Re: Mini meeting Toulouse 22-23 april / Execution Engine

Herman Bruyninckx wrote :
| > I think we have to decouple the terms of "functional" robotics components
| > and "object oriented" components.
| 
| Any suggestions? I must confess I have no real clue as to what you
| mean exactly :-)

Nothing new there, we already mentioned that in the last meeting :)

I just wanted to come back again on the fact that e.g. the
Hardware Abstraction Layer can (should...) be build with a component
approach. But those components would not be the same as the components
on top of that layer.

Should we define "robotics components", which would implement particular
functionalities (and be modeled with GenoM and CORBA and ...) and 
"software components", which are classical reusable software entities
which we can use to build the "robotics components".

Make sense ?


| That's an interesting view point... One developer is the one who builds
| the architecture; the other builds the application? Or did you have
                                         ^^^
actually the "robotics component".
In a previous mail from yours you had defined another user which builds
an application with a set of "robotics components" (and I agree with
that separation).


| In your figures I have the impression that the Execution Control at
| the User level is (or could be) _architecturally_ the same as the
| Execution Engine inside the component. But maybe I read the figures
| differently than you...

I think the "execution control" can be seen as an ORB (it is
actually part of the "system level"). The execution engine has no
equivalent in the CORBA world, since it is part of a (robotic) component
and CORBA doesn't say anything about the internal structure of
components.