Updated self+world-model document

[Linas Vepstas]

Attached is an updated version of the self+world-model document. The technical content has not changed much, since whatever previous version you might have seen.  However, this version includes some glitzy new graphics and a re-written introduction, with the goal of allowing a less technical audience to follow the proceedings.

It seems I've finally got the spare time to focus on this, again, so I hope to add proper technical content "real soon now".

I hope that this document might be exactly what a lot of newcomers to this mailing list are looking for.  It does not provide a complete overview of opencog, but it does exhibit one corner of it, as it is currently applied to robotics. It does name-drop the names of many or most of the major opencog subsystems, even though, in the end, it just focuses on one small, narrow area.

--linas

[Apil Tamang]

Hi Linas,

We appreciate the work you put, and it is definitely more informative (conceptually) than any other single piece of document I've found on opencog.

A question I have is where does this infrastructure exist (or is being developed) within the existing code-base? The impression I have of core opencog at the moment, is of an enabling framework. There is something called a 'CogPrime' circulating amongst the wiki, Ben's books, and this discussion group. Then there is the system called the 'HEAD' which I presume works on a real robotic head. And then there're some chatbots working in a blender platform. How are these concepts and code-bases linked to each other? 

Any insight would be appreciated.

[Linas Vepstas]

On Wed, Nov 16, 2016 at 10:03 AM, Apil Tamang <apil.tama...@gmail.com> wrote:

Hi Linas,

We appreciate the work you put, and it is definitely more informative (conceptually) than any other single piece of document I've found on opencog.

A question I have is where does this infrastructure exist (or is being developed) within the existing code-base?

 

Chapter 3 is a line by line review of the actual code. The files and the line numbers are given there.

 

The impression I have of core opencog at the moment, is of an enabling framework.

yes 

There is something called a 'CogPrime' circulating amongst the wiki, Ben's books, and this discussion group.

There is no cogprime code.

 

Then there is the system called the 'HEAD' which I presume works on a real robotic head.

you don't need a physical robot. It will work on your computer just fine.  It displays in a blender window.

 

And then there're some chatbots working in a blender platform. How are these concepts and code-bases linked to each other? 

The chatbots+blender is very much like HEAD, but more stable and easier-to-use, fewer bells and whistles, but also much less mess.

--linas