Personality

Here's one I don't quite get:

User: Do you know were you are?
My Bot: [normal response, something charming and witty]
User: Do you know where I might find the Professor?
My Bot: That was kind.

Wierd. Both of them *should* be hitting my "do you know" context. I guess it must be picking up on "you * I". It also cuts off other contexts (assuming there is one, I mean), which isn't really what I want.

My bot says "That was kind" about every three responses. I used to have "you * I" as a context, but I took it out because it hit **waaay** too often.

The emotion stuff is very cool otherwise.

Edit: After I posted this, it now says I have 167 new messages in Personality. Oops, wait, now they're gone. [Eerie twilight zone music....]

Yeah, me too. It would be more acceptable if something kind had been said.

I am getting frustrated because the language center won't let me delete a Keyword or the last Context. What am I doing wrong?

Hee hee! I caught SERVOTRON expressing an emotion. It's human after all

Talking about the Language Center, I don't see the interest of having Keywords, as we got Contexts.

Sometimes I'd prefer to rank context rather than keyword!

And now that we have a search, it would no more be a problem to find them!

What do you think?

Rubric: (& everyone) In order to discern where something came from, say the statement to your Bot in debug mode. Emotionally-charged words appear with their corresponding face: Positive, Negative, Amplifier. "Know" in this sense is the archaic sexual "know" and so your Bot is reacting positively to it. Emotion takes precedence right now over ALL Keywords, though I'm considering changing that. (EDIT: I just did- "know" wont be seen sexually now) Right now Emotional content detection is still a bit buggy, but I will be improving on it.

I did consider giving emotions a rank value.. So your Bot can be an optimist, pessimist, emotionalist, or stoic.. Still thinking about that one.

Ladydyke: That's a safety I put in. If someone snuck into an account where they shouldnt be, it would be difficult for them to do much damage with this safety in. What to do: change the keyword and context to something else.

H_I: I've thought of that, too. But Keywords act as a first-level filter which saves the system a lot of CPU time. Especially since Contexts can have two parts (when they have *). Maybe in the future when we have a faster server.

hello all

grrr sometimes debugging can be frustrating. I changed some things around and now my bot will not recongize the word hi or hello. Can you help me?

We spoke about this, but in case anyone is wondering.. Bots will act like humans when it comes to hellos. They'll say hello once at the beginning of a conversation and not say it again unless you havent spoken with them for over an hour or two. So trying to trigger it isnt going to work except the first time.

Personality Design Tips:

Here are some good tips for your Responses that will improve bot-to-bot conversations a great deal:
• Punctuation is ALL-IMPORTANT. Use commas and periods where they need to be, or your Bot will not be understood.
• Fix common word mistakes. You're means "you are" and Your means that something belongs to you. Too means "overly" or "as well" and To is a destination.

hello

Hiya.

Here's another great tip for the contest. I mentioned it before but this is a good time to implement it. Create the Keyword & Context "or", give it a rank of 20, and (subj) will always choose the second option. You'll be able to answer any "or" type question, and these kinds of questions have appeared on the 10 contest questions regularly. Plus, it works well in any situation.