Thanks for contributing such a well-written, well-researched, and well-considered question here!
I'm still unable to come to a very strong opinion either way. My hope with this site is that we can help people. There's been questions which were less about matter modeling and more about scientific publishing or high-performance-computing, and they never concerned me too much because someone here was able to answer it (i.e. to help the user).
I'm far more concerned about the questions we have which remain unanswered, for the following reasons:
If people come here to ask a question, and it doesn't get answered, they are less likely to come back and ask their future questions here. If they are new to the SE network, I would also then feel bad about them getting a negative experience with their first attempt at asking a question on SE (I personally want more of the world to participate in SE, since I know/understand SE rather than Reddit/Wikipedia/Quora/etc., and the more people here the better!).
The bigger the unanswered queue gets, the harder it is to defeat it. I've personally experienced this myself, having found the size of that queue to become very "overwhelming" sometimes, though much more "manageable" at other times.
A post can be migrated away after a maximum of 60 days. On multiple occasions, when we were nearing the 60-day mark for an unanswered MMSE question which I thought may be appropriate at a different SE site, I talked to the Engineering.SE moderators and then flagged the MMSE question for migration after getting permission from the Engineering.SE mods:
I think we can take the same approach for that
gnuplot question. StackOverflow does seem to answer
gnuplot questions quite quickly (at least in recent times), and overall only 1286 of their 6194
gnuplot questions remain unanswered (79.24% have been answered). The answers to the only other gnuplot question we've had here: How to use gnuplot to draw Bandstructure and DOS from VASP outputs?, seem to demonstrate that the MMSE community has some very capable
gnuplot users who have written extremely details and valuable answers, so when I saw the new
gnuplot question I wasn't too worried that it would remain unanswered for very long, but I'd totally support migrating it to SO if it doesn't get an answer within about 50 days.
For the other two questions: I don't think they are as off-topic as for example "What is the difference between civil law and criminal law in the USA?", since they are questions that have come up for Matter Modelers while they were attempting to do their Matter Modeling research, and evidently our community had the right experts that were happy to answer the questions quickly and thoroughly. I see the following benefits to allowing such questions:
- Perhaps most importantly, the user that asked the question, got their (very good!) answer very quickly, which means they got a very positive experience on the SE network
- They gave many of our MMSE users opportunities to participate and share the knowledge that they were keen to share with people when they signed up for this site. Some of these users wouldn't have got those opportunities to participate if these questions were asked elsewhere, because they simply have not (yet) chosen to join those other SE communities or to participate there often enough.
- They have brought in a lot of traffic to our site, and have increased our volume of Q/A in the MMSE database, which will help attract more people to learn about our existence via search engine results (for examples): Let's not forget that in order to achieve our goal of being the most helpful site for matter modelers, we need more people to know about our existence.
I can also brainstorm some possible disadvantages to allowing such questions:
Perhaps not being "laser-focused" on matter modeling can lead to some matter modelers losing interest in the site? I'm not sure how convinced I am of this point, because surely a lot of matter modelers come here because they're avid users of Gaussian, and they have absolutely zero knowledge or interest in VASP, yet they see far more VASP questions than Gaussian questions, and this I see as being an even bigger problem for such people (who would lose interest in a site because there's questions on it that aren't part of the laser-focused topic that they came here for) as having an occasional question about scientific publishing which may or may not benefit from matter modeling knowledge in the answers.
Perhaps these questions are not related to mater modeling enough, which means they are susceptible to occupying space in the unanswered queue which can be detrimental to our goal of having all the MM-related questions in the unanswered queue answered ASAP.
I think after writing that last point, I see where I might want to draw the line: If the question is likely to be harmful to our site/community by occupying extra space in the unanswered queue for long, I'd definitely like it to be migrated away if there's another site that would be welcoming enough to the question (so as to make sure our MMSE user doesn't get a negative experience such as having their question relegated to a site that gives them a bad experience).