We recently got a question on Physics.SE about how to properly simulate multi-body collisions that occur within a time step. The OP says he was directed away from Stack Overflow, but it seems off topic for Physics.

The next obvious candidate for where this question should go is this site, but the question is about a pretty simple programming problem. As much as we've decided that "heavy computation" is not the optimal description for this site, it seems that that wording was chosen to suggest that questions like this one about collisions are not what this site is about. Is that correct?

And the main question: Are simple simulation questions like this on topic here?

(It's likely that this particular question is appropriate for Stack Overflow and will stay there. But I'm posting this because I think it will be good to know what our position is on this kind of question.)


I've voted for Geoff's answer, but also thought I would answer one of my own - especially since this has come up on a question on the front page.

I think questions about "simple" simulation are absolutely on topic here, for a few reasons:

  1. SE site precedent: If you look at one of the sites I'd consider a logical "sister site" to this one, CrossValidated, there are a large number of simple simulation questions revolving around statistics. These questions are fairly active, are answered, and can potentially be quite useful - for communicating concepts, teaching good simulation practice, etc. That's also ignoring Stack Overflow itself, where I've asked myriad idiotic questions.
  2. Traffic. I think the defining struggle of any new SE beta site isn't going to be "fend of a sea of newbie questions". It's going to be "Don't become a ghost town". New questions, even basic ones, help drive site traffic. I've yet to see folks - including myself - driven off a site like CrossValidated because of the occasional easy, straightforward question.
  3. "What's an expert?" I'd rather not see this site use some sort of community standard "I know a beginner question when I see it" approach. Does a subject matter expert just moving into more advanced simulation count? Someone whose experienced with sophisticated models, but ones that don't need HPC, and is taking baby steps into that area? An experienced simulationist trying to figure out if the techniques they know apply well to a new field? What's advanced for one field might be basic for another.

I'd rather see an active, helpful, vibrant community that has to deal with the occasional easy question than one that repels newcomers who might at some point turn into valuable contributors.


My opinion is that any scientific simulation question, no matter how simple, where the question is about the numerical methods, software usage, or even appropriate modeling (from a computational tractability perspective), should be on topic for this forum. I think it would be valuable community outreach for computational science experts to help people who don't intend to be computational scientists. (For example, I have experimentalist friends who occasionally ask me what numerical methods are appropriate to use in their Matlab scripts, and then they use the appropriate canned routine I recommend.)

Multi-body collisions sounds like a scientific problem, and it looks like a numerical methods question, so I think we're the community to answer it. (Specifically, it seems like what the poster is looking for is some sort of numerical method for solving ODEs that detects collision events; there are numerical methods out there that do this sort of thing.)


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .