Re: [Core] Components design

On Fri, 7 Sep 2001, Anthony Mallet wrote:

[...]
> Your remark make me think of this:
>
> There can be *several* modules providing basically the *same*
> functionality to the rest of the "system" ("system" should be in the
> glossary ;). For instance, I want to be able to run several localization
> algorithms at the same time.

Indeed! This is more than an architectural thing: it's really a
"software pattern/framework" that we have to develop, because the need
for using/coordinating multiple sensors `at the same time' is very
high. (At least in our research :-)

> An implication of this is that the algorithms encapsulated by the modules
> should *not* make *any* assumptions on what is the underlying system.
> This probably means that functional libraries should be built as a set of
> functions that will just work on the data without taking care of
> recieving or sending them (the latter beeing handled by the system).
>
Yes. But there will be exceptions, in the sense that some modules will
be system-dependent (motion control etc.). But limiting the number of
system-dependent modules is an interesting design effort. Basically,
we are building a "run time" (or "virtual machine") for robot control,
i.e., the infrastructural support on top of which robotics algorithms
can run without having to bother about the hardware of the system.

Herman

-- 
"I decry the current tendency to seek patents on algorithms. There are
 better ways to earn a living than to prevent other people from making
 use of one's contributions to computer science."  D.E. Knuth, TAoCP 3