Newcomers

Ah, I didn't even think of augmented complements. I didn't realise it did that. Darn it. They need a check box for automatic emotion adjustment or something.

It is true, you could write your own with AI-script. However, as you said, since you can only use strings and no integers, to get the same range of the current system would take 11 responses for every keyphrase. Ugh. Of course it could always be made simpler, but when bots get complex, I can't imagine that would be much fun.

Also, I notice that if I have duplicate sentences, but they adjust emotion differently dependent on the current emotion (e.g. [-1:1] adds 2, while [2:3] adds 1), it doesn't even bother importing the second sentence. This requires minor adjustments, which I can imagine would bring about the occasional "No response" problem if programming your own system and forget about it.

EDIT: Just came across another problem. My AI-script isn't actually changing the emotion. Am I doing something wrong? I have tried:
<?PF emotion: -1; ?>
and
<?PF emo: -1; ?>
My phrases are something like this before importing:
Yadda yadda [2:3] <?PF emotion: -1; ?>

NOTE: Using debug, it DOES read the AI-script fine, but it just doesn't seem to change it.

EDIT 2: What... the... hell... If I import from a txt, and change all the "emotion" to "emo", or visa versa in the browser, then the ones I changed start working?! Otherwise it is as if they are ignored?!

I've come up with a little trick to overcome the deletion of duplicate keyphrase problem. I set up a local plug in in the keyphrase, ie:

are you (really|absolutely|positively| ) sure?

and then mix up the order inside the plug in to get different keyphrases that work exactly the same way:

are you (really|absolutely|positively| ) sure?
are you (really|positively|absolutely| ) sure?
are you (absolutely|really|positively| ) sure?
are you (absolutely|positively|really| ) sure?
are you (positively|really|absolutely| ) sure?
are you (positively|absolutely|really| ) sure?

It appears the engine only looks for exact matches, and none of these are exact matches.

im new and dumb to this stuff

@Sam, ah that is smart. Alright, looks like will have to be doing that. Thank you.

Anyone have any answers to the emotion problem I listed above? Since my bot is heavily dependant on fine emotional control, I cannot imagine going through every single response and flip-flopping the emotion name after every import. -.-" Especially since they should be working (and doing the same thing) in the first place!

@DarthRevan, haha, I feel ya. I have worked with this site for a small while, but nothing too majestic. Now that I am trying something more advanced, it is a whole new ballgame of strange and (as I am finding out) bug-ridden.

@Sam Little
Since that | ) at the end of each of your in-line plug-ins matches nothing, you don't even have to think of actual words for the plug in. Just use (a|b|c|...| ) with as many letters as you need to overcome the duplicate parsing routine in the AI Engine. Even if someone types in "are you b sure?" by mistake, the bot would still answer reasonably.

That will save a lot of wear and tear on my Thesaurus. That "positively" doesn't really fit in the sentence
"Are you positively sure?" but changing the base keyphrase to

are you (really|absolutely|d| ) sure?

would work just as well.
Thanks

If anyone needs help I will try my best to give advice.

i need help with amber fur. i am a total noob at this.

Deanmachine722, just talk to the bots, similar as you'd talk to a human person. You'll see rather soon what a bot can do and what not.

echon: "why are all the bots phrases all jumbled when talking"
-----
You surely have already had a look at the FAQ page (http://www.personalityforge.com/faq.php#14)?

Please specify in what way your bots are talking funny.

@Dean, I am no expert myself, but I will let you know what helps me out.

The book of AI: http://www.personalityforge.com/book.php
Try reading through the entire thing to help get a general idea of how things work. Generally, lots of really cool stuff is possible if you are creative. It is hard to be creative if you don't know how the system works, which it is why to have at least a basic overview.

Also, when first starting, get some "general phrases" in to help match unknowns. Then when you want more specific things, add the more specific phrases with a higher priority. For example, at first, something like "i * like" could trigger a vague response whenever the user states something they like, later overshadowed by things like "i * like popcorn" which can have a more specific response.

This may seem obvious, I don't know, but it helps to have "filler" when a bot is still pretty limited and under heavy development. Reading transcripts helps you to figure out what specifics should be added first.

Anyway, that is my two cents.

To add to what Thagen said, the phrase
"i * like" will match
'I really like peanuts' and
'I like peanuts' --but it will also match
'I don't like peanuts' so you need a companion keyphrase like
"i * do not * like" which would match
'I don't like peanuts' and
'I really don't like peanuts' and
'I don't really like peanuts'

etc. etc.