Julia

TL;DR

One-liner: Julia gives you Python’s ease with C’s speed - no compromises.

Core Strengths:

• As fast as C, as easy as Python
• Built for scientific computing and AI
• Multiple dispatch as core paradigm
• Excellent interop with Python, C, R

Philosophy

Julia was designed to solve the “two language problem”:

• Fast by default - JIT compiled, no need to rewrite in C for performance
• Multiple dispatch - Functions specialize on all argument types, not just the first
• Math-friendly - Unicode operators, matrix syntax, broadcasting
• Composable - Packages work together seamlessly

Julia is what NumPy would be if it were a language.

Quick Start

Install

# Using juliaup (recommended)
curl -fsSL https://install.julialang.org | sh

# macOS
brew install julia

# Or download from julialang.org

Verify (latest: 1.12)

julia --version  # julia version 1.12.0

First Program

Create hello.jl:

println("Hello, World!")

julia hello.jl

Julia REPL

julia
julia> 1 + 1
2
julia> sqrt(2)
1.4142135623730951

Language Essentials

Variables & Types

# Variables (no type declaration needed)
name = "Alice"
age = 25
price = 19.99
active = true

# Type annotations (optional)
x::Int = 10

# Arrays
numbers = [1, 2, 3]
matrix = [1 2 3; 4 5 6]  # 2x3 matrix

# Dictionaries
user = Dict("name" => "Alice", "age" => 25)

Control Flow

# if-else
if age >= 18
    println("Adult")
elseif age >= 13
    println("Teen")
else
    println("Child")
end

# Ternary
status = age >= 18 ? "adult" : "minor"

# for loop
for i in 1:5
    println(i)
end

# while loop
while count > 0
    count -= 1
end

Functions

# Standard function
function greet(name)
    return "Hello, $name!"
end

# Short form
greet(name) = "Hello, $name!"

# Anonymous function
add = (a, b) -> a + b

# Multiple dispatch (the key feature!)
function process(x::Int)
    println("Integer: $x")
end

function process(x::String)
    println("String: $x")
end

process(42)       # Integer: 42
process("hello")  # String: hello

Broadcasting

# Apply function element-wise with .
x = [1, 2, 3]

x .+ 1        # [2, 3, 4]
x .^ 2        # [1, 4, 9]
sin.(x)       # [sin(1), sin(2), sin(3)]

# Works with any function
f(a, b) = a + b
f.(x, 10)     # [11, 12, 13]

Structs

# Immutable struct (default)
struct Point
    x::Float64
    y::Float64
end

p = Point(1.0, 2.0)
# p.x = 3.0  # Error! Immutable

# Mutable struct
mutable struct MutablePoint
    x::Float64
    y::Float64
end

mp = MutablePoint(1.0, 2.0)
mp.x = 3.0  # OK

Gotchas

1-based indexing

arr = [10, 20, 30]
arr[1]    # 10 (not 0!)
arr[end]  # 30

First run is slow (JIT compilation)

# First call compiles the function
@time sum([1,2,3])  # 0.05s (includes compilation)

# Second call is fast
@time sum([1,2,3])  # 0.000001s

Arrays are column-major

A = [1 2 3; 4 5 6]  # 2x3 matrix

# Iterate column by column for performance
for j in 1:3, i in 1:2
    println(A[i, j])
end

Global variables are slow in loops

# Slow
x = 0
for i in 1:1000000
    global x += i  # Accesses global
end

# Fast - use functions
function sum_loop()
    x = 0
    for i in 1:1000000
        x += i
    end
    x
end

When to Choose

Best for:

• Scientific computing
• Machine learning (Flux.jl)
• Numerical simulations
• Data science with performance needs

Not ideal for:

• Web development (use JavaScript, Python)
• Mobile apps
• Small scripts (startup time)

Comparison:

Aspect	Julia	Python	MATLAB
Speed	Fast	Slow	Medium
Syntax	Clean	Clean	Verbose
Cost	Free	Free	Expensive
ML libs	Growing	Mature	Limited

Next Steps

• Julia Documentation
• Julia Academy
• Think Julia
• JuliaHub

Ecosystem

Package Management

# In REPL, press ] to enter Pkg mode
] add DataFrames
] add Plots
] status

# Or programmatically
using Pkg
Pkg.add("DataFrames")

Popular Packages

• Data: DataFrames.jl, CSV.jl
• Plotting: Plots.jl, Makie.jl
• ML: Flux.jl, MLJ.jl
• Scientific: DifferentialEquations.jl, JuMP.jl