Interpreted

Julia is probably my favorite language for numeric, which is most of what i do. But it has a somewhat narrow range of applicability, as it's pretty shit for anything that falls outside of its use case.

>as fast as C
Sometimes yes, but usually it's a bit slower. Julia will usually offer a FAR better development time to performance ratio though.

As someone in the target audience for julia, this is a VERY good thing. Numerics is generally a bit more exploratory than "normal" programming.

>On the downside, the plotting library is inferior.
You can switch between a bunch of backends with plots.jl, including pyplot. So the question is, are you retarded? The plotting libraries in Julia is REALLY good, with the only drawback being the compilation time whenever you start a new interpreter.
>Also the lack of library might be a obstacle to do non trivial things without reinventing the wheel.
Julia has a crazy good library system, with the ability to call python libraries as well. Again, are you retarded?
>However trivial things are more straightforward than with numpy IMO
A lot of "trivial" operations are build directly into julia, negating the need to import numpy every time. Other than that, its usually very similar. But i would say julia is somewhat easier in this regard as well.

>Julia will usually offer a FAR better development time to performance ratio though.

Attached: 1531536214089.png (434x327, 43K)

>interpreted
>as fast as C

you can only choose one

>using a language named after a JAV star

This might seem redundant to you, but i'm in physics. I very often need to write up some quick program to solve a problem, for which i want the solution as quickly as possible so that i can move onto other things. This means that i want to minimize both the runtime AND the development time. Usually the code is never ran again, so i don't really give a shit about the code as long as it provides the output i want.

you can repl compiled languages

This, plus its number of libraries is pretty small.

It has a nice object system too. It's sort of a simplified CLOS, which makes sense because it was made by MIT niggers who worked with Lisps.

JIT != interpreted
Still, altrough the compiler often works literal speed magic it can't do that all of the times, so you sometimes (enough times to be annoying) end up with some slow pieces of code you need to keep changing until the compiler likes it

Post full name and/or links retard

I use Julia, R, and octave all together because I'm too lazy to learn how to take integrals in R and graph in Julia

> 1-based indexing

Attached: penn_jillette.jpg (300x400, 31K)

>interpreted

Explain

>that's a bad thing

base 0 indexing is moronic in a high level language

mfw this faggot still tries shilling his shit language on Jow Forums .. will you ever stop?

>faster than C
hahahahahahaha

Attached: consider suicide capcha.jpg (300x57, 7K)

It uses JIT compilation.

Oh look, it's another hipster meme language.

This, and Julia isn't interpreted.

that's my name

>it's pretty shit for anything that falls outside of its use case
Why? Just the lack of libraries? Is anyone doing stats in Julia?

The developers refuse to disclose the origin of the name. When someone pressed them, they just said that they took the name from their earlier project also called "Julia".

Petty.

Did your parents give it to you?

javlibrary.com/en/vl_star.php?s=ayua

She has a pretty face but usual for a Japanese. Is she hapa?

Attached: pppd348pl.jpg (416x536, 94K)

>>interpreted
>>as fast as C

Attached: Computer.gif (640x480, 1.26M)

My daughters name is Julia.

Based and [0]pilled
Low testosterone and [1]pilled

I always thought it was named after Gaston Julia

ok but can it do multithreading properly?

Python runs on c so if you're faster than python you're faster than c too

Attached: 355.jpg (680x1000, 72K)

I wouldn't say it is moronic but it is unnecessary mimicry of hardware

It would even be relatively trivial to implement a hardware addressing mode with 1 as the first offset, so it could be a zero overhead abstraction.

Yes, Japaneses get some insane fetiche for JAV stars begin Japanese + European/American/Russian Race Mixers

Julia is trash as many people have already outlined in this thread.
Lisp is where it's at. As fast as C (for real instead of the julia meme way), proper metaprogramming and proper repl with actual image-based programming. Not to mention conditions and restarts which makes julia feel like a child's toy (actually that's really what it is).

I think you mean Scheme.

No. Scheme is a better language but has no viable implementation. Lisp does. Scheme implementations doesn't have conditions and restarts, have really limited images (in the case of racket, there's pretty much no image: it does no better than python), have garbage error messages so despite not even being able to examine the system at the point of failure, there's no way to know what failed in non-trivial code, and are generally quite buggy.

Attached: Screen Shot 2019-02-07 at 4.52.40 PM.png (268x262, 188K)

>python but fast
How does it compare to Nim?

walmart uses clojure

Clojure is neither scheme nor lisp. It's also just a crippled lisp with jvm interop. They use it because the few real lispy features of clojure are probably a win, but only if they can keep interacting with their huge java-based tools. It is possible that abcl would have been a better fit for them, but that they'd never heard of it (while clojure has become quite popular for some reason).

what's crippled about it?

>Is anyone doing stats in Julia?
Not them but everyone uses R, people try to say python or julia online. Scrape with python, analysis nope. Probably someone doing it the hard way because their idiots.

Which CL implementation is faster than Julia and at what tasks? If you name one, explain why you think it is faster.

Is R that much better than Python for analysis? Although I am not a professional statistician and don't know R, Pandas and statsmodels have everything I've ever needed.

one-based numbering fucking cancer and not considering zero a natural number ruins lots of cool numerical properties.

>the first element is at the zeroth index
Only programmers will defend this.
The bugs caused by forgetting to subtract 1 from the length of an array when indexing the last element have alone wasted thousands of hours.

What editor do you use?

>What editor do you use?
second poster, I use emacs
I bet first poster uses vim

Acme

best girl in eva?

kaworu

Misato

So we have
0-indexing, emacs, kaworu
1-indexing, Acme, Misato
Not even sure which team is worse.

>>interpreted
>>as fast as C

Back to Jow Forumseddit, kiddo

Let's try to resolve this. What OS do you prefer (name the distro for Linux)? Favorite soft drink? 6502 or Z80? C or C++?

OpenBSD
Sangaria Ramune
Z80
C

Temple
Faygo
6502
Holy C

tYpIcAl.

نيغا the what ماذا

Attached: 1491526537030.jpg (250x250, 7K)

My name is Yuliya :3

best belmont
youtube.com/watch?v=ayaWIKM9iRE

Attached: image.png (496x688, 466K)

Fortran amd Sepples are both legitimately faster than C.

Holy fuck

Go off yourself, your retarded brainlet.
Nobody will miss the pajeetware you would've produced

> one-based numbering fucking cancer

true, because having a zero offset makes perfect sense

> not considering zero a natural number

b-b-but it is NOT a natural number.

> cool numerical properties

yes, exactly. these are only *cool*, not real.

Programming is 100% not math.
T. Math masters student strongly focused in algebra and mathematical logic who can't figure out how to program

How does a math major not figure out how to program? I don't get it. I was one, and some of the other math majors I studied with couldn't program at all. Even though I tried, I never figured out what they had a problem with. Most of them were girls (female), though. Being pretty good at mathematics, especially at math olympiad problems, but completely sucking at programing was a girl thing.

>as fast as C
much slower than C

imagine fanboying for a fucking programming language
how much of a beta you have to be

because math isnt programming

Not a helpful responsive.

LO N D O N

I don't care if it's helpful, it's the answer
You can reason about programming using math but that isn't the way practical programmers think most of the time

Programming is not math.
Not a single operation in programming is math,
and none of the objects are mathematical either.

>Not a single operation in programming is math
Print(1+1)

>Not a single operation in programming is math
for (i = 0; i >= 10; i++)

programming paradigms are based on formal systems which is maths

>programming paradigms are based on formal systems
nope
programming paradigms can be based on anything, including "shit I just made up because whatever" (OOP)

>repl means interpreted
The absolute state of Jow Forums

Attached: 1232931963267.png (403x428, 103K)

Some people like to use computers to actually compute things (in the literal, pre-computer sense) and they they need math for that.

I don't know a single mather who knows how to program. Those that can write scripts (at best) write such awfully written, inefficient scripts that it makes me want to stab them, and they're a very small minority in the first place, the others being borderline afraid of computers.

Thanks for proving his point.

SBCL at virtually everything not hardware-accelerated by julia (but the same or faster using appropriate libraries).
CCL likewise, except for big numbers.
Benchmark games are not a good benchmarks but they help illustrate that in general. They show that SBCL is as much as 10x faster than julia in several cases, when excluding the outlier reverse-complement (SBCL being more than 1000x faster). CCL is generally about as fast as SBCL, sometimes faster and sometimes a bit slower, except it's very slow for big numbers. On the other hand it has a better GC (precise instead of conservative, false retention in SBCL is a real problem for numerical tasks). Julia's GC is also precise, which is great, but the amount of C interop in the core makes it hard/impossible to implement classic optimizations like moving/compaction. It tends to suffer from large GC pauses as a result, compared to both SBCL and CCL.
Any language should be approximately as fast as julia for numerical code when using libraries because that's just what julia does in the background anyway. The fact that it's integrated is nice but really no deal breaker when CL has mgl-mat and the likes.

It has most features from a real lisp removed (mostly due to sprucing forced immutability everywhere. No clos is a big one), it has no real image-based capacities and no restarts.

>provides the output you want
please leave physics and academia while you're at it. you're going to be a shit SHIT GODAWFUL scientist if you keep up that practice

>one-based numbering fucking cancer and not considering zero a natural number ruins lots of cool numerical properties.
There are no *cool* numerical properties to using zero indexing which is why counting arrays/vectors in math always starts at one. And its also why all math programming languages like Matlab, Mathematica, R all start array counting at 1

The only reason starting arrays at zero index became popular is because the C language became popular due to unix becoming popular. C is a systems language that should never have been used for user space applications and so it counts memory by offsets like assembler. So all the C family languages like C++, Java, C#, Perl, PHP, Go, Javascript all copied zero indexing. Most all languages before C such as Fortran, Cobol, Algol, Pascal, Smalltalk, APL were indexed starting at 1

Attached: image008.gif (310x117, 2K)

This but unironically, mathematician genocide soon

The Algol/Pascal family aren't inherently 0- or 1-indexed. In Pascal, any ordinal type can be used as an array index. 1..10, 0..9, 'a'..'z', Mon..Fri could all be array bounds.

That said, there is one type in Pascal that does specifically use 1-indexing, which is string types. The string "hello" for example would be a packed array of char [1..5].