First they came with Mathematica

>First they came with Mathematica
>and we fought back with Sage

>Then they came with Stata
>and we fought back with R

How are we going to fight this?

Attached: fuck proprietary.png (230x230, 32K)

fight what

proprietary scientific software

I don't care what is used, but kill it. Anyways, i thought scipy was the general replacement?

topkek, you wish

Scilab?

>How are we going to fight this?

A few lines of python can do everything matlab can do.

Prove me wrong.

Pro tip: You can't.

A studio for meshing and running FEM solvers.

i know nothing about matlab but i heard that NOO octave can replace it sometimes?

>Importing a mesh is 4-5 lines
>A couple FEM libraries and some glue code and bada bing bada boom you're cooking with gas

i know i know, thats why salome is used

This. Engineers I work with who use matlab are 90% diversity hires or relatives of my boss who got nepotismd in. Anyone with a thee digit iq and a pulse can replicate matlab without the henously expensive license using python.

>paying for matlab

Attached: 5209438602672.png (413x418, 131K)

>paying at all
>not using quality interface like a patrician

have you heard of octave?

Regular MATLAB is handily beaten in terms of functionality, capability and speed by Python + Numpy if you install it with LAPACK/BLAS.

Now, Simulink on the other hand, is without competition...

1.) I have never used MATLAB outside of one class where I was required to use it since it was all the class was about.
2.) GNU Octave already exists.

GNU/Octave works pretty well unless you need something specific from toolboxes or dicom support.

Python >> matlab
You can do everything with python and more.

The problem is that many people of the academic world just use matlab/octave and refuse to learn something new. So sometimes there are a few more matlab scripts developed for a very specific area.

Unless you need all the simulink mambo jambo, ganoo octave is pretty good.

Attached: 1508470376444.gif (849x458, 153K)

This ain't about learning something new, except for meme python spacing it's still the same shit, but python is nowhere near the simple accesibility of matlab interface where you literally get great data management interface with handy GUIs and you don't have to install thousands of libraries and search for documentaries for each.

Instead you literally fucking write retard tier sentences like weighted fit in their documentation center and it tells you how you do the shit. Scientists do programming only to do science, MATLAB performance is in fact better than Python NumPy on shit like FFTs or some PDE procedures, there's literally no reason to do it in python and if you really need heavy duty performance you go for C++ anyway.

and you don't?

sage an alternative to maple, not MATLAB.
Unless sage evolved a lot recently, it is a formal calculus software. And MATLAB 's formal toolbox is just a gimmick.

Sage has the shittiest documentation I've ever seen which is why I don't use it regularly.

Now R's opponent would be python Pandas I guess, and python numpy would try to rival MATLAB.

I still prefer MATLAB for work though. It is proprietary but tiers above its rivals. And I don't mind using proprietary tools in the workplace. I'm just an engineer and if the higher ups feel like spending money it is none of my responsibility.

The point is that there are many things that you can do with python that are very difficult to implement in matlab. That is because python is a general purpose language with MANY libraries to extend its abilities.

So you can got for Matlab for matrix work and change to R for statistics and then change to bash for scripting, etc... and get confuse every time because of the differences in sintaxis and libraries . Or you can just do it all with Python,

For lightweight work is the best choice 95% of the times. The other 5 % is something very specific.

>publicly traded engineering firm
>risking a seven figure lawsuit to save thousands
Dumb nigger.

>tfw matplotlib fuck up your graph and you have no clue where to even begin fixing it

Attached: IMG_3909.png (1018x728, 1.11M)

julia kind of completely shafts matlab

replacing matlab is pretty much what julia was created for

Python

For interactive data analysis, plotting and stats R + ggplot shits all over matlab.
Python beats the shit out of it for machine learning and almost everything else.

I suppose there are some niches for it involving obscure proprietary drivers or whatever but I can't really imagine paying for matlab these days. It's shittier than other free options in the vast majority of areas

FORTRAN + LAPACK

FEniCS. Pic related is stokes flow in 18 lines of python

Attached: stokesexample.png (1024x478, 222K)

Show me the python simulink and i'm sold.

This.
Tidyverse in general is amazing

GNU Octave

Fortran

The reason people use Matlab is for its hardware interface you dumb fuckers.

Real data scientist coming through.
The language that is absolutely and eternally overrated is Python.

>R
Best for interactively diving into data and making plots, reports and shit. It's even better for interactivity than Python + Jupyter because the semantics of R, e.g. emphasis on pure functions, multimethods allow you to overload on the fly, piping together pure data frame transformations incrementally without having to constantly reconnect/reload your dataset, or even incrementally building up plot elements with ggplot2, etc. Also has nice reporting libraries like knitr.
>Python
Only good for prototyping ML algorithms. That's it. Often said to be more "general purpose" but that point is lost because it's never used in production data science, where everything is Scala/Java due to Hadoop shit. Pandas is completely inferior to R's dplyr/tidyr. Seaborn, plotly, etc. don't have the nice grammar of graphics DSL that ggplot2 has. Knitr beats the shit out of anything you can do with Jupyter, it's insulting to even bring it up. Used by brainlet webdevs who think they ML now.
>Matlab
Used for integration with Simulink and that's it. Simulink is legit god tier but Jow Forums fizzbuzzers don't actually do real work.

learn to opengl or something.

As these anons said, simulink and hardware interfacing.
Control systems classes in EE programs are practically taught out of Matlab's corresponding toolbox.

bump

if a proprietary software is working good, ill buy it

Fuck off dirty commie.

didn't answer my question

oh, nope I don't wish for an open source project similar to a proprietary project.

you know why? Support.

numpy. For Simulink, there is no alternative.

Get octave. Its open source and free. All matlab libraries and code work. Wtf are we fighting - are you a first year/freshman EE?

Scilab has a simulink alternative that's at least good enough for education though.
Also, Octave is actually really nice when you compile it with all the extras that aren't in the binary releases.

The whole "Proprietary software is well supported" idea is a massive meme, have you actually had to put up with shit like NI's software?

I only ever had to use matlab in the first semester of university to make some small custom Lego "robots" do whatever specific task we decided to design them for. It was convenient, and we got the license for free.

not mocking you but did you never learn how to write FEM solvers? I did that second year

wtf. you all got to use matlab toolboxes? i went to a medium level uni and my professors made us actually write our own tools

I just use python and R.

MATLAB is so much easier then python it is not even funny, you immediately recognize that python was not designed as a language to Interact with matrices, it is significantly more convoluted.
The fact that arrays don't start az 1 and that a Matrix isn't the default data type for everything makes it clear that it will never be able to achieve what MATLAB can.

It actually isn't even funny how superior MATLAB is to python when it comes to ease of use and expressing mathematical algorithms.

>A few lines of python can do everything matlab can do.
So can a few lines of C...

Matlab is nice in the way that it has a lot of readily available built-in functionality, but it falls short whenever your use-case falls outside of that, when making some optimized algorithm for solving some problem etc. As such, crusty old maths professors who are scared of computers love matlab, but everyone else goes elsewhere for their computational needs.

Another MAJOR drawback of matlab is that it's closed source, meaning that if there is a bug / retarded approach to an algorithm hidden behind the curtains, the user has no way of finding that out. And as a result may end up with bad data.

>opposing proprietary software, which only exists to facilitate monopolization, is communism
Jow Forums is the place to go for political ignorance.

Thats the way, only pussies use toolbox

Octave. And you're about 10 years late.

t. babyducked engineer student

>can't use a real programming language like C or ASM
>calls others brainlets

Attached: 1514592191675.jpg (600x600, 57K)

Yeah. You need to do a lot of extra work in C to do this type of programming efficiently. It's really doable with the libraries available now, but when you're doing exploratory coding with your data, that gets really cumbersome. Couple that with the fact that Matlab/octave and R use really efficient versions of these libraries on the back end, so it's barely slower than doing it in c, and it's just not worth it 90% of the time. When it is worth it is when you run in production, but by then you write a c/fortran/asm extension for the slow parts.

>than
your whole argument crumbled there, ignorant cunt

?
I didn't use "than" in my entire post.
I am also not a native speaker, so sorry if I made a mistake.

Why is Octave so slow compared to MATLAB?

>>tfw retarded

>needing matlab
>ever
Git gud nigger.

>Mathematica
>Sage
This isn't really the case. Sage was conceived and still is mostly about pure math whereas Mathematica was conceived and still is better suited for applied, the reason it is so popular among physicists. Sage has been the answer to Magma used by its creator before getting fed up by unsolvable issues owning to Magma being closed software.

>Stata
>R
Not the case either. R is an up-to-date S which predates Stata for nearly a decade.

>Matlab
The obvious answer will be Octave/Scilab that are Matlab clones. But an even better answer will be Julia which attempts to pick Matlab and massively improves it.

>Data scientist

More like data janitor

it's made by freetards

Matlab's hideous programming-language semantics. Pure boomerville.

t. brainwashed american

ME here
get the fuck out with that string shit

in FEA the "cool graph" that looks good in picture is maybe the last thing that counts

More like meme janitor

>using matlab FEM solvers
>keknics
>kid thinks his naïve FEM solver is worth shit

ok imbeciles, lets just say you should be using a specialized FEM software or at the very least use Fortran/C libraries to manage the important operations in case you wanna DIY, and better know what a Raviart-Thomas element is.

DifferentialEquations.jl
JuMP
nuff said.

>hideous programming-language semantics
C

C is the standard for anything good in a programming language

C is glorified assembly

isn't it beautiful?

That's why it is good.

Is there an alternative to the wolfram language?

I am using Matlab since 2007 or so. However, with the maturation of my projects I ported my stuff to C++/CUDA/python.
Well, there are some pros and cons:

+ its quite useful to test new ideas (proofs of concepts) because implementation of algorithms can be easily done
+ good for linear algebra, its more comfy to use it with matrices than numpy
+ great for fast visualization of data
+ support for windows, linux and mac
+ doesn't need any fucking indentation like python
+ you can compile C or C++ procedures as .mex files and import them directly in Matlab (this is a good solution for the problem with slow loops)

Now some cons:
- expensive, especially with add-on which they call toolboxes. For testing there are cracked versions on torrents.
- very slow with (for) loops
- code can be more elegant than python but gets messy very fast for bigger projects (this is probably because its used by scientist and not professional programmers)
- its harder to port it and install your solutions somewhere else
- in the current hype of neural network and CUDA computing python is getting much more attention and support, although Matlab has also cooperation contact with NVIDIA.
- most of the toolboxes available in Matlab are already implemented in python and available for free
- in the scientific community, if some project is coded only in Matlab is means that it is not mature and is associated with poor coding abilities and untermensch class.