Comment by TrackerFF | Hacker News Reader

Comment by TrackerFF | Hacker News Reader

I'm happy to see that Academia has embraced Python the past years. When I went to college, almost every science class which required coding / scripting used MATLAB. Luckily stats classes either used R or Octave, but MATLAB was generally heads above anything else, in terms of usage.

I have observed that some businesses and groups, that sprung out of Academia, still heavily use MATLAB. I guess that's logical - as their users come from MATLAB-heavy environments.

But still, it kinda sucks because:

- MATLAB is still proprietary, and costs money

- The larger the codebase becomes, the more incentive there is to stick with MATLAB

- You somewhat limit yourself on potential candidates that can (or WANT) to work with MATLAB.

The only true benefits I see from using proprietary languages like MATLAB, would be very specific and optimized packages / libraries, along with good support...but then again, being dependent on vendors for updates is rarely a comfortable experience.

I have used MATLAB extensively. I exclusively use python these days. I would say in terms of quality MATLAB is 8.5/10 and python is like a 6.5.

I am always searching google to fix something not working on python. In matlab, just refer to the docs.

I would say the exact opposite. I found MATLAB painful to work with at university. Using python for everything after that has been a breath of fresh air.

I work in an industry R&D group. We're a Python shop, for the most part. I've noticed that when we hire recent graduates, it pretty much doesn't matter if they've used MATLAB or Python. If they're likely to become good coders, then they come up to speed on Python equally quickly.

If not, then MATLAB tends to have more canned apps that they can use as-is or modify, but they're also unlikely to get deeply into computational work anyway.

My daughter took a college class that involved some programming, and they were told that they could use the tool of their choice, MATLAB, Python, whatever. I suggested: Use whatever the teacher is using, which happened to be MATLAB.

A benefit of MATLAB is that for basic use, it all comes from one place, and there's pretty much exactly one way to set it up. This practically eliminates the day wasted getting the tools working on each student's computer.

With Python, unless you give them strict instructions to use a particular repository such as Anaconda, or an online tool, they will discover a myriad of blogs with conflicting advice on how to get Python running, with the most opinionated having the worst advice.

With that said, so long as I've followed a consistent recipe for getting Python onto a computer, It's always been great. I've been quite happy with WinPython.

Regarding the part about wasting time on setting up stuff - because I've been plagued with that on every laptop I've used (due to conflicting/competing software and DLL hell):

repl.it or google colab.

Won't work offline, but it takes 5 seconds from launch/load to coding. Those tools are a godsend. I haven't opened an IDE or touched Conda for Python in probably 3 months now.

What I'm learning, at least from my kids, is that students are always online when they're doing anything with a computer. So, they are typically comfortable with using online tools.

Reply on news.ycombinator.com