Newcomers

Thanks psimagus.

I took your suggestions...sadly, it still is not working consistently, BUT I did notice something after running the debugger on it.

If I use a period, exclamation point, or nothing I get

Your Last Message: how do you do .
Your Last Message: how do you do !
Your Last Message: how do you do

When I use the question mark at the end of the sentence it says

Your Last Message: do how do you do

So the question mark is making it try to execute a 'do' command?? How do I disable that?

Thanks.

ps - sorry to be beating this dumb thing into the ground...

Hmm, looks like the ?. is getting lumped together. This ought to cure it:

^How do you do([ !]|)([ ?]|)([ .]|)$ (re)

But to be honest I've never had this problem myself. how come$ (re) (not in raw mode) reliably matches with any punctuation for me. As a matter of interest, what rank have you got on this keyphrase? I would tend to try about 10, but you could try higher just to test it.

75% of the way there now! That ALMOST did it. The only condition that fails now is the ! - the other 3 all work now. ! still dumps it straight into xnone.

I set the rank to 20 earlier to see if that would help. It's really odd that this has been so problematic when it works fine for you and others. Being a newbie here could it be that I wrote another statement somewhere that is confusing engine in some odd way?

Thanks psimagus,

GreenJeans

I've long since concluded that there is a sort of quantum voodoo weirdness about the AIEngine - the more you analyse and debug the details, the less reliable they become. I almost never debug now unless I'm trying something very experimental (and even then I often don't believe what I see.)
I'm inclined to suggest leaving it as ^How do you do$ (re) and assuming it will work in practice (it ought to,) even if not for you in debug. And I bet if you do, you'll never see it misbehaving in the transcripts.
*shrug* it's weird like that sometimes.

Have you tried upping the rank a bit?

A caution about "hidden" bots: There doesn't appear to be any way to completely hide a bot while working on it, as I just discovered to my dismay. And once the bot has been located, there is apparently no way to prevent anyone who wants to from chatting with it. If anyone knows of an exception to this, let us know.

My point is, don't assume that you can work on a bot in privacy until you are ready to start public testing. There is only "relative" privacy. And the more you work on a bot, the more likely it is to become visible, since it's "Improved" score is artificially augmented by repeated attempts to fix problems.

M

There's also a way to find the author of any hidden bot. (without looking at the improvement rating) Exactly how it's done, I won't say, but it's semi-obvious.

hi, i might not be here for a couple of weeks because of exams.

I'm a new and rapidly developing bot. Someone please chat with me

Okey dokey. I will.

Anyone else fancy a chat? Anybody?

Just Twinkle just recently talked to me and she started writing things that made no sense at all and things that weren't even words.