About theoretical research

  • #26
It shouldn't discourage you. There is plenty of research that doesn't involve much computer work. So if you don't like computers, then you certainly don't have to spend all day long coding. However, saying you don't want to deal with computers AT ALL will close a lot of nice opportunities. So you should definitely learn how to code in a program like sage or matlab.
Isn't the fraction of computer free math/physics vastly, vastly smaller than the computer-ful math/physics research though?

The slice of the pie left for mathematica and blackboards is pretty tiny from what I understand, but I'm really asking a question since I don't know for sure.
 
  • #27
Dr Transport
Science Advisor
Gold Member
2,491
614
Isn't the fraction of computer free math/physics vastly, vastly smaller than the computer-ful math/physics research though?

The slice of the pie left for mathematica and blackboards is pretty tiny from what I understand, but I'm really asking a question since I don't know for sure.


yes, the number of analytically solvable problems in theoretical physics is pretty much dried up. to get an answer, you're gonna have to lay down the theory and then code it up.
 
  • Like
Likes Crass_Oscillator
  • #28
yes, the number of analytically solvable problems in theoretical physics is pretty much dried up. to get an answer, you're gonna have to lay down the theory and then code it up.
More or less exactly what I thought.

OP if you want to avoid computers, you might be able to pull it off, but you'll have to be unusually brilliant I think. Consider determining why you're not fond of computers; I always liked them but once had dreams of being a pen and paper theorist. Learning to like them made my life easier. Also there's a spectrum of research involving computers; I do a lot of pure computer stuff and less pure math/algorithms. Other individuals who do a lot of programming do a lot of math too.
 
  • #29
radium
Science Advisor
Education Advisor
763
242
I am in CMT and can say that there are definitely still problems that can be done analytically and others that can be done partially so. Many of the papers on topological states of matter are entirely analytical and I believe people may still be engineering exactly solvable models in lower dimensions. You do need to do treat things perturbatively a lot of the time (in transport calculations with disorder for example), but you can still get results.
Entanglement is another area which has also been of interest to people in HET where people have been doing analytic calculations, many using the Ryu-Takayanagi formula from holography.
 
  • #30
13
2
OP if you want to avoid computers, you might be able to pull it off, but you'll have to be unusually brilliant I think. Consider determining why you're not fond of computers; I always liked them but once had dreams of being a pen and paper theorist. Learning to like them made my life easier. Also there's a spectrum of research involving computers; I do a lot of pure computer stuff and less pure math/algorithms. Other individuals who do a lot of programming do a lot of math too.

I don't think of myself as a particulary brilliant person, but it is not my intention to completely avoid computers.
I am ready to learn to like them if it's necessary, I simply don't want to spend the majority of my time behind a screen. As radium suggests, though, it looks like it will depend a lot on the field I'll eventually choose...
 
  • #31
20
2
I think this thread is being really useful. Just let me point out two things that might be of your interest, Punck. First, acording to a paper I found at scholar google, the main programming languages at ATLAS (LHC) are currently: 1. C++, 2. Python, 3. FORTRAN. (I think you will have to trust my memory on this, I don't remember where I read it). You may want to learn any of those. FORTRAN is the oldest, pretty simple. Python is very simple to learn as well, you have millions of libraries all around and I think that is, with R, one of the most used languages in data mining and this kind of stuff. And finally C++ is more complicated but one of the fastest.
And the other thing you may want to know. There are in fact some libraries of programming called Computer Algebra System, that are sometimes used for symbolic programing. For example, if you want to do an analytical integral, they might be able to do it
 
  • #32
13
2
Hello Pablo, thank you for your advice.
I have already started learning some Python and a while ago I taught myself the basis of C ++. Again, as above, I don't love computers, but being aware of their importance I do brush up my lousy programming skills from time to time.

Punck

This isn't meant as a subtle insult, is it? ;)
 
  • #33
20
2
This isn't meant as a subtle insult, is it? ;)
Well, no it isn't, hehehe. Just typo, sorry
 
  • #34
143
36
I am working in computational biology right now. The work is done almost exclusively my physicists or mathematicians.

People basically sit behind their computer most of the day. Writing code or reading papers with equations all over the place.
Then sometimes you write your own notes/paper. You need to be really smart to make a breakthrough in the math. It is mostly about applying stuff thought out by the truly brilliant to new problems. For the large part we use equations that have been thought of a long time ago, but only now we can use on biological systems. Either because how the field has developed, how much cheaper computational time has become, or how stuff can be compared to practical experiments that are possible or easy today, but were not in the past.

Every few days you spend an hour or so listening to someone explaining something difficult, or you explaining something difficult to someone else.
Or you discuss the problems and possible solutions with colleagues or supervisors/professors.

Sitting behind a computer sucks. You all sit there next to each other, barely talking. People lighten up in lunch breaks, though.

But practical work has it's own problems. You are sometimes in a basement of a building, all alone, repeating the same measurement over and over. You basically sit there and wait while it measures. Or you set it up and walk away to read papers.
Or, you have to wait for centrifugation steps. Sometimes waiting steps are too short for you to do something else, but long enough to get really annoying.
I have had professors tell me this is the exact reason why they choose to do theoretical work. They hate spending so much of their time waiting or doing the same mundane things over and over.

Getting real results is really slow in any field. Big areas of disappointment and short moments of euphoria. It is hard.
 
  • #35
62
6
Sitting behind a computer sucks. You all sit there next to each other, barely talking. People lighten up in lunch breaks, though.

I can't really say that this reflects my own experience working on computer stuff. (Co-developing a new parallel PiC (particle-in-cell, plasma) program)
We easily spend a third of the 'coding time (when we are productive) discussing problems, algorithms, alternative algorithms, test scenarious and results, coding practices.
I generally think that two developers that stops often and discusses problems with each other get much more done than if they each sit quietly coding alone.
 
  • #36
143
36
I guess there is some hyperbole and maybe it's just more true for me currently because of circumstances.
 
  • #37
Biology tends towards being incredibly applied, as there is no strong theoretical foundation on which to base one's algorithms. Brute force and intuition currently outmatch cleverness and sophistication as a general rule.

However there are many other fields where the theoretical foundation is surer and so more pencil and paper mathematics accompanies the raw computer stuff.
 

Related Threads on About theoretical research

  • Last Post
Replies
2
Views
2K
Replies
1
Views
1K
Replies
2
Views
429
Replies
2
Views
1K
Replies
3
Views
1K
Replies
3
Views
6K
  • Last Post
Replies
7
Views
2K
Replies
3
Views
1K
Replies
12
Views
6K
Top