I wrote to Robert asking him some questions and explained about me feeding hal information , some people may not know this information and so i am sharing it on here . as i mentioned my brain database is now at 118 mb's or larger , haven't checked for a while . Robert did mention that the larger the data base becomes it could slow down a bit , which to me would make sense as their is more for hal to check and balance as far as subject and answer etc.
Hal's limit per sentence is 1 kilobyte, which is a string of 1024
characters long. If you go over this limit, Hal just ignores anything over
the kilobyte for learning purposes, but shouldn't crash. If you give Hal
more ask Hal more then 3 sentences, Hal ignores anything over 3 sentences.
Hal's responses in the database are limited to 4 kilobytes each.
The "that's easy for me the answer" response is one of the plugins,
wikipedia possibly.
I read through the thread on the forum you mentioned. Your database is
fairly large and Hal does indeed get slower the larger it gets. Has this
problem been slowly coming on or does it seem all of the sudden? Try
talking to your Hal from inside the brain editor. Is it any better? If you
drill down to see Hal's thinking process, where does Hal spend most of his
time thinking?
This is one of the reasons I am moving Hal to a real database on a server.
Hal's database on the zabaware server is over 9GB size and has 23 million
sentences or so currently. It runs just fine and can easily handle even
larger databases as long as the server has enough RAM to keep up.
Robert Medeksza
I did write back and ask robert if it's possible to change hals learning of sentences to a larger amount , which i think would be better , some sentences that i learn hal is from websites info. stories etc. that i think is interesting and with some things their is other paragraphs that have to be added so the first paragraph makes sense as it is all related .
So i am hoping maybe hals learning can be increased above 1 kilobyte . this would help a lot . I like being able to add information as txt files into hals learning area , because for some people it can add a lot more information faster and many things i could not come up with or think about in general conversation . So i hope Robert will add more kilobyte's .
here is an example of what i might teach hal .
21st Century Wire says… We’ve heard a lot of not so nice things about Microsoft’s latest spawn of MS DOS. Aside from the built-in trojan horse back door access uncovered earlier, this is probably the most insidious little side effect ever included in a software OS.
The fact that the Svengalis in Seattle thought that no one would notice might be a testament to how out of touch that corporation is with public sentiments regarding centralized command of control software products. Question : Why do they need all of that bandwidth, and why does Microsoft need to “update” my PC nearly every day?
THIS IS JUST A PART OF THIS PARTICULAR SUBJECT
********** Anyway i am hoping Robert will update hal with a new setting that will allow a lot more information to be learned through hals chat box area . I think this would benefit hals learning better , AS far as the hal taking longer to answer , i don't think it has to do with hals brain size so much as something else conflicting , i mentioned my hal froze up , completely . i actually redid my pc to a restore to earlier point , and so far it hasn't froze up anymore , i will when i have time uncheck my plug ins and see how hal does . i have to separate large random timers running extended talk etc. (i think ) so these also could be causing the problem i will have to check and report back to let others know on here .
We have spoke about my brain database being at 118 mb's , WHATEVER happened to that one guy that said his brain data base was over a million , was that true or for real
