The new eva-embodiment directory [was Re: [opencog-dev] Re: Stop merging to the opencog repo

[Linas Vepstas]

Hi Nil,  Hi everyone,

Sorry, yes, its intentional,  got discussed in a meeting here last week, and therefore "everybody knew about it" ... except, of course you :-)

To recap, this is https://github.com/opencog/opencog/pull/2641 which resolves a circular dependency between opencog and the robotics code.   Its now arranged so that all of the thinking/understanding/behavior is in the main opencog repo, while all of the ROS nodes providing sensation (vision, sound) and all motor movements (3D animations) remain in the https://github.com/opencog/ros-behavior-scripting directory. 

This allows, at least in principle, other, non-ROS sources of sensory inputs, and also maybe someday a better, more advanced motor-control interface; right now it remains closely tied to the existing blender animations.

The moved code is in https://github.com/opencog/opencog/opencog/eva directory. In a few days time, select portions of the chatbot will end up there as well, although hopefully most of the nlp code will stay in the nlp directory.

I've provisionally named the directory "eva" because, at least for now, it is tightly tied to the eva/sophia blender animations.  Perhaps in the future we will be able to refactor this into a more generic, disembodied intelligence, with sensory input from wherever, and motor control over something.  But for now, its not entirely clear how to split this up, and so the directory is named "eva" instead of "embodiment".

One prominent subdirectory is "model" which contains a provisional "model of the world" (which includes a model of the robot itself).  The code and ideas here need a lot of elaboration.

Another directory is "behavior" which currently contains a mish-mash of hand-authored stimulous-response type scripts, and the plumbing to enable these.  This will get replaced/updated by a better authoring system that Cassio is leading (a chatscript-like interpreter)

I am not quite sure where or how PLN will get wired in -- it will get stapled into the language parts, somehow.  I won't be doing this work.

OpenPsi is attached to this stuff, but its sort of a dr-jekyll/mr-hyde type system, uncertain as to its role in proper society.

I've started a refactoring of the space-time server code, but am running out of time.

All of this code moving-about will probably generate a long list of "buglets", breaking how things attach and talk to each other. I will appreciate all help in diagnosing and fixing these.

--linas

On Fri, Mar 24, 2017 at 8:37 PM, 'Nil Geisweiller' via opencog <ope...@googlegroups.com> wrote:

Actually never mind, it comes from

https://github.com/opencog/opencog/commit/3b2637e6c0c774ac06ecc779a265564fa616be0e

which I suppose has been done purposely. It was looking very suspicious at first because it contains commits as old as 2y.

Nil

On 03/24/2017 02:29 PM, Nil Geisweiller wrote:

The github history on the opencog repo is seriously f-up. Please do not
commit anything before I figure out what went wrong.

Nil

--
You received this message because you are subscribed to the Google Groups "opencog" group.
To unsubscribe from this group and stop receiving emails from it, send an email to opencog+unsubscribe@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.
Visit this group at https://groups.google.com/group/opencog.
To view this discussion on the web visit https://groups.google.com/d/msgid/opencog/f20bb1a4-6ea0-d6bc-56e0-22d07c9f5da2%40gmail.com.

For more options, visit https://groups.google.com/d/optout.