How were things before ROS ?

asked 2019-12-25 23:33:07 -0500

electrophod gravatar image

Just curious to ask this question, but I wanted to know how were things which we do today using ROS been done without ROS? ROS is really a beatiful thing and knowing this would only help me appreciate it moee and more! TIA

edit retag flag offensive close merge delete


Everyone was developing every part of the robot themselves; internal/external communication, sensor drivers, behaviours, specific apps etc. Everyone was trying to make decisions on their own about solving every problem while developing.

Orhan gravatar image Orhan  ( 2019-12-26 04:00:28 -0500 )edit

Yes, basically what @Orhan writes.

it took (sometimes) years to set up a robot control architecture, and even after those years, you typically only had partial support for all the "other" stuff that was not your main topic or interest.

I've personally seen PhD candidates spend 3 years on the software of their robot and only in the last year being able to do some experiments. Their subject was not control software, but something else entirely.

You may want to search for some papers about robot control architecture or 'robot framework' from before 2012. That should give you some idea of how many there were, and how many were just waiting for "others to add support for more robots/hw/sensors/etc".

gvdhoorn gravatar image gvdhoorn  ( 2019-12-26 07:32:01 -0500 )edit

this is a relevant paper. Basically, the wheel got reinvented 10000 times. YARP, Player/Stage, and others were a good first effort, but ROS (and more recently ROS 2) seem to be massive improvements.

allenh1 gravatar image allenh1  ( 2020-01-02 16:41:39 -0500 )edit