On 11.12.2013 15:13, Alexander Duda wrote: > On 12/11/2013 02:25 PM, Jakob Schwendner wrote: >>> Our policy is: >>> * stop the component >>> * reconfigure it via properties >>> * restart the component >>> >>> In some special cases if there is no other way than changing properties >> while >>> the component is running the dynamic properties should be used. >>> Can you please point out why this is the case here? >> The use case would be to change the max speed value for the servo, >> without >> having to restart. E.g. your source component only provides >> positions, but >> not speeds, and you want to set the speed depending on e.g. some sort of >> system state. Using an input port here wouldn't do justice to the >> nature of >> the data, I think. As Matthias pointed out, this seems to be a safe >> thing to >> do now, and is meant for exactly these cases. > I have my doubt here. > * basically, you want to combine two commands (speed and position) > with the help of dynamic properties which sounds kind of bad. I thought we disussing a maximum_moving_speed configiration option?!
> * if you have a system state change you most-likely reconfigure your > hole network anyway. Even then, why restart/stop this comonent? > * dynamic properties make it incredible complex to test your components Why? -> every configuration is hard to re-apply so long Syskit supports no real replay... > * there is no logging if you change properties from vizkit Logging of properties is afaik possible > * we have no tooling allowing to reconfigure tasks dynamically > according to log files It's not done even for normal property reconfigurations, you have to log them and manually apply. (same for connections and so on)). The Toolchain supports no "start and use this component as live" it in general supports only rebuilding data-streams manually. > > For me the hole dynamic property thing is a really bad idea and not > needed in 99% of the cases. Don't agree, even for you cameras (like exposure) for example it's needed to reconfigure this without restarting... > > Alex > Matthias P.S. maybe we should move this discussion if there is a need to to the next rock-meeting -> i stopping here now for the public one... -- Dipl.-Inf. Matthias Goldhoorn Space and Underwater Robotic Universität Bremen FB 3 - Mathematik und Informatik AG Robotik Robert-Hooke-Straße 1 28359 Bremen, Germany Zentrale: +49 421 178 45-6611 Besuchsadresse der Nebengeschäftstelle: Robert-Hooke-Straße 5 28359 Bremen, Germany Tel.: +49 421 178 45-4193 Empfang: +49 421 178 45-6600 Fax: +49 421 178 45-4150 E-Mail: [email protected] Weitere Informationen: http://www.informatik.uni-bremen.de/robotik _______________________________________________ Rock-dev mailing list [email protected] http://www.dfki.de/mailman/cgi-bin/listinfo/rock-dev
