Julia has been around for about 7 years, but the price of base Matlab has been quite stable. Actually if you take into account the bundling toolboxes, it is getting more expensive. Since Julia can do most of the stuffs Matlab offers (at least in my case), I consider Julia as a substitute of Matlab to some extents. It puzzles me that Matlab is still maintaining a high price.
Matlab has plenty of “free” competition beyond Julia: SciLab, SciPy+Numpy, Octave are even older than Julia. There’s probably enough Matlab-based projects out there for which moving to an alternative is more expensive (and risky) than paying for Matlab’s license.
There’s also plenty of things that Matlab can do and Julia can’t: Simulink (at least the graphical interface), and easy conversion to C and VHDL/Verilog are two examples among many.
Personally, I can’t wait for the world to leave Matlab behind and move to Julia. It won’t happen overnight, unfortunately.
So what exactly do you want advice on?
1 Like
MATLAB is a polished, established software. Universites and academics have both short term and long term contracts. But really it’s because of familiarty. Many people are just stuck using the tools they are used to and it takes a lot of effort to switch to another language. Moreover, from my personal experience these are some points that work against Julia:
-
Julia is fast, but it requires you to have a little bit of knowledge of computer science. I know folks that copy/paste scripts in different folders rather than having one script. They are not experienced in computer science paradigms, and definitely don’t know about type stability and multiple dispatch. So when they try Julia, they write bad (often unstable) code which dosn’t give them the speed they were looking for. This leads to my second point…
-
The time to first plot (or the speed of first run). This is a common gripe and you’ll see threads on this almost every week. Just today, my PI asked me to run a simulation and plot the results (which after compilation takes < 1 second). However, because I had just launched a fresh copy of Julia, it took “forever” for the results to pop up. It was a frustrating since I was screen sharing and we were just both sitting their silently. AFAIK, there is an interpreter in the works (actually there already exists a fully functioning interpreter… right?) that will help with this.
-
A polished IDE. Juno is amazing and the VS code extension is great. But like I said, for those that are not experienced programmers or were not trained in computer science, they barely know how to launch a terminal. It’s easier for them to launch a self-contained software that has all the functionality from a point and click level. For example, curve fitting in Matlab (a GUI) vs using
LsqFit.jlin Julia.
I slowly find this is changing. My lab/mentors/PI/colleagues all WANT to learn Julia. The above gripes are not dealbreakers for most. It’s just a matter of allocating time and resources to learn it.
3 Likes
If Matlab has less customers, the price should actually increase and not decrease.