What's Jow Forumss thoughts about julia lang?

Attached: 1545008655581.jpg (2010x1206, 270K)

1-based indexing is an unholy abomination.

Julia is meant to replace Python. Most people who use Python in data science are casual programmers who dont care about the quality of the programming language they use and want something that looks easy. Python does look easy to the novice. But any good programmer knows that Python is a shit language. Although Ruby is a better language than Python it uses blocks which are really shitty replacement for methods and high order functions. Julia borrows the only good thing about Python which is its comprehension syntax which is much more usable than blocks. Julia syntax is closer to Javascript. A lot of people like Javascript because its a clear and simple language. But Javascript is not useable as a general purpose scripting language (node does not qualify as a standard library). I would argue that Javascript is not a simple language anymore with ES6, ES17, etc.

and it's bad, too

I've been trying to get Makie.jl to work with it but it always gives me some shitty fucking error and it's frustrating

> 1-indexed arrays

Attached: highronda.jpg (720x708, 60K)

>Meme
Install R
Then... Install Python.

lol why would anyone ever install R

Emoji programming and undergrad tier libraries

Attached: 1545028713916.png (831x138, 119K)

>better than python
>if it doesn't outperform c# it'll have to sell me on some other features to be better than lisp
>otherwise it's just a decent clojure
>somehow i get a feeling it was inspired by fortran
>syntactic macros are nice
>syntax is very clean
>the built in python stuff and package manager worries me slightly, but that's all fixable
>dynamically typed seems good
>concurrent programming is a must in today's age

over all the high range of probability being at least c+ and the low range of probability being B+.

I'd give it five years to try to beat out python ahead of c#. Maybe seven if microsoft keeps being inefficient.

A programming language named after a woman's name is bound to be complete garbage

It's the best data science language
>actually scales because no GIL
>fast llvm jit backend
>still dynamically typed so works in a shell

I agree. On top of that overloaded operators have weird precedence rules.

For example if I overload • to represent dot product between vectors then a+b•c, where a is a scalar and b and c are vectors, will first evaluate a+b and then the dot product. This will result in very slow and suboptimal machine code, because (a+b)•c will involve n additions and n multiplications whereas a+(b•c) only involves 1 addition and n multiplications.

Basically user-defined operators have lower precedence than anything else, forcing you to litter your code with parentheses and if you forget to add a pair, you'll suffer huge hard-to-find performance penalties.

This is most definitely a seriously superior language to python. Are you sure you are aware what makes a good programming language?

Attached: proprietary-vs-foss.png (1008x630, 44K)

I've read that in Julia you can change the syntax however you want. is that true?

>JIT compilation times are a huge issue still, with no easy solution in sight.

Not for Terry, CIA nigger

thats probably referring to Lisp-like macros in Julia

I tried it out a bit and the syntax and core designs are good for data science for sure. It's lacking good libraries which made it hard to justify using currently.

why does everyone still use python then

>> 1-indexed arrays
Just so you know, zero indexing became a 'standard' only because C took over as the lingua franca of the IT industry so of course all the C family languages like C++,Java, C#, Perl, Javascript, PHP, etc all copy zero indexing. C is a systems programming language that should never have been used in userspace, zero indexing only makes sense as a means of counting memory space as offsets from zero. If you look at programming languages previous to C you will C many started arrays with a one-index.

The following languages are one-indexed: Matlab, Mathematica, R, Algol, Cobol, Fortran, PL/I, Smalltalk, APL, AWK, Lua.

Ada and Pascal have optional starting numbers for arrays.

technically in lua you can index starting from whatever you want

How does Julia have lisp macros? I thought such a feature required parenthesis everywhere like lisp? Why hasn't every language copied it if it's possible in a normal language?

Python start to taugh in schools massive to non-CS non-IT as first language, some people start to build data analysis/sci libs for python, a lot social science,agro,meds research start to use python as compute language over SAS,R,matlab, CS people wrote more python libs,python JIT, now new students in non-IT learn python for data science.

docs.julialang.org/en/v1/manual/metaprogramming/index.html

Other language prefer types or begin half-baked macros system

I had someone recommend it, but I completely lost interest once at the words "dynamically typed".

> The green line went back in time.

>Ada and Pascal have optional starting numbers for arrays.
Better way to describe it is that you can use any ordinal type as the index for an array, so 0..100, 1..255, -273..1000, a..z, monday..sunday, etc., could all be used as indices. Imagine a WorkSchedule type as an array [monday..friday] of DailyTask types. In other languages, you would end up using some convention like 0 = monday, 1 = tuesday, etc., but this lacks type-safety.

>julia is dead...

At this point my brain is so marinaded with Python data science stack I don't even give a fuck or want to learn a new thing that does the same thing for any reason.

Attached: 1472922780198.jpg (600x600, 55K)

>n additions
I will kill myself if you can demonstrate to me that the julia compiler cant optimize that out.

Learn R+dplyr and then you won't be tied to Python anymore. Use magrittr as much as possible (aka %>% pronounced: pipe).
Only then will you be liberated and can use R, Python, or Julia for Data Science (hell even Matlab, or Fortran also)

I would never inflict R upon anyone. Its literally PHP tier

Except it's literally not

Attached: languages-1-768x576-1.png (768x576, 62K)