What a request actually says (HTTP)

The phone line with eight orders travelling down it. This lesson opens one of those orders and reads what is written on it, line by line.

Before you start. You know that before a website sends anything, your laptop and the site open a phone call that delivers every byte in order, from TCP vs UDP, and the 3-way handshake. You also know that on most sites that call is then made private, so nobody on the line can read it, from The whisper code (how a call goes private). This lesson is about what gets said once the call is open and private: the words themselves.

The idea in 30 seconds

An HTTP request is plain text: a first line naming the method (the verb), the thing wanted and the version, then header lines, then one blank line, then an optional body.
The server keeps no memory between requests, so every request has to carry everything the server needs to know.
The method says whether you are asking (GET) or changing something (POST, PUT, DELETE).
Every reply starts with a three-digit status code whose first digit says who is responsible: 2xx done, 3xx look elsewhere, 4xx the client’s mistake, 5xx the server’s.
Cache-Control says how long a stored copy stays good, and an ETag lets the browser ask “has it changed?” and get back a bodiless 304 Not Modified instead of the whole page.
A cookie is a header pair: the server sends Set-Cookie, and the browser sends it back as Cookie on every later request to that site, which is how a memoryless server recognises you.

Read the story · Skip to the numbers