letter from bionic robots developer from Germany

[nortikin]

Gerard writes:

I looked at the github issues link.

Yes they are doing similar work to me. Good to see others pushing it :)

We are also looking at how best to leverage Blender for the design
phase and the browse based system for the operational phase.
We are driving many robots with vision and depth sensing, and wish to
use blender for integrating with the proposed design VERSU the current
design of a 3D object.
This is really the key.
So, It would be great IF there was a way to run blender as a server
for processes and the vertex outputs can be piped over http.
After all The Nodes are just pumping vertexes

I doubt anyone is using blender this way. But it is highly
advantageous because the art directions of Physical object creation
can then happenin blender, and the physical making of the 3D object n
the real worls can also be driven by the same nodes that worked in
blender.

This is really the only way to make this usable.

Please can you forward to the other guys and see if they have some
technical ideas on this in terms of running as a server and  also
piping the vertex translations out

i think his tasks much wider than my possibilityes and i don't know his finantial status or so. Just for discussion.

[ly29]

It is interesting and certainly pushing the possibilities much further than we do.

[zeffii]

let's see some pictures, diagrams. I've read this a few times and still feel uninformed as to the nature of the data they want piped out of sverchok. Robot movement data or geometric object data or both, in what kind of format do the incoming instructions come.

[ghost]

we have none. :) Your demand for it is understandable.

currently we are working up the quadrature encoder equipped motors. at the same time a FBP based simulation system. at the same time the mechanical design and the chemical polymers

at a high level the positions of the "objects" in 3d space are streaming as events. this is all you need to knwo really whatever the data "shape" or "format" is in does not matter because its simply a data maping exercise on the way through.

then there is the sensor data. Depth fields etc. at this stage that is voxels. Again they can be shaped in whatever vertex data structure you need.

[zeffii]

FBP being flow-based programming ?

so: (input: per unit time = verts+edges+faces) --------> ( throughput: unclear ) -----> ( output: unclear ) (input: per unit time = sensor data) --------> ( throughput: unclear ) -----> ( output: unclear )

[ghost]

yes its fbp

we will have actual data streams soonish. but cant promise when. as you can see we have lots of other aspects to take care of.

Once we get data streams coming OUT, i will then be able to show you.

[zeffii]

OK:) until then please understand reserved enthusiasm (from me at least)

[ghost]

We are also very grateful for the enthusiasm. So appreciate you saying this. When we have data and something going we will get in touch.

[zeffii]

closing for now.