I know this is a frequently discussed topic but I am not happy with the current stance of treating Computational Physics related questions as "Off Topic".

A couple days ago I asked this question (Molecular Dynamics Software for Coarse Grained Polymers) about the suitability of different molecular dynamics software for coarse grained polymer simulations. The question was marked as off topic.

I have been reading the previous discussions on this topic and I cannot see a clear consensus but I would like to put forward some of my thoughts:

  1. There are no other places in Stack Exchange where this question fits. SuperUser, Overflow, softwarerec all focus on general computer tools or programming and development. I don't expect my question on coarse grained polymer simulations will get much response there. I want Physicists and researchers to answer my question.
  2. Not allowing computational physics related questions unnecessarily splits the content of this site. Just because my question could fit on another stack exchange site shouldn't be reason to remove it from here. The question is related to physics research and this is the Physics site.

  3. The help and tour pages say that questions about "Experimental technology used in physics or astronomy" are allowed. I would consider molecular dynamics software to be part of the technology that is used in Physics.

  4. It seems that the main concern is that the site will be spammed with coding or general tools questions. I can see why this is a concern but I think the community is being overly strict and cautious and as a result is cutting out potentially worthwhile and useful questions.

I would like my questions status to be reconsidered, but more generally, I would like the Physic.stackexchange community to reconsider how strict it is on questions related to computational physics.

  • 8
    $\begingroup$ Couple of points about "no other place". First that is widely agreed to be irrelevant to the topcicality of the question: there simply isn't a guarantee that even a good question will find a home of some site. Secondly, you didn't explicitly list Computational Science, which is a fairly obvious candidate to consider. $\endgroup$ Commented Jun 2, 2017 at 2:35
  • 3
    $\begingroup$ Computational Science does seem like a good spot. However, that doesn't change my thought that physics related questions should be allowed on the Physics site. The existence of other possible places for the question should also be irrelevant. $\endgroup$
    – user668074
    Commented Jun 2, 2017 at 2:42
  • 1
    $\begingroup$ BTW—Now that I've looked at you question I doubt they'll accept it on Computational Science either because of the network wide distrust of recommendation questions. $\endgroup$ Commented Jun 2, 2017 at 3:14
  • $\begingroup$ @dmckee, That's unfortunate. I'm not familiar with the Computational Science site. It looked promising. On their help they say ask questions about: "Software packages or languages used in computational science". Where do you see that my question might not be welcome? $\endgroup$
    – user668074
    Commented Jun 2, 2017 at 3:18
  • 1
    $\begingroup$ Because it is a recommendation question. See their meta post scicomp.meta.stackexchange.com/questions/427/… for some thought. It's not obvious to me what (if any) firm result they came to and I am not terribly active on the site. $\endgroup$ Commented Jun 2, 2017 at 3:22
  • $\begingroup$ Questions on experimental apparatus tend to be viewed fairly poorly on this site, so one should not necessarily believe the letter of the law, so to speak. $\endgroup$
    – Jon Custer
    Commented Jun 2, 2017 at 15:04
  • $\begingroup$ On the plus side, you're one vote away from re-opening. So you do have some support here, myself included $\endgroup$
    – tpg2114
    Commented Jun 3, 2017 at 1:44
  • $\begingroup$ This is absurd. Of course computational physics is part of physics. $\endgroup$
    – AGML
    Commented Jun 11, 2017 at 16:24

4 Answers 4


I think this question should be on-topic. To the extent that it is a software-recommendations question, it's precisely the kind of recommendation question we want to encourage: it is well-researched, it has a clear statement of goals and of the OP's situation, and clear requirements on the answers, which also make it clear that the answers are expected to be more than simple links to software.

However, I think a closer reading of the question shows that it isn't so much a recommendation question, and it is much closer to asking for the relative advantages and disadvantages between the available tools, which I think puts it squarely on-topic via OP's point 3,

The help and tour pages say that questions about "Experimental technology used in physics or astronomy" are allowed. I would consider molecular dynamics software to be part of the technology that is used in Physics.

Moreover, while comparison questions can often be not-all-that-constructive, here OP provides plenty of criteria to avoid a Gorilla vs Shark situation - and, indeed, it stays distinctly away from the four criteria of badness listed there:

  1. Nobody needs to know the answer to this question.
  2. It’s not nearly specific enough.
  3. It is difficult to learn from these questions.
  4. It drives away experts.

None of these apply here, I think.

  • 2
    $\begingroup$ Re: your second paragraph, if the question isn't really a recommendation question, it should be straightforward to edit it so that it doesn't read like a recommendation question, and that's how I would like to see a question like this handled. $\endgroup$
    – David Z
    Commented Jun 5, 2017 at 3:07

This question should be on topic.

In most mature subdisciplines of computational physics (moleular dynamics, polymer physics, computational fluid dynamics, numerical relativity...) there are only a handful of different codes, and everyone in the field is more or less familiar with all of them. The codes will generally differ practically and algorithmically, and be appropriate for different sorts of problems.

To design and interpret research projects in computational physics it is critical to understand the relative strengths and weaknesses of these specific codes. If such questions are off-topic, than computational physics is off-topic.

By analogy, "should I use GROMACS of LAMMPS for this problem" is not like asking "which brand of hammer should I use to fix my fence?", but rather like asking "should I use a hammer or a screwdriver"? This makes the question very different from asking for a recommendation between, say, vim or emacs.

The argument that the question is basically fine, but simply phrased too much like a request for a recommendation, is empty. First of all, there is quite likely a single most-suitable code for a given problem for reasons that give physical insight. For example one CFD code might handle, for example, highly turbulent flows better than another. Second, understanding which code is best for which circumstance is itself part of computational physics. Third, even if the above were not true, it is very easy to parse the question into a clearly-on-topic form without needing to explicitly modify, let alone to close, it.


Your question is a recommendation question. That's a category that has seen a lot of history going back to the early days of Stack Overflow, and they are generally frowned on throughout the network.

Some of the turning points in the discussion of such things appear as blog posts

and there was endless discussion on (then) meta.stackoverflow (now much has been moved to meta.stackexchange), but I don't want to dig up links.

Physics is a little unusual in allowing some recommendation questions subject to limitations and the willingness of the community to curate them. It is not at all clear that "Which software package should I use?" falls into the exception to the general rule that have been carved out here on Physics.

However, all hope is not lost. Many questions that are off-topic on the site (including this one if the community rules that way in the end) are welcome in the h bar (our general chat room) and you have the requisite rep to be a full participant there. True, you'll have to hope that someone who knows something about the subject looks in on chat, but it is better than nothing.


I think question should be allowed. The answer with an explanation could be very informative. That should be important.


You must log in to answer this question.

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