Chemists, lab managers, etc. in industry: which programming language, if any, do you get the most use out of in your job?
I work as a lab tech at a small company and we live and die by our IT guys when it comes to the data entry/processing part of the job. I find myself wishing I had a little BASIC knowledge so I could have at least an inkling of how our macros work (when they work). What kind of fundamental programming skills are good for someone looking at a lifetime working in labs?
>>9034956
All of them. Currently working on a project that uses C, C++, C#, python, perl, matlab, mathematica, and a hint of fortran. For the love of god don't do this. I want to die.
Check out python. I also highly recommend fucking around with openbabel for chemistry shit.
Isn't python the dominant language for academic and scientific stuff?
There is always C/C++ But that's mostly for libraries. They often have python bindings.
>>9035260
Yeah, but it's slow.
>>9035263
CPU time is cheaper than your time, bro.
>>9035272
Very true. Python makes it easy to get an experiment off the ground. I've been working with ipython notebooks and they're great
>>9035303
>>9035263
It may be fast with high level calls like those for numpy (Matlab syle operations and arrays)or systems like theano and tensorflow (these one mainly for neural networks).
You can compile cpu intensive parts with numba a sort of jit compiler using llvm targeting both cpu and cuda and even compile python via c or c++ with cython where you can use c/c++ types with python syntax for full speedups
There's plenty of way to speed up python with little effort and a few profiling.
>>9035263
Thou art quite a smarten ladd mi laddie,.
>>9035263
python script kiddies amirite
>>9034956
Aerospace industry, C, C++, Fortan.
>>9035391
I'm studying AI, so most of the time the optimization work is already cut out for me. Most models have some decent implementation in scikit, or if its a neural net then there are more than enough libraries for that.
>artificial intelligence
It sort of feels like child's play with neural nets. Throw shit to the wall, see if it sticks.
That being said, good performance is ultimately up to the programmer. Good design, and proper algorithms will yield the greatest improvements in performance. Compilation/run-time speed-ups are the icing on the cake.
When it comes to programming skills, most languages are the same. Some make some things easier for you. I think Python does this.
>>9035500
I agree, and python correspond to what would mean and aimed originally B.A.S.I.C : Beginner's All-purpose Symbolic Instruction Code and it's great in that purpose easy enough but still powerful and being better than basic in the two domains
Python can suck a dick.
R for life.