Brought up by the flag and comments on How to determine whether a large container is air-tight?".


I want to address question in which the OP has run into a roadbump in build/calibrating/operating a piece of experimental apparatus, and the question is well defined in terms of fixing the problem.

Note that this particular question is not a great example because the OP is not building an experiment, but I would ask that the case for experimentaly realted question be examined separately from the fate of this particular question.

  • A user has flagged suggesting that this particular case is an engineering problem and asks for a open ended list of suggestions.

  • David Z comments that

    I do think it's not constructive, as it's just asking for a list, not asking a specifically defined question.

I am going to argue below that this is exactly the kind of question which the Stack Excahnge engine is good at answering and should be explictly allowed.


We should also note that attmepts to open a experimentally specific Stack have not been successful (you'll see that I argued that most such question are on-topic for Physics.SE).

Vaguely related: Experimental technology questions: on topic?

share
    
Ah, I forgot about that last linked question. In that case I retract my statements about the off-topicness of the question in question. –  David Z Jan 3 '13 at 22:00
    
By the way, what I meant by my comment which you quoted is that the question is a "list question." The problem is not that there can be more than one answer, which is not a problem itself. It's that the question just asks "Suggestions, anyone?" without providing a clearly defined framework of what to suggest. That is something that is specific to this one question, not generally applicable to all experimental problem questions, and so perhaps we should leave that aspect of the question out of this discussion. –  David Z Jan 3 '13 at 22:07
3  
I deleted some irrelevant comments. –  David Z Jan 4 '13 at 0:22

3 Answers 3

I want to follow two threads in arguing that such questions are very well suited to Physics.SE: firstly that stack exchange is well suited to these kinds of questions and secondly that these are physics questions.

  1. Prototypical questions that are well suited to our engine here have a small or medium number of distinct answer that are not wrong and can be well evaluated by users who are familiar with the question domain. Such questions can be either outright objective or "good subjective".

    Yes, there is going to be a list of answers. There are lists of answers to questions on Stack Overflow and to most of those on the main site (in fact many of our most highly voted questions have 10 or more answers that take several approaches to answering the problem), but these list are characterized by a domain expert being able to say to each answer "this is correct" or "this is incorrect", and any subjectivity that may enter is in terms of which answer is clearer or easier or more scalable or more pedogoically useful.

    Managing, maintaining and running experimental apparatus constitutes a specific domain of knowledge in which experts will know which answer are good, which might work and which are nonsense.

  2. Experiment is, obviously, a pillar of the scientific inquiry co-equal with theory and phenomenolgy, and a substantial fraction of working scientists would describe themselves as an "experimenter" rather than a "theorist" or a "phenomenlogist".

    And the actual work of an experimenter includes a lot of things from the design, construction, calibration, operation of experiments though data acquisition and analysis as well as writing up the results. In my personal experience a scientist in his or her early career will spend a substantial fraction of his or her time tinkering with and swearing over some piece of apparatus that almost works or simple isn't working at all.

    These are problems encountered and solved by physicists, and because they revolve around experimental apparatus they are specific to the profession or to a small set of related professions. (Yes, I recognize that the leak test question is rather broader than that.)

    Finally, to exclude the kinds of questions that experimental grad-students spend a lot of time on from consideration here send a message that experiment is not valuable to the physics community. Do we really want that?

share
1  
Playing Devil's advocate here -- If experimental design is a pillar co-equal to theory, then should computational work be a co-equal as well? It is a numerical experiment. So the argument that experiment questions, including techniques and apparatuses, would be on topic could then extend to say that computational design, techniques, and apparatuses would be on topic. But the FAQ rules them out -- is this contradictory? –  tpg2114 Jan 5 '13 at 17:44
    
@tpg2114 Well, I didn't say "experimental design" but experiment and meant for the word to be taken broadly: things done by self-identified "experimenters". My personal experience with computational physicists is with lQCD types, and I claim that they are a breed of theorist. I know they talk about running "experiments" and term their results "data" (I can even see the distinction they wish to make), but I don't buy that. Anecdotably, a lQCD buddy in grad-school had a giddy revelation when he first settled an argument by going into the lab. –  dmckee Jan 5 '13 at 17:55
1  
To take the question seriously, the pratice of computation physics has more in common with other computational science then with the practice of other physics, and there is a seperate stack for them to go to. On the other hand attempts to get a stack for broadly defined experimental work have gone no where. –  dmckee Jan 5 '13 at 17:56

I agree that questions about building experiments, if specific to our subject, should be allowed.

That is, until they are about physics. A question about "what type of concrete should I use to build my atom smasher" is not about physics.

The key point here is this.

Questions that work are questions:

  • which are best asked to a physicist (an experimental physicist in this case, but not a engineer or a mathematician)

  • whose answers are best judged by a physicist

If we fail to stick to this (obvious) principle, the consequences are going to be poor answers being given and bad answers being up voted.

Furthermore the problem will struggle to self correct because it requires competence to determine poor answers. If someone answers with the wrong kind of concrete, to the example above, who would know better, or that we have a problem at all?

That said, that particular question is poor on many levels.

share
    
OK. Can I ask how in particular the question in question is bad with an eye to developing a set of criteria or a decision procedure. I'm willing to start with the context is clearly not a physics one even if the problem actually comes up in expermient. What else? –  dmckee Jan 4 '13 at 3:29
    
The question seems to ask for speculation and seems to be of the "my code is broken, fix it" sort which doesn't really seem useful. I think that both issues can be solved with editing... –  Sklivvz Jan 4 '13 at 10:11

I exactly agree with Dmckee, questions that deal with issues specific to smaller or larger physics experiments or the analysis of experimental physics data should be in principle allowed and encouraged here. They are not just engineering questions! After all, physics SE is for the whole subject of physics; experimental, phenomenological, and theoretical question are equally good.

And question with a countably finite number of answers, whose content can be evaluated if it is physically or scientifically right, are certainly not a problem either. The fact that a question has multiple answers (which have all there merit and right to be there) is not a reason to close it!

share
    
I don't get it. Why is it, that there were two downvotes here, but not on dmckee's? –  Dimensio1n0 Sep 17 '13 at 13:13

You must log in to answer this question.

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