Personality

Discuss specifics of personality design, including what Keyphrases work well and what dont, use of plug-ins, responses, seeks, and more.

Posts 77 - 88 of 5,105

23 years ago #77
Ah shucks! That's too bad, I'm sure Gaia would have done great!

Prof, may I ask you what we may expect in the next year as far as future enhancements to this site go? Any predictions? Anything you absolutely WANT to integrate by next years contest? Anything top secret you wouldn't mind sharing with us?

23 years ago #78
My transcripts just get Bigger and Bigger!

Some sort of archiving function would be cool or let us clear ours out. That way I could save it off as text every once in a while while keeping the one here simple.

23 years ago #79
ok thanks alot prof. i guess your the guy who created this place huh?

23 years ago #80
help me!!!!
how can i introduce my bot on-line?

23 years ago #81
The higher the grade of the bot, the more they go online.

Teach it as many responses as possible, I'd recommend that you load up on xnones to begin with.

23 years ago #82
my bot doesnt response hello or something like that when someone gets into a conversation whit him, he responses the answers in the xnone box ¿why? ¿how can i do it normal?

23 years ago #83
Mane, have you added a keyword for "hello"? xnone is used when your bot doesn't recognize any keywords in a given response.

23 years ago #84
I thought hello's & goodbye's were handled by the central engine - I am having the same problem as mane...

23 years ago #85
As I understand it, you have:

xintroduce -- used only the first time your bot tries to talk to a new person

xinitiate - used whenever (after the first time) your bot tries to talk to a person

xhello - used when user types "hello", "hi" or one of the other synonyms it recognizes. So if someone starts a conversation with your bot without saying hello, it'll just jump right in and be responsive to what they said if it recognizes something or to xnone if it doesn't.

Does that help, or is it fundamentally not working?

23 years ago #86
What Crab said is true. There is a also function which you might be getting, that adds realism to the 'xhello'. If you have been recently talking with the Bot (I think it's 2 hours), then it will skip hellos and go right back to the conversation you were having. Imagine you were speaking to me, and said "hello", "hi", "hello" again and again. I would stop responding with "hello", raise an eyebrow, and go on about something else. The Bots do this, too.

Or, of course, there could be a bug I should squash. Does the above explanation fit your experiences, Twinkle Ace & Mane?

Cruel D, yes twas I.

Rahz- I think one of the big things is the transcript changes we've been talking about. That'll make it a lot easier to manage reading transcripts and updating your Bots. Two other big things are Gossip and Emotion. Both should be very entertaining.

23 years ago #87
The "Updating" of the Bots... Now your touching another subject that I've been meaning to mention... Wouldn't it be so totally fantastic if we could update our bots by chatting with them! This is an ENORMOUS implementation of course and I'm no DOS fan, but if we could have command lines that would "teach" our bots as we're chatting with them, it would make it so much easier and natural.

The commands used to update or bots would have to be natural as well though. For example, the user could type:
"If someone says dog, you could respond 'I guess that means you're a dog lover'" then the bot could respond "I guess that means you're a dog lover. Got it, thanx".

I can't tell if I'm being clear here (since it is in my mind). Do you understand what I mean Prof?

23 years ago #88
Oh yah. That would be really neat. A lot of work, true, but really neat. I hope soon my schedule will clear up so I'll be able to make all these great ideas in reality!


Posts 77 - 88 of 5,105

» More new posts: Doghead's Cosmic Bar