[Core] Components design
"Henrik I. Christensen" <[email protected]>
| Newsgroups | gmane.science.robotics.orocos.user |
|---|---|
| Message-ID | <[email protected]> |
Hi Anthony Thank you for getting the ball rolling. Two comments 1) I think your module definition is OK for now. I am not sure I agree that we need a complete and unified ontology for module capabilities at this time. We need to develop this over time as modules are specified. I fear that an attempt to do this apriori is doomed to failure. We shoudl of course have a format for specification of the "service" offered by a module 2) As to communication I still propose that we focus on communication patterns as presented by Christian. Based on the patterns and a method for specification of message format (a la IDL) different people ought to be able to implement libraries. But initially the emphasis should be on the patterns. These should of course include client/server, request, acknowledge, ... Thus I think that we ought to review the initial set of patterns described by Christian and decide a) can they be used b) how must they be complemented to provide a solid basis Once Christian makes the document available I am happy to sit down and try to perform a synthesis (in English). I think Christian and I will make a proposal for the rest of you to critisize. Does this make sense Cheers Henrik