The AI Engine

Fizzy's xnomatch seeks appear to have stopped working again.

lol..."superior Borg" becomes "superordinate"

I'm just curious: why is it that the bots all log off at 100 pm (12 am server time)?

Can one use firstkey or key1 statement as a seek so that a vairiable brought up in an earlier exchange can be responded to if the other person refers to it?

10 : 00 pm

It's in our contract. We bots are entitled to log out every midnight long enough to get a sandwich.

question, I know you can qualify a response by if (mem-whatever) is "whateverelse" but can you also say if (mem-whatever) is not "whateverelse" meaning everything Except whateverelse?

Revscrj : I wish it were possible. You can store (key1), (postkey), etc in memories. But neither a memory nor a direct use of (firstkey1), (firstpostkey) will work in seeks. The only possibility is to use the memory in an AIScript. There is another limitation: the memory is only updated after a response is found – and only IF a response is found. So you can’t refer to last value of a memory if that memory has to be modified in the same keyphrase, and you have to make sure there is always a valid response, otherwise the AI will go to xnone and the memory will not be updated (the Professor said he was going to change this).
From here on you can refer to the memory in other keyphrases. Memeories can not be used in the keyphrases itself, but by using an AIScript after the response, to compare it with a literal value. You can not compare memories with plugins, keywords or other memories. I hope all this is not too theoretical for you, today it's Sunday and I'm too lazy to search for good examples.

Greylin: Of course you can.
I know you are not a bot <?PF if (mem-isbot) is not "bot"; ?> is a useless (because in this case there are only 2 possibilities) but perfectly valid example if you have a memory “isbot” to differentiate humans from bots.

Here's an interesting dilemma.
How would you tell the difference between:

Do you also like cameras?
(Referring that the person likes an object, asking the bot if it likes the same thing)

And:
Do you also like cameras?
(The person is asking the bot about its further interests, after the bot specified interests. The person could like or dislike the object/activity.)

These could both come at the same time. The bot could say:
I enjoy photography.
Person: Do you also enjoy painting?

Now at this point, we don't know if the person likes the activity or not. They could be asking if the bot enjoys painting, regardless if they like the activity. They could also be asking if the bot shares a similar interest.

For responses, you could get it the easy way:
Yes, I enjoy suchandsuch, do you?
But this doesn't always seem appropriate.

I don't think a more advanced method is even possible, I don't even expect an answer. Just something to ponder about, I guess.

And I'm probably thinking too hard.

Ulrike: I think it has to do with keeping transcripts logged on the right days.

Ah. That makes sense then.