What Are the Differences Between Theorists and Experimentalists in Physics?

[Elwin.Martin]

Hi, I was wondering if anyone could help me better understand the differences in a Theorist and an Experimentalist. My understanding is something like this:

Theorists typically spend their time attempting to model a given system abstractly, perhaps first in concept alone and then mathematically or perhaps diving straight into a mathematical model.

Experimentalists, while not prohibited from having their own models and ideas about the behavior of a system, spend their time primarily on performing (and improving) experiments to find data. Their efforts would be spent not only on the collection of data, but also on experimental design (kind of like some engineering, maybe?).

I don't know that this is accurate, but this is the impression I've been given. From this (probably false) impression, I've decided that Theory is what I would like to go for, but it would probably be good to know what Theorists actually do and what Experimentalists actually do.

I was also wondering what kind of REUs exist, since I'm still an Undergraduate student and I would like to find research. I would think they would be experimental in nature, since it would be very difficult to be developed to the extent where a student could be working on theoretical problems, but again I am not sure.

Thank you for your time,
Elwin

****Edit: Does anyone have any commentary on t'Hooft's guide to becoming a Theoretical Physicst? (http://www.staff.science.uu.nl/~hooft101/theorist.html) I have lots of free time this and the next semester, and I'd like to use it productively.

 

[Nabeshin]

The experimentalists spend a lot of time constructing experimental apparatus and trying to find all the sources of noise which are contaminating the result they're looking for. A significant portion of time is spent complaining about how poor the apparatus is and how it is likely from the 1950's. Ultimately, you will after a summer be able to show how experimental element A responds to criterion X, Y, and if you have a lot of time, Z.

Theorists, on the other hand, generally spend most of their time banging their heads against a keyboard and trying to track down missing semicolons. A significant portion of time is also spent refilling a mug with coffee. As a result of a summer of work, they will likely have several graphs demonstrating the feasibility of procedure X, but unable to actually apply it to get meaningful results.

On a more serious note, almost all of the 'theory' jobs for an REU will ultimately be computer oriented, and most of your time will be spent wrestling with code. Experimentalists, on the other hand, seem to be more about tinkering and trying many things and just seeing what comes out -- often times conclusions aren't really reached, and at best you have a characterization of whatever process it is you're investigating. That is, you show how the experimental element (a certain circuit, mirror, seismic isolator, etc.) responds to certain conditions (temperature, pressure, current, vibrations), and essentially your analysis stops there.

 

[Monocles]

I'd say that that is a fairly accurate characterization of the difference. If you look hard enough, you can find REUs doing theoretical work (not with computers) as an undergrad, but as you said it usually requires a lot more background than an undergrad typically has. I would conjecture that one of the better ways to go about doing this is to get to know a professor well who works in that area and impress him/her enough to get them to take you as a summer student. At least, this is what worked for me (by doing theoretical work with a professor at my university and then having that professor vouch for me to get another professor to take me), and finding data points on undergrads doing non-computational theoretical work is difficult. You might look into the Caltech SURF program - that is how I ended up spending last summer applying non-commutative geometry to a problem in string theory. You have much more freedom in what you work on than in most REU programs.

 

[Elwin.Martin]

The experimentalists spend a lot of time constructing experimental apparatus and trying to find all the sources of noise which are contaminating the result they're looking for. A significant portion of time is spent complaining about how poor the apparatus is and how it is likely from the 1950's. Ultimately, you will after a summer be able to show how experimental element A responds to criterion X, Y, and if you have a lot of time, Z.

Theorists, on the other hand, generally spend most of their time banging their heads against a keyboard and trying to track down missing semicolons. A significant portion of time is also spent refilling a mug with coffee. As a result of a summer of work, they will likely have several graphs demonstrating the feasibility of procedure X, but unable to actually apply it to get meaningful results.

On a more serious note, almost all of the 'theory' jobs for an REU will ultimately be computer oriented, and most of your time will be spent wrestling with code. Experimentalists, on the other hand, seem to be more about tinkering and trying many things and just seeing what comes out -- often times conclusions aren't really reached, and at best you have a characterization of whatever process it is you're investigating. That is, you show how the experimental element (a certain circuit, mirror, seismic isolator, etc.) responds to certain conditions (temperature, pressure, current, vibrations), and essentially your analysis stops there.

Okay, I wouldn't mind doing some computational work ;O I've got to learn ROOT eventually, right? (Kidding, I don't even know C++ yet...)

Thanks for the reply, I didn't really consider research being inconclusive as an acceptable end. I figured that it was basically failure not to have any results to show for (though I suppose a response to a given set of conditions is results). I have a slightly shallow concern about being able to publish something (eventually), so hopefully I get *something* done, but I suppose it would be alright if it was interesting and I got to learn about the process of research first hand.

 

[Elwin.Martin]

I'd say that that is a fairly accurate characterization of the difference. If you look hard enough, you can find REUs doing theoretical work (not with computers) as an undergrad, but as you said it usually requires a lot more background than an undergrad typically has. I would conjecture that one of the better ways to go about doing this is to get to know a professor well who works in that area and impress him/her enough to get them to take you as a summer student. At least, this is what worked for me (by doing theoretical work with a professor at my university and then having that professor vouch for me to get another professor to take me), and finding data points on undergrads doing non-computational theoretical work is difficult. You might look into the Caltech SURF program - that is how I ended up spending last summer applying non-commutative geometry to a problem in string theory. You have much more freedom in what you work on than in most REU programs.

I wish I was competitive for SURF. The only thing I have going for me would be that my recommendation would come from a Caltech Honors BS and that I'm taking a course in QFT.

Would you mind if I asked you some questions about the geometry you did with string theory applications? I understand that my mathematically knowledge is going to need a significant expansion at some point (I'm not even familiar with the majority of PDE techniques that one typically uses, I just know the stuff I can guess) and I always look for opportunities to ask what I could be learning.

 

[Nano-Passion]

Theorists, on the other hand, generally spend most of their time banging their heads against a keyboard and trying to track down missing semicolons. A significant portion of time is also spent refilling a mug with coffee. As a result of a summer of work, they will likely have several graphs demonstrating the feasibility of procedure X, but unable to actually apply it to get meaningful results.

What type of theory jobs are you talking about here?

 

[Monocles]

I wish I was competitive for SURF. The only thing I have going for me would be that my recommendation would come from a Caltech Honors BS and that I'm taking a course in QFT.

Would you mind if I asked you some questions about the geometry you did with string theory applications? I understand that my mathematically knowledge is going to need a significant expansion at some point (I'm not even familiar with the majority of PDE techniques that one typically uses, I just know the stuff I can guess) and I always look for opportunities to ask what I could be learning.

Arg, the forum just ate my post. The biggest hurdle into getting into SURF is finding a professor there who is willing to work with you, since you need one in order to even apply. But otherwise the acceptance rate is something like 70%. What the SURF website suggests and what I did was to ask a professor at my school who knew me well and recently published a paper with a Caltech professor I was interested in working with to put in a good word for me. That got my foot in the door and then the rest was easy. So, I think it is more likely that you could get in than you think.

What non-commutative geometry is is a mathematical framework invented largely by Alain Connes. The launch pad is Gel'fand duality, which basically states that you can look at the space of continuous complex valued functions over nice topological spaces instead of the topological space itself and not lose any information. This space forms a nice object called a commutative C*-algebra. Then you study the non-commutative C*-algebras, which are not continuous functions over any 'actual' topological space, but we pretend that they are and translate a lot of topological and geometric language into the algebraic world so that we can 'do geometry' and 'do topology' with an algebra.

Most of the applications of NCG as far as I know are in physics - condensed matter, quantum field theory, string theory, and quantum statistical mechanics for the most part. But it also has applications in some fields of math such as algebraic geometry and number theory. Because of that it seems like a great playground for someone like me who cannot decide whether he likes physics or math more.