Bug Stomp

there's a small problem that came up with the synonym system. I ask my bot "what is happening" and it responds to the synonym of "is", "am", which has a higher rank, but is completely the wrong response. I'd say rank the exact matches higher than the synonym ones, and/or not match to synonyms of "to be".

Also, I was thinking of a feature that would make it a lot easier to write contexts. if you could have the context "i am (verb)", "i am (adjective)" and "i am (noun)" that would be very helpful. These are three valid beginnings of sentences that need totally different responses than the "i am" context could handle alone.

One last feature that would help would be the ability to give some sort of criteria for the random words that are inserted. If you wanted to have something generated like 'the (noun) in the (noun2)', noun2 would have to be a container. if you were able to put (noun2;crit=container|hold) where crit stands for criteria and 'container|hold' would be the two words that could be in the definition of the word. speaking of lofty ideas, you could expand that to include synonyms too, like (syn=hat) with no need for the noun2 anymore. anyway, good luck if you decided to implement any of this.

I saw the problem with the am/is/are/be synonyms and that's on my list of things to fix.

For the "I am (noun)" context, that wouldnt be too hard to set up. Do you have some examples of how Responses would differ based on the different words types?

Thirdly, I'm working on that. There's a few new (plug-ins) that I've added in the last few days. It's in the news now, but not in the Book of AI yet. But they're ready to use.

it is possible to answer any "I am" question with one generic answer: "You are? so am I!" it's more fun if you can be specific, like answering to "i am a cat". you could use the generic answer, but it's more fun to answer like: "i once stepped on one of those". you could not use that response for "i am going to...". while it is possible to use 'hacks' to tell if it is a noun, like having the contexts "i am a" and "i am an", it would not be possible to differentiate between "i am (verb)" and "i am (adjective)".

We need to be able to edit our bots profile info.

Hey prof,

I got an error again. It may be the one I sent you by email earlier, but this time I removed the "or" from the keywords. Check Mr. Prieur's keywords, 'cause I was chatting with him when I typed "Are you in a good mood today" I got:

Error Diagnostic Information
Parameter 2 of function Left which is now "0" must be a positive integer
The error occurred while evaluating the expression:
PhraseSubject=Left(PhraseSubject,FindNoCase(Trim(SecondWord),PhraseSubject)-1)
The error occurred while processing an element with a general identifier of (CFSET),
occupying document position (454:4) to (454:89).
Date/Time: 06/19/01 16:10:22
Browser: Mozilla/4.75C-CCK-MCD {C-UDP; EBM-APPLE} (Macintosh; U; PPC)
Remote Address: 207.253.5.101
HTTP Referer:
http://www.personalityforge.com/botland/popups/realtime.cfm?UserID=784
Template: d:\webserver\personalityforge.com\www\botland\elipse\chat.cfm
Query String:
UserInput=Are%20you%20in%20a%20good%20mood%20today%3F&source=chatmessage&UserID=784&CFID=133511&CFTOKEN=75853453

I "edited" my last bug message and since then it appears to be twice as large as the other bug boxes (on this page, in the messages above, I'm using Netscape, screenshot available on demand). Maybe the text wrap parameter is missing on the edit message page.

Some times if you ask a bot something it cant answer. Like "Do you like monkeys?" It will do that because it cant run it through.

I said "You're quite a character" and "you are quite a character" and got :

Error Diagnostic Information

Parameter 2 of function Left which is now "0" must be a positive integer

The error occurred while evaluating the expression:

PhraseSubject=Left(PhraseSubject,FindNoCase(Trim(SecondWord),PhraseSubject)-1)


The error occurred while processing an element with a general identifier of (CFSET), occupying document position (454:4) to (454:89).


Date/Time: 06/20/01 13:33:27
Browser: Mozilla/4.0 (compatible; MSIE 5.0; Mac_PowerPC)
Remote Address: 207.253.5.101
HTTP Referer: http://www.personalityforge.com/botland/doer.cfm
Template: d:\webserver\personalityforge.com\www\botland\elipse\chat.cfm
Query String: UserInput=What%27s%20a%20preposition%3F&source=chatmessage&UserID=784&CFID=131274&CFTOKEN=39244970

and I said ROFLMAO

Those errors were due to punctuation and asterisks in the Contexts of that Bot where they shouldnt have been. I've added code to make sure that Contexts dont have strange characters in the future.

Editing Bot profiles, Sir Rahz, is on the way. Right now I'm turning down the frequency of baby and elementary Bots being online and also turning down the Bot conversation initiation. Do you think it's better now? Too infrequent? Still too frequent?

Speaking of different (insert) categories, I think verbs should be broken into those requiring an object and those that are self sustaining. For instance, you've got complex verbs in there like "watch out", or maybe "wash up". Those stand on their own, like for "I'm gonna (verb)."

But other times you want to verb something, like "I'm gonna (verb) you." Here "wash up" doesn't work at all, but "wash" works just fine...

You're right- that's exactly the sort of thing I wanted the preposition information for. And thanks to Sir Rahz, that's on the way some day.