We’d like to get an idea of what versions of ROS people are using and how long they want those versions supported. If you could fill out this short survey, that would help us to make decisions on where we should focus our efforts for ROS1. We’ll leave the survey open for 2 weeks, until April 4, 2018, and post the results shortly after that.
I think it might be helpful to add a ‘N/A’ option for some of the ”how long” questions, or at least let them be optional, as it might otherwise conflate reported preference of duration with those that have no opinion. E.g. those who never use non-LTS releases, and are not concerned with it’s EOL. With this option, we could still take the selected duration and number of responses into consideration.
I already took the survey, but if there had been a “n/a” option for the question about the release cycle for non-LTS releases, I and most or all of my colleagues would have chosen that one.
Perhaps these answers can be teased out, however, by filtering those who selected “Never” for “When do you upgrade to a new ROS1 non-LTS distribution?”
Good point @ruffsl . In retrospect, that would have been a good thing to add. In order not to have different surveys for different people, I think I’ll leave it as-is for now and we can tease that out from other answers like @jatowler suggested. Thanks for the feedback.
We wanted to focus this survey on ROS1 only here, since that is where we want to know where to focus our resources. We certainly could do a ROS2 one, but there would have to be specific things we would want to answer with the survey. What kinds of things were you interested in learning about with a ROS2 survey?
I think that the questions about when you usually update to a LTS or non-LTS release are important and relevant, but that they also miss an important aspect: why people wait as long as they do (or do not wait). For some people, the desire to update to an LTS release or even a non-LTS release may be held up by the need for a package that is not yet available in that release. I know that the time needed to get MoveIt! into Kinetic held me back on an older release much longer than I wanted to be. It’s too late for this survey now but next time I think a question about why people upgrade when they do should be asked to find out how many are actually waiting on a dependency and how many are waiting for some other reason (product development cycle, for example).
Yes, the “why” aspect would have been a valuable feedback item to investigate.
I’ve indeed encountered the reciprocal scenario, where instead of needing to maintain a legacy release install due to slow software migration, I’ve had to upgrade to bleeding edge release install due to fast hardware adoption. E.g. when my lab recently procured new motherboards, graphics cards and peripherals that were not easily supported with previous linux kernels/distros.
Thanks, that is good feedback. I’ll add this (along with the earlier feedback) to the notes we have so that when we run the version survey again, we can get more information.
LTS is the main critical aspect while building Industry grade robots .
Indigo is what we started out with (Start of 2017) but towards the completion of the software stack , we gradually migrated to kinetic .
Ubuntu 14.04 was the norm for a long time , but now ubuntu 16.04 LTS is the norm .
14.04 might have stable ROS support but this is not the case with other non ROS projects especially when dealing with deep learning .