AskBot, the site providing ROS Answers, is a pretty obvious clone of Stack Overflow (SO). It does a pretty decent job but has a few shortcomings: it is much slower than SO and other Stack Exchange sites, it presumably costs money (who is paying for this right now, OSRF?), and doesn’t seem quite as well supported as SO. I think it would also be easier for the community to have fewer such forums. There is already a lot of duplication between ROS questions on SO, Stack Exchange Robotics, and ROS Answers.
It does seem possible to migrate existing Q&A communities to Stack Exchange, see my question on the SO meta page. So we should be able to get all questions, answers, and comments migrated. Karma and badges would probably be lost.
So my question is: Does it still make sense to maintain a separate ROS Q&A forum, or would it be better to migrate ROS Answers to https://robotics.stackexchange.com/ ? Thanks in advance for sharing your thoughts!
Well actually… Stack Overflow is a clone of AskBot; ROS Answers actually pre-dates Stack Overflow sub domains.
This isn’t a priority right now but we’re starting to talk about what the next generation would look like so I am interested in people’s thoughts. We have discussed potentially doing a SO migration but it is fairly complex and we would need a point of contact at Stack Overflow to make it happen. The process would be high touch as we would need them to create a new community and help with creating a migration tool. One possible option I’ve discussed is doing the migration with the OpenCV Askbot website added to sweeten the pot and make it worth Stack Overflow’s time. My big ask is that we get our own ROS sub-domain and not just glob onto the existing robotics subdomain.
A second alternative, and we’ve discussed doing this for Ignition/Gazebo, is standing up a second Discourse website solely for Q&A. This is a bit more fraught as Q&A isn’t a first class function in Discourse.
Anyway, I would love to hear what people have to say. Also if anyone has a business development point of contact I would love an introduction to get a better sense of how this could happen.
More seriously though: Askbot is showing its age. And it’s becoming somewhat obvious the maintainer isn’t really maintaining it any more (I’ve recently given up and closed all my outstanding issues).
If I remember correctly, @tfoote has mentioned on a nr of occasions that using Discourse for a Q&A site is possible, just not with the current tier (or at least, it wasn’t back then, I’ve seen some questions / categories on ROS Discourse where you sometimes see a “mark as answer” button pop-up).
Not sure whether this is still an option. I believe Discourse itself also uses Discourse for Q&A.
Ah, this indeed.
I would also find this important.
We’ve (or at least I’ve) tried to keep ROS Answers really about ROS. So no questions about which motor controller boards work best with RPis, or how to do triangulation of distance sensors based on signal strength of UWB sensors. There’s Robotics SO for that.
If the opposite would happen (ie: ROS as just a tag on SO), it wouldn’t help discoverability of Q&As I believe.
I also strongly support not just chucking ROS stuff into Robotics SO. Discoverability of answers for common ROS problems would plummet, making the user experience worse.
Well actually actually, AskBot (est. July 2009) is a clone of Stack Overflow (est. April 2008). But ROS Answers predates the robotics SO subdomains, that’s correct.
+1 to bundling (and reminding people of the size of the community). Could we also rope in AWS robomaker folks?
I think we should migrate, since it will offload the maintenance to some other entity. I’m a big fan of focusing on core competencies, and I’d rather see Open Robotics focus on robotics and not infrastructure maintenance.
i’ve just started learning ros yesterday and i found myself most of the time here on ros answers, and the first thing that came to my mind, why isn’t this a stackexchange community!! hope you guys reach an agreement with the stack people and go forward with this amazing idea
The migration is still planned. It got put on hold for a few months due to the acquisition of OSRC by Intrinsic, but OSRF is moving forward with it. We just got the process restarted a couple of weeks ago. We intend to give the community 30 days’ notice prior to the transition date, so you will be informed in time. In the mean time, if you’re interested in helping, please start contributing to answering questions about ROS and other OSRF projects at Stack Exchange, but also keep answering questions at ROS Answers until the transition is complete.
I can only imagine the transition of ROS Answers is challenging. Thanks for that!
As one of those who have some answers to provide to (some) questions (call it “experienced-users” just for convenience sake), challenge I see is that there are now 2 different websites that new questions are asked on, so it’s on us the experienced-users to somehow check/get notified for the questions they might react to, on both 2 sites.
To that end I suggested in ros-infrastructure/answers.ros.org#220 to encourage new users to ask only on the new site, stackexchange, but Tully’s answer reads that there seems to be understandably much more considerations need involved than I thought so doing what I suggested wouldn’t be easily done.
So personally as an “experienced-user”, I just stay on answers.ros.org for the most of my time (primarily because the site requires moderation. And there are often 10+ waiting in moderation queue when I open. That’s a separate problem answers.ros.org#210), for now. That’s just my opinion.
answers.ros.org will be put into read-only mode at an appropriate point during the migration. You’re free to continue using it up to that point, but once it becomes read-only users will need to ask their questions on Stack Exchange, and will be directed to do so by a visible notice on the site.
Super late to the party here, but I’ve found getting people as close to the source code as possible tends to help everyone (implementers/users and maintainers/devs). Because of that I’m somewhat partial to using github Discussions for Q&A. A good example of this in use is: iRobotEducation/create3_docs · Discussions · GitHub
Not suggesting anything should change from the current plan, just wanted to add it as a potential thought for package maintainers in the overall federated ecosystem.
In my experience, a migration from one system to another should generally be avoided at all cost, unless there is a clear compelling reason, and a very clear payback (or cost when not migrating).
In this case, I think the benefit could be a reduced maintenance burden for the infrastructure team, but I fail to see how a migration would lead to more and/or better answers.
In any case I would very much prefer that the knowledgable people spend their precious time documenting core ROS 2 concepts, rather than wrinting scripts to migrate posts and user accounts etc from forum A to forum B.
I took a look at ROS answers and there have been 18 questions asked in the past 24 hours, that’s hardly dead. Answers tend to wax and wane with the time of week and the quality of the questions. The Stack Exchange team, along with our own internal developers, are presently working together to move over the vast majority of ROS Answers content. We will make a Discourse post explaining the transition when everything is ready. We don’t have an exact date, but this transition should be fairly soon, probably sometime this late spring / early summer. As you can imagine this migration is a non-trival effort and takes some time.
We believe this transition will greatly benefit the ROS community. AskBot, the software that runs ROS Answers, while maintained, is not under active development and as such it is difficult for us to bring new features to the ROS community. By moving Answers to Robotics Stack Exchange we will be able to make use of the improved moderation capabilities available on that platform and get answers from a wider robotics community. This reduced moderation burden, and larger community, should enable us to spend more time answering questions and grow the Answers community.
The answers on this thread also give some background.
Let me substantiate this with an example. Have a look at this simple question. I tried hard to answer it. But it is impossible. For someone like me (having a significant c++ knowledge, but being a mechanical engineer, i.e. coding is a necessary tool and neither a passion nor something daily used on a professional basis), it is impossible to grasp the code concepts in a reasonable time frame. The code is hard to read, and there simply is no documentation.
Maybe if you’re a computer science major with ample code reading experience, or maybe if you’re in a team with colleagues that have been using ROS / Gazebo since the early days and know it inside and out… But if you’re on your own: good luck.
Time will tell. I hope you are right. But I very much think this is wishful thinking.
As long as people cannot easily become knowledgeable enough to answer questions such as the abovementioned one, there will be many people asking questions, and not many people able to answer.
I think a wire got crossed here! The statement you’re responding to and this discussion in general was that of making it easier to take care of the question-and-answer community infrastructure, not about adding features to ROS itself. It’s an orthogonal conversation - I’m sure that everybody who works on or with it will agree that ROS 2 needs more documentation
Maybe you were coming at this from a resource allocation standpoint, but I don’t think that the people spending time on migrating ros answers are the same people who would be needed to write the better documentation - it’s not really a fair or constructive point.
I did not interprete “new features” in above quote as “new ROS features”, but rather “new forum functionality that will lead to better answers” (i.e. “greatly benefit the ROS community”).
I tend to disagree that the current forum itself and its lack of features are the bottlenecks to getting more and better answers. I think the real bottleneck is that very very little people know enough about the internals of ROS 2 and Gazebo, to easily answer questions that go beyond the basics.
Anyway, I am not against the migration per se. I think the “reduced maintenance burden” and “general look and feel” and “people like SE more” etc are all valid arguments.
I just wish there would be some sort of a plan or strategy to get to better documentation on core functionality.
