Re: [O] General advice beyond Org

[Marco]

Hey fellow orgers,

I started as post-doc almost 10 years ago, I made a point in having
linux on my desktop and only use software I see fit - and know - such
R, and LaTeX to write papers. The occasional word/excel document I
opened with libreoffice/wps office or more frequently remotely on a
windows server / on a virtualbox, same with the occasional
windows-only piece of software I needed to use.
After 8-9 years of mockery, my whole departement is now switching to
LiNuX/LaTeX/python/R/jupyter, with the exception of a few specialized
softwares which are windows-only and will run on selected
machines/servers. Point is, also the secretary has to migrate, she was
the greatest advocate against the change.
No luck yet with emacs + org with my colleagues, but hey, one step at a time ;)
If your research and daily work involves programming/modelling I
actually don't see any point in staying on windows. Unices have LOTS
of advantages. The best general-purpose research tools nowadays are
free/libre software (since 3D FEM was named: openfoam, moose,
opencascade/salome, code_aster just to name a few... ), that is, if
you are developing something and not just running models, their steep
learning curve pays off immediately. If it's just a matter of running
specific models maybe there are other specialized softwares which may
be more efficient (e.g., avizo for 3D microtomography, eclipse /
petrel for geological modelling and reservoir engineering). In the
realm of statistics/machine learning/AI it's all linux-based free
software (think about keras + tensorflow with R and Python, but loads
and loads of other similar stuff out there).
So in my opinion in a research environment there are 2 kinds of
activities: computational/programming stuff, where unix and free
software have a clear advantage both on desktops and of course on HPC
(of course it depends how it looks like in your domain AND in your
research group), and on the other side wiriting documents and
interacting with others (reports, proposals, data exchange, papers).
You should have 100% control of the first category, but still have to
factor in cooperation with colleagues (i.e.: if everyoen is working on
matlab and you are supposed to extend and re-use their developments...
probably you should too. At least try sci-lab and octave); but you can
definitely compromise on the second.
Cheers and good luck,
domanov