I'm a few months from a Ph.D in Physics (theory)...

I find it rather curious that you've only learnt Fortran while nearly having obtained a Ph.D.
2nd year of undergrad and i've got Python, Matlab and a bit of C++ under by belt as part of the curriculum.

Matlab and Python not FORTRAN, I never heard that this is used in industry

Attached: 1526920892665.gif (640x443, 3.88M)

Learn something about architecture
Jow Forums-science.wikia.com/wiki/Computer_Science_and_Engineering

>engineers doing heavy simulations also use it
Aerospacefag confirming this, all our bespoke cfd / csm tools are written in Fortran

which University?

while technically true not practically true
If you're a PhD and get a research position, nobody's going to care if you produce shit code--it's the expectation. Granted if you actually know your shit about computing and understand the physical limitations of memory and binary encodings (i.e. IEEE floating point), that'll go a long way to ensure that you're not a total shit.

saving mum

>because the curves diverge
If not heat death, then what?

well I am the guy with MSc. in EE, in fact I did my MSc. in developing FPU units kek, I assure you most developers don't and don't need to understand IEEE 754 standard at all

>I'm a few months from a Ph.D in Physics (theory). I hate physics now
the absolute state of academia.

they certainly don't need to understand the relationship between the exponent and the mantissa, but they at least need to understand that there's a lack of precision that can eventually result in calculations approaching NaN

Don't go into webdev/appdev and be a codemonkey. Those are for kids and dropouts like me.

You'll be very well suited for Otherwise, get into Data Science/Machine Learning.

Just join a fintec user, they lap up maths-phys and comp-phys PhDs so long as you're not stupid.

Instead of worrying about programming languages work on your soft skills; as a physics PhD myself those I knew that failed to earn bank were the ones with zero social skills and too much autism to accept sufficient answers over correct answers.

It is quite normal at that stage. And it has been like that for ages.

It's the absolute state of physics PhDs, departments, and the whole subject culture. They take a year longer than (almost) anyone will fund you for, they usually have little to no definition, an enormous problem space, typically a very small viable solution space (which you only find out with a lot of research), and the field holds itself to a horrendously high standard if you want to publish in good journals and/or are in a good group. Your only chance of having a 'normal' PhD experience is being an experimentalist, as at least once they get their kit working they can pump out a few papers quickly.

I've worked in both maths and engineering departments after doing my PhD and postdoc in maths-phys, all at similar (top) quality institutions, and physics is just another ballgame. PhDs are treated badly everywhere, but in physics it's an order of magnitude worse. Even the academics have a rather bad time of it. And it's generally their own doing. Even the PhDs that stay in physics hate the field by the end of their PhD (or usually within 2 years of starting).

With a physics degree (especially experimental physics) you have a good change of getting a good job in industry. Some universities seem to resent this option and seem to want to actively block our attempts to recruit PhD graduates in order to keep them as badly paid post docs.

You don't need to learn shit: just apply to some company that hires any idiot with a Ph.D., like Google for instance. Seriously.

Can Dr. Anons ITT comment on (EE)CS PhD experience?
I'm set on getting a Master's in Robotics/ML but contemplating whether I should go all the way.

Yes, very easily. With a physics PhD of any sort it's very easy to get a job across just about every field.

One thing I noticed when I was still in academe is that some groups have the very very bad practice of trying to hold on to their researchers indefinitely, sometimes through underhand methods.

Despite my general complains I came from an exceptional research group, both in terms of academic capability, and in terms of mentality. We had an iron rule that nobody could do a postdoc in the group (or in the institution) after the PhD, as it hurts long term progression chances too much; the general point was everybody who got a PhD from us should be on track for an early career professorship (n.b. if you're american UK professors are nothing like US professors, it's a tiny minority of academics).

Others, and this includes first hand knowledge, do everything they can to bind their good PhDs to the group as paper machines, to push group credibility, and then drop them. I've know of group leaders and professors systematically writing poor references and undermining their students in order to damage their career prospects. I've even known professors to fabricate negative rumours at conferences and the like to do this (not limited to physics btw). So it's not so much universities as the research groups and culture therein, but it's horrific never-the-less. In the same vein, I know research groups that will refuse to allow their PhDs to spend travel money on anything outside of the narrow academic context, e.g. poster showcases for industry. Again, I had the luck of a research group that gave you tonnes of travel money and forced you to present to industry, MoD, policy makers, and everyone else, but, as I've found out since, it was a real exception.

Sorry, whilst I did do some scientific computing and modelling I was basically a theoretical physicist, and, to a limited extent, a systems engineer as I wrote a few recommendations papers in that domain pertaining to emerging technologies.

I have had friends in Comp Sci at PhD level and above though, and have some observations: For the love of God be careful when picking your supervisor, comp sci seems to have very low quality supervision with very limited contact hours. The Chinese seem to be particularly bad at the institutions I've been working at.

Make sure you know your maths, if your maths is weak, do more maths. You will need it, even if you don't think you need it, you do. Especially if you do ML.

If you want to do anything that requires implementing to prove your hypothesis make sure the group has funding/means for it. I've met a few roboticists who, as a last step, need to demonstrate capability outside of simulation, and suddenly realise there's no funding for the equipment they need.

And, whilst everyone hates it, makes sure you plan your project well. Set deadlines, use project management tools, make Gantt Charts and find critical paths, keep a centralised record of knowledge, in particular recording your decisions and assumptions throughout the project. It is very easy to get lost in a PhD, it is very easy to lose motivation, it is very easy to have untracked and unnoticed project drift (damning if your PhD has a defined remit and external funding). Students hate planning, but it really really helps.

Lastly, manage your supervisors. Academics have no managerial training, they are not going to keep you (and the n other PhDs they have) on track, they are not going to record information you give them sensibly, they probably won't remember what they said to you last time you spoke. It is your job, from day one, to manage them, and to get what you need out of them. Being demanding is okay, making contacts and collaborations is vital.