Some very interesting thoughts
Got some pretty strong opinions flying around in that discussionâŚ
Indeed and some interesting insights into the short comings of the new guy ROS 2.
I spent very little time with ROS indigo and kinetic, basically proving they could be installed, but not attempting to learn more than the basics, no developing anything.
Then Keith started posting his successes with and corrections for the Hand On ROS Melodic, which got me interested in ROS again but I didnât want to learn the âdead endâ ROS so I attempted to follow the book but do everything in ROS 2. It was a killer idea to convert while trying to learn, and I got totally distracted and discouraged trying to fuse IMU and Encoder data. Then my IMU died so I dropped ROS 2 completely thinking it was essential. (I think my mapping may actually be better without the IMU.)
One thing that struck me square in the face was the simplicity of launch files in ROS, and the giant effort required to build launch files in ROS 2. I ended up punting, and just running things from the command line and putting them in the background. ROS 2 eventually added the same xml and yaml file launching that ROS had, but I never learned that aspect. Still need someday to revisit that.
Yeah, there is a lot of complexity to go around and I canât fault your logic. For me though, I think that ROS1 does have a maturity and depth of information (i.e. way more has been written explaining ROS1 than exists for ROS2) that is attractive, compared to ROS2 so I will stay on the ROS1 path for now.
Didnât realize that. Yeah, the python launching always seemed weird. But Iâm sure there were specific reasons for it.
Still, ROS2 is the future. I hope the ROS community doesnât go through the protracted transition that Python did moving from 2 to 3 â although looks like thatâs exactly whatâs happening.
/K
âAinât broke, donât fix itâ is the norm - itâs just human nature.
And thatâs why I stopped contributing to a large number of fora. I just donât have the patience or bandwidth to pander to highfalutinâ imbeciles who are a legend in their own mind.
Maybe Iâm cheating, but Iâm still not willing to invest a lot of time in a system that doesnât know if itâs pitchinâ or catchinâ.
As for me, thereâs still a lot âo meat on those âole GoPiGo OS bones and Iâm doinâ my chewinâ there.
Maybe when I get back to Moscow, Iâll have something more interesting to contribute.
Indeed, after âa line of code a dayâ challenge on GoPiGo OS, Charlie will be the smartest GoPiGo OS robot around! (There is even a couple neural net vision lessons lurking in there!)