Teaching with Hal

[Harkle]

I just purchased Hal last night, and I have to admit I am totally newbie to all of this. But it seems to be what I am looking for.

What I want to do with Hal is have it parse wikis about subjects that I would be interested in and auto enter it into it's brain when it is idle. And after a certain length of time it would expire this information. Is there a way to do this? Any help is appreciated

[Baerdric]

quote:Originally posted by Harkle

I just purchased Hal last night, and I have to admit I am totally newbie to all of this. But it seems to be what I am looking for.

What I want to do with Hal is have it parse wikis about subjects that I would be interested in and auto enter it into it's brain when it is idle. And after a certain length of time it would expire this information. Is there a way to do this? Any help is appreciated

That would be the Holy Grail in my life.

There certainly is a way to do it, someone with time and talent will need to write a plugin that does the job. It doesn't exist yet, as far as I know, but it is certainly do-able.

1) Start a database of topics sorted by frequency of use in user input
2) Use the inactivity timer to begin a search
3) Parse out all the garbage on the found pages
4) Parse page into Statements or paragraphs
5) Write to general knowledge database.

Easy as pie... except for those two parsers...

This is exactly why I was asking about a method to read information directly into Hal by voice. Better than writing a parser, from my point of view.

[Harkle]

I have been digging through the list of plugins people have been creating for hal for the past several years. I do believe I have stumbled on something that is somewhat useful.

<Plugin> Auto Knowledge.uhp
<What does it do?> Auto Wikipedia Knowledge

<problem> It only captures up to so many characters / words at the very beginning of the webpage. Which tends to be a advertisement, and useless information than any relevant information.

Anyone have any ideas on how to fix this? thanks

[Baerdric]

I'm not ready to dig into it, but it has to be that parser I was talking about.

The parser would be built to find the tags that Wikipedia uses to define the main article section of the page, remove any HTML in that section and then break the rest up into phrases.

That's a fairly big and tedious job, in my estimation.

[spydaz]

its probably to do with the selection process, where the part that needs to be selected is the "body" of the text<< this is the part which contatins the information
but the source for the wiki pages are very complexed.

but the "edit this page" source is MUCH CLEARER

THE INFOMATION REQ: IS BETWEEN </textarea> TAGS