📖 Programmer's Bible
First Chapter: From Zero to Your Own Website › Part 2: The Network › 49 of 60 · See on the map

HTTP

⏱ 6 min read

In one sentence: HTTP (HyperText Transfer Protocol) is the set of rules that web browsers and web servers use to send requests and responses for web pages, pictures and other files.

Why should I care?

You use HTTP every time you open a web page. In Protocols, you saw a short HTTP request and response, and the status codes 200 and 404. Here you will see the parts of both messages. MDN, a website that documents how the web works, calls HTTP "the foundation of any data exchange on the Web".

Analogy

Think of ordering at a shop counter with a written slip.

  • You write an order slip: what you want to do ("give me"), which item, and some extra notes ("please wrap it").
  • The clerk writes a reply slip: a short code that says how it went ("OK" or "we do not have it"), some notes about the item, and then hands you the item.

An HTTP request is your order slip. An HTTP response is the reply slip, often with the item attached.

Where the analogy breaks down: a clerk may remember you the next time you come. Basic HTTP does not. Each request stands alone.

How it actually works

As in Client and Server, the client (usually your browser) always sends the request, and the server answers.

One HTTP request and its responseThe browser sends a request to the server with the method GET and the path /, which asks for the page. The server sends back a response with the status code 200 and the message OK, and then the page itself in the body.ServerYourbrowserRequest: GET /Response: 200 OK +the page

The request

MDN lists the parts of a request:

  1. A method: what the client wants to do. MDN says the usual ones are GET, to get something such as a page, and POST, to send something such as a filled-in form.
  2. A path: which page or file on the server you want. It is the address without the https:// and the website name. For developer.mozilla.org/en-US/docs/Web/HTTP, the path is /en-US/docs/Web/HTTP.
  3. The version of HTTP.
  4. Headers (optional): extra information for the server.
  5. A body, for some methods such as POST: the data you send.

A header is one line with a name, a colon, and a value. Here is an example request, built from MDN's examples:

Diagram: an HTTP request of three lines. Line 1 holds the method GET, the path / and the version HTTP/1.1. Lines 2 and 3 are headers: Host names the server, and Accept-Language says the client prefers German.

GET / HTTP/1.1                    <- method, path, version
Host: developer.mozilla.org       <- header: which server
Accept-Language: de               <- header: "I prefer German"

Accept-Language tells the server which human language the client prefers (de means German). A browser sets it from the language of its menus, and from any languages you choose in its settings.

The response

MDN lists the parts of a response: the version, a status code (a number that says if the request worked, and if not, why), a short status message such as OK, headers, and often a body: the thing you asked for, such as the page.

Diagram: an HTTP response. Line 1 holds the version, the status code 200 and the status message OK. Line 2 is a header, Content-Type, which says the body is a web page. After the headers comes the body: the page itself.

HTTP/1.1 200 OK                   <- version, status code, message
Content-Type: text/html           <- header: the body is a web page
... (more headers)
(then the body: the web page itself)

Status codes come in five groups

The first digit of a status code tells you its group:

Codes Group Example
100–199 Informational
200–299 Success 200 OK: "The request succeeded."
300–399 Redirection 301: the address of the page has changed permanently (forever)
400–499 Client error 404 Not Found: the server cannot find what you asked for
500–599 Server error 500: the server met a situation it does not know how to handle

Each request stands alone

MDN calls HTTP "stateless". This means each request stands alone: HTTP makes "no link between two requests". So, by itself, a server does not remember your last request. Websites remember things with cookies: small pieces of data that a server sends to your browser, and that the browser sends back with later requests. You will learn more in Cookies and Privacy.

HTTPS is HTTP sent over an encrypted connection. Encrypted means turned into a code that people without permission cannot read. You will learn how in HTTPS and Encryption Basics.

How it came to be

MDN says HTTP was developed by Tim Berners-Lee and his team between 1989 and 1991. It was one of four building blocks of the World Wide Web. The others were HTML (the format of web pages), the first web browser and an early web server.

HTTP then grew, version by version:

  • HTTP/0.9: each request was a single line. The only method was GET.
  • HTTP/1.0: described in RFC 1945, in May 1996. Headers arrived with this version.
  • HTTP/1.1: RFC 2068, January 1997.
  • HTTP/2: standardized in May 2015.
  • HTTP/3: RFC 9114, June 2022. It uses a protocol called QUIC instead of TCP to carry its messages.

🧪 Try it in your browser

You need a computer with a keyboard. These steps only look. They change nothing.

  1. In Chrome, Edge or Firefox, open MDN's page about HTTP in a new tab: developer.mozilla.org/en-US/docs/Web/HTTP.

  2. Open the network tool, as in Protocols: in Chrome and Edge press Ctrl + Shift + J (Mac: Cmd + Option + J) and click the Network tab. In Edge, if you do not see it, click the More tools button and choose Network. In Firefox press Ctrl + Shift + E (Mac: Cmd + Option + E).

  3. Reload the page with the browser's reload button. Rows appear: one row for each request.

  4. Click the first row. It is usually the page itself. A panel with details opens.

  5. Click the Headers tab in that panel, if it is not open already.

  6. Find the status code (Chrome and Edge: in the General section; Firefox: next to Status). You will most likely see 200 OK. You might see 304. MDN says this means the server does not need to send the file again. The copy your browser saved earlier is still good. Both are normal.

    By default, the tool shows each part as a label and a value, not as lines like the examples above.

  7. Find the Response Headers section (Firefox: Response headers). Look for content-type. Its value starts with text/html: the body is a web page. (Header names may be written in small letters.)

  8. Scroll to the Request Headers section (Firefox: Request headers). These are the lines your browser sent. Look for:

    • user-agent: a long line that names your browser and your operating system. (MDN calls a program that acts for a person, such as a browser, a user agent.)
    • accept-language: the languages your browser prefers, for example en for English. It may be a longer list. That is fine.

  9. You will see many other headers too. That is normal. Close the tab when you are done.

Common misunderstandings

  • "HTTP is only for web pages." No. MDN says it also carries pictures and videos, and form answers sent to servers.
  • "The server remembers everything I did." Not by itself. HTTP is stateless. Websites use cookies to remember you between requests.
  • "An error code always means my internet is broken." No. A 404 or a 500 comes inside a response, so a server did answer. The table above shows what each group means.

Recap

  • HTTP is a question-and-answer protocol: the client sends a request (method, path, headers), and the server sends a response (status code, headers, often a body).
  • Status codes come in five groups, from the 100s to the 500s: success is in the 200s, "not found" is 404, server errors are in the 500s.
  • HTTP is stateless: each request stands alone. Cookies help websites remember you.

Sources

Written with AI · Reviewed by AI · Not yet checked by a human

Found a mistake, or something you did not understand? Report a problem.