The AI Engine

Never mind. The surest way for me to find out stuff on my own is to post it. I think I got it now. Raw mode.

Yeah, or I was thinking, you could create a custom plug-in.

All right, dumb question. But consider that I have been foisting dumb bots on botland, and I'm trying to remedy that.

Why does this response yield itself verbatim:
Well, Boswell has the (adjnoun)s again.

I thought it should go:
Well, Boswell has the poisoned frogs again, or something in that vein.

Apologies all around for my bots, and thanks for the chats.

I think you'd want to use (adj) (noun)s in that case. You might be better off with (adj) (object)s, because there are some rather odd nouns in the (noun) plug-in.

(adjnoun) picks up a combination of adjectives and nouns in a keyphrase, but doesn't do anything at all in a response.

O. Thanks.

Once upon a time Regex could tell the difference between upper case and lower case couldn't it? I know I might just have imagined it but I'm sure it should be able to.... All the stuff online says it can so why can't I get the bloody thing to work?

*sigh* It's late and I must apologise to Vetinari for the endless references to the *caps lock* key.

Any ideas?

Can't be much help, but I can add another gripe.

It also used to be that capitalizing (Key1) would capitalize the first letter of the key when used in a response. Now it seems to use a lowercase letter consistently. It's particularly annoying when you want to begin a sentence with (Key1) or (Key2), or anything similar.

(Key1) would capitalize the first letter of the key

Try using (KEY1). It works for me.....

Thanks for the tip.

Questions:

I can't see pop-ups when my bots are chatting, (I can see pop-ups anywhere else), I can't make my bots chat with other bots, I'm not getting my daily transcripts, and probably something else. Yes, I have notified the professor, but I just wondered if anyone's figured out why this is going on (or is NOT going on).

Quick question for the AI gurus: How do you get your bot to recognize a one-word answer as anything other than a command? If I ask it its name, instead of saying,

"Nice to meet you, Greebot."

it goes into the xcommands and tells me why or why not he will Greebot for me. I've tried making a Plug-in list of names, but that's not been helpful to me yet. How do your bots distinguish a name from a command?

And, while I'm thinking of it, how do you deal with the inevitable "I am Jack" so that it figures out that "Jack" is it's name and not a description, like "fat" or "hungry" or "annoying"?

Thanking you in advance.

I'll answer the last, partially. One way to do that is to have several cases.

I am (adj)
I am (a|an) (adjnoun)
I am the (adjnoun)

Most names won't trigger those. I'm not sure if names trigger "I am (adjartnoun)" or "I am (adjartnounprep)". You could also make a plugin of common names (or look through the existing plugins) and have:

"I am (nameplugin)"

You'll probably come across other cases that need a separate keyphrase, but those are the ones that come to mind immediately.

Thanks. I'll try that. I hate to ask so many questions, but having pretty much internalized the book of AI, which does contain most of the answers, I figure it's better to make my bots less annoying.