Brain storming with the cave man.

[Ooglor 7]

Hi guys,

Does anybody know of a script that adds a brain editor button link to a Ultra Hal GUI skin? Also, it would be cool if Hal stayed open on the desktop and just refreshed from changes made in the "General Options..." menu. I wish that the brain editor program was integrated into Ultra Hal more. Like maybe a fully functioning mini brain editor that opened from an attached window on the skin itself.

[TechSteele]

I am not sure why you would want it open and docked at all times.

You can get to it by useing the menu button on hals Face, or the icon in the tray.
You can even have it open with hal running if you so desire.
Only thing is for any changes you make to become active you need to hit options, or close then chat, to reload any changes.
I think i even read a script that will open it on command with hal, in the forum somewhere.

I sometimes have a text editor, the brain editor, Hal, and Paintshop all open for testing.

As for adding a button, we only have the 4, with no way to add any, or so it seems, to me It looks like Its hard coded in.

[Ooglor 7]

Not opened and docked at all times. Open when you hit a button. Closed when you hit it again. As much as you are not sure why I want an easier way to run the brain editor, I don't understand why one would have to basically stop Hal entirely to select from what is really just a customized right click menu and load an entirely different program, load screen and interface, just to make a change to the brain you are running in the program you just closed.

[Ooglor 7]

I guess what I am trying to say is Waaaaahhhhh. Maybe when I get a better understanding of things I won't be such a big newbie cry baby. On a more positive note, I just bought People Putty and the add-on pack! It looks very cool. Hope to be contributing more than just questions and grouchie replies real soon.

[TechSteele]

One of us is not reading right.

 

quote:Open when you hit a button. Closed when you hit it again

 

quote:would be cool if Hal stayed open on the desktop

If you hit the Menu Button on Hals GUI it has a Brain Editor button in the list, or you can use the tray icon to bring up the list/menu and open it there (the little green tv with eyes).
You do not have to close Hal at this point.
You can apply changes to the brain or just close the editor at anytime with hal running.
If you are working on the UHP file, you can work on those and save them with a text editor at anytime with Hal open.

 

quote:just refreshed from changes made in the "General Options..." menu.

Using the same Menu button (on hal or the tray icon) you can load your changes and/or test/debug new code you are working on. By simply hitting General Options in the menu, Then OK, Then hit the Chat With Hal button, and Hal will reload with your new changes and nothing on him as far as skin or character should change.
Another way to reload any changes is to hit the Close button then the Chat With Hal one, to reload the brain.
There is NO program i have ever seen that lets you Compile Code on the fly while the program you are editing is running.
There might be some fancy Module based program where you can work on one part while another runs, but (however much id like to see more of that in hal) this is not one of those.

Some people have had success useing an SQLite editor along with hal, but i believe the editor with hal is better then any i have found.

As far as attached/docked, i guess you can resize the window.

The People Putty, good luck with it, i played with it briefly, but didnt get the hang of it. The first one i made would have made the Elephant Man look HOT [:p][xx(][:p]

[Ooglor 7]

Isn't Hal "learning" as the program is running?

[TechSteele]

Yes, Hal is "Learning" as its running.
You would only be working on a Snap Shot of the brain when you open it in the editor.

Hal is not like a baby, constantly entering data from senses of his surroundings, he may learn a couple words or line of text, and that gets saved while you work.

When you reload hal, it enters the adjustments you and he had made to the brain file.
It just dumps the Cache? in from memory.

It does update files on the fly, like some text files, or UHP files depending on what you have pluged into hal.
Every time you reload, it writes the new config/state to the HalScript.dbg for example.
This is why you have to reload hal to have the new information written, added to what ever he had written.

One way to look at it may be, you are working on a picture of a list, and he has the original, then you combine the two in the end.

If any of that made sence.

[Ooglor 7]

Gotcha' So when interacting with Hal, it isn't the entire brain loaded into RAM. I like bouncing ideas out and getting the kind of intelligent feedback you are giving. I consider the forums like a brush or pen. An artistic extension to a train of thought. I worked in a hard core industrial steel foundary here in Ohio, cutting steel for years until I got a hernia and emphezema. If I come across kinda harsh, don't be put off. It's just the cave man in me. I bought my first Mac back in 1984 and I have been picking things apart ever since. I truly appreciate creative minds to scream out to now and then. For that, I thank you.

[Art]

There's been a lot of thought given to it's learning mode.
Would it be better if Hal had a huge knowledge base (brain) from which to draw OR would it simply be better if Hal could ACCESS the info from whatever source when needed, thus saving the need for huge memory requirements, speed reduction, etc.

Searching sites like Wikipedia or other online sources is not always the best method of acquiring data and it often produces less than desirable results not to mention garbage, ads, spam, filth, etc.

Hal's training / learning must be controllable much like a parent monitors their child's learning / associations as they grow.

Perhaps there is no one perfect answer as there are no two people who think alike in all aspects of life. It's basically YOUR Hal...raise it, teach it, develop it as you see fit. Unlike a child, if you aren't satisfied with the outcome, erase and start over!

Thoughts?

[TechSteele]

Personally, I would like to see a bit of both.
I want to teach hal his personality, Likes, Dislikes, general information and so on,
But, If there is something he does not know, or if i ask him, he can go get more info then just what he has stored in his list.
Sure he can open Google, and even gets search terms correct from time to time, hehe, But, its not the same as, lets say, in movies/tv show, when you ask the computer a question it gets your information For you.

One example would be, i say to Hal "what do you know about Alchemy"
If i have entered any info he could recite it, and then if i say i need more he can download from Wiki or somewhere more info.

I do have the Wiki Learning plug in, but i have to tell him ahead what to read, and then its not easy to get back out of him.

I would like more interaction on Hals part. I do know thats a tall order the way it is now.

*edit, Oh and Saving the info he gets as well, of course. I thought id add that in.

[Ooglor 7]

Perhaps if Hal had two seperate sections. Personality and knowledge. His personality could dictate how he utilizes the knowledge. What about a set of typed commands that informed Hal what to do with the knowledge given to him. "You are a gardener." and he could respond "What do I know about gardening?" One would then tell him to learn from the provided web site or document. Maybe if he took each sentence as input from the user as he read and then could ask the user if the statement he learned was correct. As he is reading, perhaps he would go through a series of If/Then routines on his own with the information provided. That way he is not merely repeating input but has established knowledge that his personality could work into his responses. I honestly am more of an art and idea man than a coder so maybe I am speaking gibberish. I hope that is not the case.