why is R so comfy /g/ ?
it's not. it's awful. even R advocates agree that it's full of warts.
A language tends to be "comfy" if you have too little experience or basically no perspective on the subject. It takes a lot of time and exposure to lots of languages before you realize that being a zealot about any single language is ass-headed nonsense.
Because data science is a hipster field, and things need be made retard-proof for hipster usage (see also: python, ruby).
R's only redeeming quality is a huge community of mathematicians contributing to libraries. That doesn't make a language comfy. That makes it a begrudging necessity.
It's grossly inefficient and syntactically bizarre compared to the other languages a scientist is likely to use (Python, Java, etc...) so it's also frustrating to teach people because it's paradigmatically inconsistent with other languages.
You'll never see more cargo cult programming than in the *broader* R community, which itself is enormous because almost everyone who uses R is a reluctant user (using it to do basic statistical tests, make visualizations, etc... for paper submissions, but little if anything more).
>>53478303
Because it's so slow.
After 2 months of learning I still had to google "the better way to do it" every single time I encountered a new task. Debugging is also damn frustrating.
>>53479171
And this also.
Staticians should just migrate to Python. Or Scala maybe, since some R libraries depend on Java.
>>53481580
*Statisticians
![logo[1].png](https://i.imgur.com/gNR8lLD.png "logo[1].png")
>>53481580
this.
i started learning R last year for a small project; at some point realised it was a shit of a language, and did everything with Octave.
>R is kill
>>53478319
This.
I wonder what flaws Julia will show in the long terms.
>>53481580
>migrate to python
>python4 comes out
>older research no longer works
>>53483555
This, Python shares exactly the same flaws.
Tell me about R, /g/
>>53483835
You can use <- as well as = to assign stuff.
Also, there are a lot of matrix operators and tables.
Also, the package manager isn't comfy.
>>53483861
Is there a reason to use it over Java or Python?
>>53483871
It can do EVERYTHING
But learning how to do anything is hard, because there are like 10 ways to solve something, and all tutorials do it differently and you don't even know what to do. There is a lot of documentation for packages and functions, but (for a beginner) it's quite mediocre, without any knowledge of statistics you will get lost
>>53483871
only in advanced dicktronics and if you belive lexical scope rules are your ultimate fetish
>>53483973
I'm a mathematics student who can program in Java, Python and MATLAB and I'm considering teaching myself R.
>>53484104
I'm certainly not saying it's useless or anything, it's easy to get the basics, just try it and see if it's something that fits your programming/work style. It's more on the programming side than Matlab is (tried that one briefly), as far as I can tell.
Rstudio is a nice addon.
>>53483555
I'm OK with this. They should just use an old interpreter for those files instead of forcing python to freeze its syntax.
Why use R instead of manipulating data in any language you like and then gnuplot a graph from it? Or run gnuplot from your program to generate the graphs on the fly. I truly can't see the point of R anymore.
