From Sand to Website

How a computer thinks — Binary & Silicon

A computer looks intelligent, but deep inside it only understands two things: ON and OFF. Everything — photos, videos, this webpage — is just billions of tiny switches flipping between 0 and 1.

Why only 0 and 1? Why not 0, 1, 2, 3, 4?

A chip is made of billions of microscopic switches called transistors. A transistor is electrical — it either lets current flow (ON = 1) or blocks it (OFF = 0). Two states are extremely easy and reliable to detect, even when the electrical signal is weak or noisy.

If we tried to use 5 states (0–4), the chip would have to measure exactly how much voltage is present — for example 0V, 1V, 2V, 3V, 4V. Real circuits are noisy: heat, interference and manufacturing differences shift voltages slightly. A "2" could easily be misread as a "3", and one wrong digit corrupts the data.

Why Silicon for hardware?

Materials come in three electrical families. Silicon belongs to the magical middle one:

Talking to the machine — Languages & Frameworks

The CPU only understands 0s and 1s. Humans can't write billions of 0s and 1s — so we invented programming languages as translators, in layers from "machine-friendly" to "human-friendly".

The language ladder

The languages we discussed

Language vs Framework vs Runtime — don't mix them up!

The Internet — Domain, DNS, IP & Servers

The internet is just millions of computers talking to each other. Four ideas make it work: every machine has an IP address, humans use domains instead, DNS translates between the two, and servers hand out the content.

How DNS works — step by step

When you type www.example.com and press Enter, this happens in milliseconds:

How a server works — the request/response cycle

A "server" is both the hardware (a powerful always-on computer in a data centre) and the software (Apache, Nginx) that listens for requests, typically on port 80 (HTTP) and 443 (HTTPS). One server can host many websites; one big website can need hundreds of servers (see Module 6).

Hosting a website — Types, IP, OS, Control Panels & SSH

"Hosting" simply means renting space on a server that's online 24×7, so the world can reach your files. Here's the whole process, and every choice you'll face.

How to put a website live — 6 steps

Types of hosting

How IP works in hosting

What is the OS in hosting, and what does it do?

Every server runs an Operating System — the base software that manages the hardware and runs everything else. The OS handles: files & permissions (who can read/write what), processes (running Apache, PHP, MySQL), memory & CPU allocation, networking (ports, firewall), and users & security.

Control panels — hosting with buttons instead of commands

A control panel is a web dashboard that manages the server graphically: create email accounts, add domains, manage files and databases, install SSL, take backups, one-click WordPress installs. Popular ones: cPanel (most common, Linux), Plesk (Linux + Windows), DirectAdmin, hPanel (Hostinger), and free ones like CyberPanel / aaPanel for VPS.

4 ways to communicate with your server

SSH in detail — your first commands

SSH creates an encrypted tunnel from your computer to the server. You type commands; the server obeys. Connect like this:

# connect: ssh username@server-ip  (default port 22)
$ ssh root@203.0.113.25
$ ssh -p 2222 user@203.0.113.25   # custom port
$ scp backup.zip user@203.0.113.25:/home/user/   # copy a file TO the server

Making it fast — Speed, Cache, Responsive Design & CDN

Speed is money: slow pages lose visitors and Google rankings. Three weapons make sites fast — caching, optimisation, and a CDN.

How to check website speed

Key numbers to watch (Core Web Vitals): LCP — largest content paint < 2.5s (main content visible), INP — interaction delay < 200ms (responds fast to taps), CLS — layout shift < 0.1 (nothing jumps around).

What is cache, and why was it invented?

Cache = a saved copy of work already done, kept somewhere fast. It exists because rebuilding the same page for every visitor is wasteful: without cache, every visit makes PHP run, MySQL query, and the page rebuild from scratch — hundreds of times per second for popular sites. With cache, the server does the hard work once, saves the finished result, and serves that copy instantly to everyone else.

The cache layers — from visitor to database

A request stops at the first layer that has a valid copy — the earlier it stops, the faster the response.

Fixing speed & cache issues — CMS end vs Server end

Desktop / Mobile / Tablet optimisation — responsive design

One website must adapt to every screen. This is done with responsive design: flexible layouts + CSS media queries that apply different rules per screen width.

How to check responsiveness: Chrome DevTools → Device Toolbar (F12 then Ctrl+Shift+M) simulates any phone/tablet; test real devices too; and run PageSpeed's mobile test. Sites like responsivedesignchecker.com preview many sizes at once.

What is a CDN, and why do you need it?

A CDN (Content Delivery Network) is a worldwide network of servers ("edge servers" / PoPs) that keep cached copies of your site's files. Visitors download from the server nearest to them instead of your origin server.

Benefits: faster load everywhere, less load on your server, absorbs traffic spikes, DDoS attack protection, and free SSL on services like Cloudflare.

Handling heavy traffic — Scaling & Load Balancing

What happens when 10,000 people hit your site at once — a sale, a viral post, exam results? A single server chokes. Here's how big sites stay up.

Two ways to grow

The load balancer — traffic policeman for cascaded servers

With multiple servers ("cascaded" in a tiered setup), a load balancer stands in front, receives every request and distributes it to whichever server is free/healthiest (round-robin, least-connections, etc.). If a server dies, it's removed automatically — visitors never notice.

The full anti-traffic-jam toolkit

Keeping it safe — SSL & HTTPS

That padlock 🔒 in the address bar is SSL/TLS — encryption between the visitor's browser and your server.

What SSL does

How the handshake works (simplified)

Inside the chip — RAM, Cache, Processors & Apple Silicon

Back to hardware — this time deeper. How the CPU, RAM and cache cooperate, why chips slow down when hot, and why Apple says "Unified Memory" instead of "RAM".

The jobs: Processor, RAM and Cache

The memory pyramid: higher = faster but smaller & costlier. Data constantly flows up and down this ladder.

What is thermal throttling?

Working transistors leak energy as heat. The harder and faster a chip runs, the hotter it gets. Beyond a safe temperature (~95–100 °C), the chip protects itself by deliberately slowing down its clock speed — that self-defence slowdown is thermal throttling.

This is why thin laptops slow down during long exports, and why gaming PCs have big fans and liquid cooling. Efficient chips (like ARM designs, below) generate less heat and throttle later — or never.

CISC vs RISC — two philosophies of processor design

ARM is a company that designs RISC processor blueprints and licenses them; Apple, Qualcomm, Samsung and others build their own chips on those designs. ARM's efficiency is why your phone lasts all day on a small battery — and why Apple moved Macs from Intel (CISC) to its own ARM-based M-series chips.

Windows PC architecture: CPU + separate GPU

In a traditional PC, the CPU and the graphics card (GPU) are separate chips with separate memory: the CPU uses RAM, the GPU uses its own VRAM. They talk over the PCI Express (PCIe) slot/lanes.

When the GPU needs data (say, a video frame or an AI model), that data must be copied from RAM → across PCIe → into VRAM, and results copied back. Every copy costs time (latency) and energy — and all that extra work by CPU, GPU and memory chips produces more heat, pushing the system closer to thermal throttling. Duplicated data also wastes memory: the same texture may live in both RAM and VRAM.

Apple Silicon: CPU + GPU + NPU on one chip, one memory

Apple's M-series is a System on a Chip (SoC): the CPU, GPU and NPU (Neural Processing Unit, for AI tasks) are built together on the same silicon, all sharing one pool of memory mounted right next to the chip — Unified Memory.

So why does Apple write "16GB Unified Memory", not "16GB RAM"?

The whole journey in one line each