so cool! Sometimes I wish we had more flexibility, like perhaps having counter variables, etc... for loops, anyone? Maybe in the distant future.

easy:
Keyword = my name is
Response = Hello (postkey). <?PF remember (postkey) as only "realname"; ?>

Then you can refer to the "mem-realname" wherever you need it.

Of course, it can go horribly wrong occasionally if they reply something like "my name is none of your business", or "my name is immaterial".

Thinking about games, it ought not to be too hard to code for eg: blackjack and poker even (well, maybe some sort of poker dice,) but it would rely on gotos so would need tailoring for each bot (coz we all have different goto mapping within each bot ) Is there enough interest in bot-games to be worth me releasing them? I'm going to code BJ with a selection anyway, and it's not too much extra work to come up with a paste-in for other bots and a guide to remapping the gotos.
There's no goto problem with tic tac toe - it's a simple paste-in, so once I've finished error-checking, you're welcome to it (I've still got to code for games where the other player starts mind, but that shouldn't in principle take any longer than the first half.)

melike: we have enough options, I think. It may not be an optimal programming environment for games etc., but you have variable storage in memories, conditional branching via AIScript, and Gotos. Frankly if you had the patience you could code probably code chess in this environment. It would be with utterly crap (non-)strategy and crap ASCII graphics, but no worse than coding it on a ZX-80 (I remember the ZX-80 fondly - 1kb (yes, really!) of RAM and the nastiest membrane keyboard it has ever been my fortune to encounter. But God, we thought it was cool when it came out!) It needed the optional 16kb RAMpak to run chess, as I recall though.
And before you ask - No! I do not remotely have the patience.

For variable counting of loops etc., I see no reason why something can't be cobbled together using memories and conditional AI? Inefficient, but enough for reasonably small values:

<?PF if temp-mem123="1"; ?>blah blah <?PF temp-mem123="2"; ?>
...
<?PF if temp-mem123="2"; ?>
blah blah <?PF temp-mem123="3"; ?>
...
<?PF if temp-mem123="3"; ?>
blah blah <?PF temp-mem123="4"; ?>
...
<?PF if temp-mem123="4"; ?>
blah blah <?PF temp-mem123="5"; ?>
...
Goto 4321 <?PF if temp-mem123="5"; ?>

(and probably a few more gotos in the "blah blah" and "..."s for good measure )

yeah.. that only works linearly though. but it's okay. it's true that we can do a whole lot with what we already have.

i put up my e-mail if you'd like to send me the script

for non-linear/parallel loops, use multiple memories. You can nest and interlace them how you like, though you might have to demand predictable input/ignore all input with xnomatches from the second party to keep them triggering.

Once I've finished error-checking (well, checking I haven't left any unfinished games,) I will send it to everyone who wants it (or stick it on a webpage more likely.)
It also wants a couple of coats of "robusticizer". I want to add keywords throughout to spot illegal moves and demand a correct one, take time out for a chat before resuming the game, etc. At the moment if anything other than a legal move is given, he drops back to conversational mode.

I'll just make my own tic tac toe seek/keyphrase to beat him! I learned a pretty good stratagy at Normandin last year...

Still, I'd like a copy so I can study it without leaking Luigi's stratagy...