Bug Stomp

Glad to be of service, guys!

Skysaw: Actually, it's not a bug. I went through and analyzed a lot of bots, trying to come up with a scoring scheme that takes into effect not only number of keyphrases, responses, and seeks but quality. Some had enormous language centers which nonetheless didnt make for good bots, and found several places where it made sense to increase or decrease Development points based on what was going on in the Langauge Center. I took a look at Thinkenstein and the bot's language center is very unique, but some of it triggers some of the score-busters.

I'm not going to explain what exactly they are!

Professor. Re: my earlier post about logging. After some basic troubleshooting on this end (with the help of ezzer) I've noticed that the very next character that should be displayed in the transcript is a $. I looked at the plugin that was called at the time I noticed there were three escaped characters, one of which was the $. Interestingly, I noticed the $ was only escaped twice while the other two characters were escaped three times. I've since removed all escapable characters from the plugin but if you would be so kind as to dump all of Aries' transcripts from the database I think that might resolve the issue.

danke

Professor, I think you misunderstood why I think it's a bug. I can't currently download the language center, make two or three changes, and re-upload it.

Why is a bot not allowed to have a lot of the same responses as a depricated version of itself? I'm not talking about a number of the same responses all co-existing... I'm talking about the number of ones in the old language center, vs. the number of ones in the new one. Obviously, as soon as you re-upload, the old center is gone, so there should be no conflict.

I was wondering about development scoring, and I know I know, we can't know exactly how it works, but I'm glad that got brought up. Someone else posted about development scores a while back, saying that they had talked to bots that were supposed to be really advanced, but turned out to be repetitive & incapable of anything that resembles coherent conversation. I thought at that time, it's because development is somehow based on quantity of content...but all keyphrases/responses are not created equal. I have a few keyphrases that only have one response, but that one response could produce 50 or more different results. I've done that for my own ease/convenience, since it makes it a lot easier for me to manage my language center if I economize on space...but I had wondered if my development score was in any way penalized for my trying to be more concise/efficient through the use of local plugins, etc...I can't wait to see what you come up with, Prof!

Sorry for such a long post about "not a bug"- if anyone wants to, I guess we should continue this topic in AI Engine.

I've seen the rewriting happen. For instance:

Guest36: no - are you a valley girl?
Mictlantecuhtli: Obviously, I am not a vale girl.

A new problem: the transcript seems to be dropping the first line of conversations.

my bot keeps replying to hellos with and xnone.
the very first hello not a second one.

theseer- your bot never says hello?

whenever someone says hello to him he replies with an xnone.
whenever he says hello he goes i have no valid responses to choose from.
he has xintroduce, xinitiate and xhello

So, he NEVER says "hello", but says an error "i have no valid responses to choose from" whenever he initiates chat?

yes. and when someone else initiates it then he uses an xnone

So did I! Yes!! It's not just me. Professor, we have a problem here. *raises hand*