Re: My 2c

On Thu, 9 Aug 2001, Anthony Mallet wrote:

[...]
> We've already seen in this list that many people will come with their
> architecture design and ideas (and have very good reseons to do so).
> I think that the challenge is --- at first --- a matter of
> "compatibility" between all those systems, rather than trying to find
> which one is the best.

"compatibility" doesn't seem to be the correct term: it suggests too
much that our efforts would go to glue legacy code together. I want to
achieve more: build a new consistent framework (such that
compatibility follows ``automatically'').

[...]
> Don't you think that, for instance, it would be great to agree on common
> data structures?

Yes! At least for topics which have matured enough during the last 5
decades of robotics: the low-level control, kinematics, etc.

> Ok, but I think it will be necessary to define and use some common
> communication systems (or, at least, interfaces) in order to obtain some
> level of compatibility between the various systems. If you decide to use
> CORBA and I decide to use some other protocol, it won't be easy to share
> what we'll build on the top of those protocols!
That's a challenge indeed... But the existence of public standards has
always been the best answer to this challenge, even though there will
eventually always be performance issues, sooner or later.

> I don't know much of CORBA and IDL, but I think (?) they're just a kind
> of spec. From what I understand, choosing CORBA does not imply choosing
> an implementation of it

That's right. There are lots of implementations available.

> At the momment, a question that I am asking is: does IDL provide
> sufficient level of description/abstraction to suit our needs?
> From the experience of GenoM, I would say no. That is: we must extend
> somehow the abstraction of the description we'd like to share.

I would like to see some examples from your experience. At this
moment, I don't fully see where the limits are :-)

> | But I do see some useful applications for CORBA (or SOAP, or...), at
> | higher, less time-critical and machine-dependent levels. For example:
> | I want to test your motion planner on my system, or you want to
> | compare my pattern recognition algorithm with yours, etc.
>
> That's what I called "blocks sharing" above. I think it's a key issue,
> and I think we should aim at this.

Some other names for the same thing: CBS (component based design),
holonic software engineering, intelligent software agents, ...

> | > I will be attending the meeting at the start of September so we can
> | > continue this discussion over a beer or two. :-)
> | Great! But I'll leave the beer drinking to you, and take a glass of
> | wine, if you don't mind :-)
>
> Could I have both ? ;)

Leuven is a paradise in this respect!

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