@tush As much as I would love to read many people reporting their relationship with Wolfram Mathematica, at most this will account to a nice collection of anecdotes. What is real question you want to answer? If it is about strategically choosing the programming language with better job opportunities, I think that we can all agree that Mathematica is not it.

@tush Use a job search engine and type Mathematica in the search field?

That said, I have learned more Mathematica by trying to answer questions here (and trying to understand the best answers) than by my work.

@tush apply at Wolfram Research.

@CarlLange, is there anything specific about the license agreement to puts your clients off - is it too expensive or something else?

@CarlLange Even in the research world with readily available site licenses using Mathematica can (unfortunately) be a non-starter. My research is entirely code-driven and much of what I do involves writing libraries that can be used by other researchers (in principle if not always in practice), but I had to switch to python & reimplement tons of things in the process to get buy in.

I started out writing almost 100% Mathematica code since the suite of tools was so powerful that it allowed to do science without getting bogged down writing a bunch of helper code. My boss at the time was even impressed by how much it was possible to do with what they figured was better known as a tool for college kids to get help with their calculus homework. But no one I was working with had the time or energy to learn a language that had such a niche user base

This was compounded by the fact that learning to write efficient Mathematica programs takes a long time relative to other languages, just because it is so easy to write inefficient ones. Add in the fact that Mathematica expertise buys you next to nothing when looking for a software-oriented job and it was just a non-starter. At the end of the day I wrote tens of thousands of lines of code reimplementing things built in to Mathematica just to be able to escape the walled garden

It's not optimal, but you've gotta meet your users where they are

@user21 I remember people commenting that Wolfram Research is not a particularly nice employer, in terms of pay and how they treat employees. But I don't know first hand.

@b3m2a1 I acknowledge you describe a pragmatic reality, but levelling down is a shame. There must be a limit to that, otherwise you end up just Pointing and Grunting or even worse, using MS Excel.

😉

@rhermans, I work for WRI as a consultant and have not been treated badly and I am also happy with my compensation.

@user21 That is very nice to know. Thanks. I think the information quoted at the time was both some particular person's experience and Glassdoor

@user10478 You need to make a Table of your results. Even better, just make a Table when you find your results.

res = Table[suc /. NSolve[1 - CDF[HypergeometricDistribution[8, suc, tot], 0] == 0.72, suc][[1]], {tot, 60, 75}]

ListPlot[res]

if you want the x-axis values correct:

ListPlot[res, DataRange -> {60, 75}]

or include them:

res = Table[{tot, suc /. NSolve[1 - CDF[HypergeometricDistribution[8, suc, tot], 0] == 0.72, suc][[1]]}, {tot, 60, 75}]

Thank you! This was the trick I needed.

@rhermans Lots of science done with Excel & Fortran. Usually you can convince the younger cohorts to use python or C++ but even then a lot of the actual ground level analysis ends up in Excel sheets or on scraps of paper

@b3m2a1 No doubt that it happens, I see it all the time myself. However, I can not have a good opinion about that choice. Unless people are doing accounting, I would say that having excel as the weapon of choice is the opposite of being resourceful.

at school, Mathematica is used. Mainly in Physics departments. I've taken a course where we used only Mathematica for everything. I agree, outside research places where they do basic science, commercial CAS software such as Mathematica and Maple (its main competitor) are not used much because in the real world, very few need to do computer algebra and hence less need for such software.

@tush I work at ZF (automotive industry). I introduced WL via a EnterpriseCDF tool which is now widely used in my particular group. However, this is a heavily MATLAB dominated this industry.

@gdelfino Yes, Matlab is entrenched in auto and aerospace, because of simulink. They design everything there with simulink., Matlab is also very strong in control and DSP. But very weak in computer algebra. Python is trying to overtake Matlab, but this will not happen. It is simulink which is keeping Matlab popular in industry actually.

(another shameless self-plug) FYI, there is now ResourceFunction["TraceView"], a trace display function that's designed not to kill the front-end for moderate to large execution traces:

hopefully, this is actually useful to someone for tracking down weird behaviors in functions without having to worry about the front-end :)