Re: Bertrand Meyer on IDL, components, etc..

I agree on some but not all.
He argues that the interface should be extracted from the code.
I think that by first having to specify the interface makes you really 
analyze your components task and how it should be used by other components.
This should result in better code IMO. His way may lead to a mess. (Depends 
on the programmer of course.)
Regarding the documentation, I agree fully.

./Anders


On Wednesday 06 February 2002 20:39, Herman Bruyninckx wrote:
> Here <http://eiffel.com/doc/manuals/technology/bmarticles/sd/dotnet.html>
> is an interesting article by the father of Eiffel, Bertrand Meyer, on
> some issues we talked about last sunday in Pisa: the shortcomings of
> CORBA's IDL as a good component interfacing and documentation
> language; and the distinction between a ``Service Request Broker'' and
> an ``Object Request Broker''. The former can work with a `meta-data'
> description of a component's interface, without the need of a specific
> (IDL) syntax.
> Especially the section on ``Documenting components'' (with
> the remarks on the advantages of meta-data with respect to IDL) is
> relevant to the current state of Orocos: we have to define how we will
> specify and document the interfaces of our components. And maybe we
> can profit from the Mono (<www.go-mono.com>)/.NET lessons in this
> respect... Does somebody has a good pointer to documentation about the
> state of affairs about these matters in Mono?
>
> Herman
>
>
> _______________________________________________
> Orocos mailing list
> [email protected]
> http://mail.mech.kuleuven.ac.be/mailman/listinfo/orocos

-- 
Anders Oreback                         Phone: +46 (0)8 790 6204
CAS/CVAP/NADA                      Cellular: +46 (0)70 749 2103
Royal Institute of Technology        Mailto:[email protected]
SE-100 44 Stockholm,SWEDEN      http://www.nada.kth.se/~oreback