How useful is a Physics degree?

I graduated from a small liberal arts school with a degree in Physics (and a minor in CS). I actually would have done Computer Science, but the university's program didn't teach the things I was interested in -- but let me be clear: computers were my primary passion.

I got a software engineering job offer after graduation, but I decided to go to grad school instead and I'm now finishing a masters in Computer Science. When I started, I was worried that not having a BS in Computer Science would handicap me compared to other students, but it didn't seem to (or at least grad school is hard enough for everyone that I didn't stand out).

My work and my interests are in Free Software, and what I'm finding out in applying for jobs is that companies (especially those that deal with Free Software) care very much about your contributions to Free Software. It's obviously easier to get a handle on someone's skills if their entire portfolio of work is available in public git repositories.

In working with different projects, I've also gotten to know developers who work on the teams that I'm applying for jobs on. I feel like this gives an extra advantage that's not easily available in other fields.

The summary of this is that the degree doesn't matter so much. What matters (at least in the world of Free Software) is what have you contributed and who have you worked with.

Another data point: a number of really good software developers at Red Hat have non-computer degrees or no degrees at all.

Owen Taylor (GNOME) - Physics

Jerome Glisse (AMD driver) - PhD in Biology

Matthew Garret (Kernel, power management) - PhD in Biology

Ben Skeggs (Nouveau driver) - no degree

Really tho--if you have a passion for CS, but studied Physics, you have the capacity to excel. I fell in love with the ordered-ness of computers and self-taught myself programming in my teens. I went to college and studied Physics knowing I would likely not continue as a career physicist. During college I found that between my basic programming knowledge and physics requirements I had covered a large portion of the CompSci major (except for the upper division classes). I tested out of some and then took the others to end up with CompSci + Physics degrees.

Since graduating, I've benefited from my CompSci degree in areas like data structures, runtime complexity and parsing--but the rest of my CompSci skills were self-taught before or afterwards ("the Internet" was not in my CompSci curriculum).

I believe I've benefited more from my Physics training. Specifically--the Socratic method of looking at root causes in a systematic and problem-simplifying manner. And I've never been daunted by a hard problem or one that needs theoretical analysis for an elegant solution.

Given all of the above, some of the best programmers I've met have no degrees--so passion and willingness to self-learn trump all.

Hell yes. It will mean that clueless recruitment agencies won't even bother contacting you or doing anything for you, because they don't see you as suitable for anything related to IT. Whether that's a good or bad thing is questionable. It does however mean that getting the foot in the door, i.e. getting into a position that requires a Bachelor in IT is a lot more difficult, in them not wanting to take any risks. The only thing you need to do is prove to them that there is no risk.

Between casual IT work (Sendmail, Exchange, Adobe, Linux/Mac OS, Windows Server and end-user support skills) next to my degree and my Physics degree, I was able to pick up enough skills to get a job. It helped that my Physics degree forced me to learn all of Assembly, C, LabVIEW, LaTeX, Mathematica, MATLAB, Python (Project Euler, actually, is the thing that started forcing me to really know Python) and to a lesser extent Bash and Scheme.

Being able to put ten languages on top of Perl, PHP, JavaScript and others I knew already or know outside of my degree, probably really helped in getting a position programming in a niche language. Given the ten programming languages you're likely to learn in a Physics degree (with a bit of interest) and considering that's likely to be more than you'll get from a Bachelor in IT or maybe even computer science, you're in luck :)

So, in a nutshell, and riffing of another comment here, learn version control and unit testing and know at least one programming language like Python or JavaScript inside out and you're pretty good. Brush up on your design skills and you're done - for now :)

If you learn C/C++ programming, you'll be able to get a job at an investment bank and won't have to worry about money. If that's something that attracts you, study (and write code) in a field where Monte Carlo simulations are used, like lattice gauge theory, which is a lot of fun and will help you understand QFTs. Here's some links to get started:

http://latticeguy.net/mypubs/pubs.html

Some C code:

http://thy.phy.bnl.gov/~creutz/z2/

If you want to do something like a web startup and learn Ruby on Rails, then Physics won't help you much. You'll just be on equal footing with CS guys, if you put in the time.

PS: Most physicist I know are absolutely terrible software engineers. If you're going to write larger programs, please read some books about how to write larger programs!