its a good point actually... The students will know java really well and some will know ruby really well too, and I was assuming I could just assume they could pick up python. maybe a bad assumption...
Well look - that's the long term vision, I know now that we will not get that done in even multiple semesters but we will learn and teach a lot as we strive for it :)
python language is easy to pick up. the trick is the environment... complex tools, different packages, standards always changing... and ROS is a very old framework using deprecated python features on top of that.
In general I think the software level needed for doing robotics is much higher than what people usually... so the goal of robotic platform writer should be to make things simpler. but ROS sadly fails at that, making everything distributed and asynchronous makes everything so much more random and unpredictible.
If I understand correctly your summary for Robot API, you want the students write the web server that connects to ROS ?
yep : ROS on the robot : topics communication for sensors. ROS services to expose that. ROS services are just xmlRPC, so then you can use flask directly, serialize your json to ROS format and send with xmlrpc.
yeah people usually overlook services because of this comunity focus on topics (which for me sucks since they are lossy. cant develop complex protocols on it)
with the company I m currently in, because we need highlevel interaction wiht the robot and we coudlnt find anything already there... But my contract run out soon, so I ll need to find other means to keep working on that...