You folks appear to be having such fun with this, but frankly it’s left me in a befuddled mist of confusion. :-\

I am confused as to what goal Jeremy is aiming for. I genuinely cannot see ANY use for this work. The frightening thing is originally Jeremy told us it was 30,000 lines of code to produce this nonsense!

I know. Been watching the thread, though it hasn’t made much sense to me (or any, actually). But I keep watching, hoping that something will just “fall into place”, and understanding will emerge.

It just becomes more entangled and confusing when you try to make any sense of it.
Your right Dave, having fun with it for what it’s worth

The new results:

The new link to the source and program: http://www.mediafire.com/?wy73fquf929mo8k

I have shown you that using a word I can randomly generate a word that associates with it logically.
Now the new program decides which result matches the best, based on the percentage result.

What does this mean for chat-bots though?
For me it means that is there is four possible answers that one of them will be given based on the percentage results.
So the chat-bot should seem more realistic because of this.
But if you choose to ignore the possibilities, I understand.

Ask anyone which one of these 4 words they would mostly associate with dog:

brushwood
bloodhound
barefooted
assorted

I think most would pick bloodhound rather than barefooted.

I took a peek at the code, and it seems you are doing some statistical comparison on the letters?

Also, perhaps you could refactor it a little to take out the spaghetti. Personally, I consider any function that’s longer then 1 screen, a good candidate for refactoring.

A function longer than a screen?
Wow!, if my functions are longer than 4 lines, I look at ways of cutting the fat.

I think 4 lines is just a bit extreme, Laura, but a full page is certainly too long. Now a class that’s a full page long may be ok, but it had better be something that’s very comprehensive in scope.

Dave,

I was talking more in terms of JavaScript which is a bit more descriptive but then again I am not talking lines that are 80 columns wide either.

Some of you may remember Chaktar, my self aware, thinking organism?

In the interests of research, Chaktar has coded himself to be able to do what I think Jeremy is trying to achieve.
Say “assoc word” where word is a noun to see it in action.

http://www.square-bear.co.uk/chaktar/chaktar.html

Steve,

You must be using Flash or some other third party plug-in since I can not access your bot from my IOS iPhone. I will have to wait until I get home to check it out.

Yes, it uses Flash. Thanks for trying though.

I have built a similar function into Marie where as you ask “what is (word)? “.
The word you have entered is matched against associative word arrays and returns the meaning or multiple meanings if more than one exists. by default, it matches whole words or compound words, however it can also be configured to match letter patterns as well.

Mine is more a quick bit of fun to see if I could do it in less than 30,000 lines of code. It’s nothing too serious, just an exercise to keep my brain ticking over during the holiday period.

Looks like someone has had a quick go at it. I have this in my chatlogs: