I recently noticed this question:

What is the format for "local catalog" files used by JSkyCat?

Apparently it was migrated from Astronomy SE, and was on topic there according to the comments. But on the other hand, Physics has had a long-standing policy of not allowing software questions like this one. (see also this) But as dmckee pointed out, there are some software questions about tools very specific to physics which would not get any useful reception on another site.

Should we adjust our policy to keep this question (and others like it) about specialized software questions, or just close it in accordance with the existing policy? Or should we start allowing software-based questions more generally, as long as they are related to physics?

UPDATE: Here is an example of the kind of question I would like us to decide about:

Visualization of dynamics of a scale-free network of Ising spins

Update to the update: actually it's not such a great example, because this is really a software recommendation question, and recommendation questions of all sorts are generally off topic (except for a particular set of book recommendations). But it's still worth discussing.

I was literally about to post the same thing as your edit... – tpg2114 Dec 27 '12 at 2:55
Good that we're thinking along the same lines :-) Feel free to put in an answer addressing that example. – David Z Dec 27 '12 at 2:56
Here is a good example of a question that (IMHO) should not have been closed according to the current or proposed policy, because it is clearly about the physics rather than the programming: physics.stackexchange.com/questions/25346/… – Nathaniel Dec 29 '12 at 9:46
up vote 9 down vote accepted

I exactly agree with Dmckee, who says that questions about spezialized software used and of interest only for physicists should be kept and allowed, and therefore the Physics SE policy adjusted a little bit.

In addition to the Monte Carlo simulation example of such a spezialised program, Prof Strassler explains what the Black Hat program does, which is used to calculate the standard model background in different evaluations of LHC data. It takes advantage of the new twistor methods to calculate amplitudes.

I am sure there could be found other exmples of such important spezialized programs needed by physicists in astronomy and other subfields of physics generally. It would not make sense to ask questions about these at a random programming/computational site, where people are by no means familiar (and probably not strongly interested in) with the physics context wherein such software is applied. So such questions about specialized software exlusively applied by physicists and no one else, should be allowed here.

As things stand discussion of the twistor methods as physics belongs here and of implementing them in software on SciComp.SE and there isn't a well defined place for questions about existing software making use of these methods. I take it that you are suggesting that those currently unassigned question belong here? That is certainly a position I would be comfortable with. – dmckee Dec 25 '12 at 19:00
Hi @dmckee, yep that is what I wanted to say ... – Dilaton Dec 25 '12 at 19:17

We have accepted questions relating to the meaning of software objects in the particle physics Monte Carlo generator Pythia, and my feeling at the time was that this was a tool by- for- and of-physicists and of no interest to anyone else.

So the question here becomes "Is this JSkyCat thing of interest or use to anyone other than astronomers?"

If the answer is "No." then it is certainly a question for us. If the answer is yes, then we have to think.

So what is it, any way?

Hm... you raise an interesting point. I edited the question in response. – David Z Dec 24 '12 at 23:20
For what its worth, no I don't think JSkyCat would be of interest to anyone outside of astronomy. – ehneilsen Dec 25 '12 at 20:45

I think it's a good idea to allow these kinds of questions, as long as they are local enough to physics. A few people have said that it is "off topic" and not relevant. But since this isn't an only theoretical physics site, I think questions relevant to the acquisition and storage of data are relevant. When most data analysis is performed on the computer, some questions which are 'software related' are bound to arise, and as dmckee pointed out, no one else can really answer them.

Citing an example from experience... Astronomers tend to use a file format called FITS. Now there exists extensive documentation about FITS files, but unfortunately not about one tiny aspect of FITS files local to radio astronomy, and I had to extensively hunt through the web to figure it out. If a question like that is asked now, I don't think it should be closed. Frankly, I don't think anyone else apart from astronomers will be able to answer it accurately.

It could be argued that questions like this are very specific to one branch of physics... but as a contrary, plenty of theoretical physics questions are relevant to only that branch of physics (For example, this question).

I'm in favour of letting a (very narrow) range of software-related questions be asked.

Good astronomy example, +1 – Dilaton Dec 25 '12 at 14:04
"But since this isn't an only theoretical physics site" Yes, in the "big science" experimental fields specialized (and sometimes confusing) software tools are a big part of what we do. – dmckee Dec 25 '12 at 19:02
@dmckee - Exactly! Like I said, we can't allow ourselves to become pseudo-StackOverflow in the programming content we answer, but we should certainly tolerate some. Specifically, the some that won't/can't be answered elsewhere. – Kitchi Dec 25 '12 at 19:55

I personally feel that we should not allow any software questions -- I feel that our site ought to be mainly conceptual. So I am completely against Or should we start allowing software-based questions more generally, as long as they are related to physics?. Questions about software that is just related (and not confined to) physics are better asked on SU.

That being said, I don't mind if we allow software qs on software that is only of use to physicists/astronomers/students of the two disciplines. But we should be careful about these, many such qs can be too localized. And they may detract from the site if there are too many of them.


It seems to me that the spirit of the original meta question you linked to wasn't so much "should we allow software questions?" as "should we allow questions that are only about computation, without any specific physics content?" The answer to that question was "no", but that doesn't mean we need a blanket "no computation questions" policy.

Currently the FAQ says

If your question is about writing a simulation or other program, and it concerns the algorithms, execution, or presentation of the results (rather than the underlying physics), it is probably more appropriate at Computational Science.

I think this is a reasonable policy, but it should be interpreted such that questions are on topic if they're about software but aren't just concerned with algorithms, execution, or presentation of results.

In my view the JSkyCat question should be allowed because although it isn't strictly about the physics as such, it's asking for knowledge that's quite specific to physicists (or rather astronomers) and for me that's enough to make it not just about algorithms, execution or presentation.

@CrazyBuddy it's asking about turbines that are coupled because of the flow of gas between them. Gas is a fluid :) – Nathaniel Jan 16 '13 at 9:58

Of course, we shouldn't allow such questions. While, I was revising Astronomy imports, I arrived at that one. I thought it was off-topic (and indeed, the flag was mine). So, I voted to close.

We don't need such questions. Because they're dealing with some kinda measuring data-collection-like stuffs that we don't like. I think most of the users also won't like it.

We could close it... I don't think a prepoll would be of so much importance :-)

I really agree with dmckee's answer. Questions which are more likely supporting Physics could be allowed. But, not the question which you've pointed out. It seems like It'd be understood and responded only by some 10s or 20s out of 16k users. I think we don't need such questions (like the above)...

The "of course" in your answer was a bit misleading to me (as the question is "Should we allow..."). But I completely agree that this question in off-topic for the current Physics.SE (and we should keep it off-topic imho) – Bernhard Dec 24 '12 at 12:02
@Bernhard: Err... Sorry for that, Benhard. Revised now :-) – Waffle's Crazy Peanut Dec 24 '12 at 12:03
So questions about measurement devices are also off limits in your opinion? If I wanted to ask how Laser Doppler Velocimitry is different from Planar Laser-Induced Fluorescence for imaging experimental flow fields and/or how they work, you would consider that off topic? I agree software questions are off topic, but I wouldn't go as far as to say questions dealing with measurement instruments or things dealing with data collection are. – tpg2114 Dec 24 '12 at 12:13
@tpg I think these kind of questions is really borderline. Questions should be really about the physics of the measurement technique then. – Bernhard Dec 24 '12 at 12:23
@Bernhard I agree, "How do they work" is probably too basic. My issue is only with the sentence "Because they're dealing with some kinda measuring instruments, data-based things - stuffs that we don't do" I don't think it's clear what was meant as I can think of physics-of-measurement system questions that would be on-topic. – tpg2114 Dec 24 '12 at 12:35
Only because this answer will show up in searches days/weeks/years from now and the original context may not be known. So answers should be clear and on-point so there's no misunderstandings later. – tpg2114 Dec 24 '12 at 13:20
Crazy Buddy, I edited the question in response to some of the points dmckee brought up, so see if you would like to change your answer in response. – David Z Dec 24 '12 at 23:21
There is, I think, a difference between "on topic" and "questions we like," which are both different from "useful questions." Some of it comes down to whether you value the site primarily for entertainment or utility. – ehneilsen Dec 25 '12 at 20:51

Most software questions are better off on StackOverflow and other related sites. The folks over there are far more knowledgeable about such things, even if the software employed is specialized. Migrating such questions would actually be more helpful to the user. Maybe asking the OP before moving the question would be the way to go. This gives the OP a heads-up and a chance for clarification if need be.


You must log in to answer this question.

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