Just text messages

When computers talk to each other via the internet all they do is send around messages.

You can think of these as simple text messages: They may be encrypted, they may have some binary content attached to them (such as an image, or a movie), but essentially they’re just text messages.

HTTP is the protocol that browsers and web applications speak in order to know how exactly to formulate these messages. If you’re not familiar with the term “protocol”, it means a clearly specified language, a set of formal rules for how to talk to each other.

When you open your browser, type a URL into the address bar, and hit return, your browser will send a text message to another computer associated with the hostname that is part of your URL. The computer on the other side will receive this message (called the request), and send another message back (called the response).

For example, when you go to this is (a simplified version of) the request that is going to be sent by your browser:

GET /location.html HTTP/1.1
User-Agent: Mozilla/5.0

In essence, this message says “Get me the page /location.html”. Simple as that.

The webserver that hosts our homepage then looks up the content of that page, and sends another message back to your browser containing the content of the page, which is, of course, HTML:

HTTP/1.1 200 OK
Content-Type: text/html

    <title>Ruby Monstas - Where we meet</title>

That’s it. Your browser will extract the HTML that is sent as the body of the message and display it.

Let’s look at the request and response in more detail. What does all of this stuff mean exactly?

An HTTP Request

HTTP defines that the first line of a request must contain three bits: a verb, a resource, and the version of the protocol. Out of these, we only really care about the first two bits. The version is only relevant in way more advanced contexts.

The verb, also called the “method”, defines an operation. GET of course means just that: “Get me this thing, please.” The “thing” we’re looking for is defined by the second bit on this line: the resource, also called the path. In our case that’s a web page, i.e. some HTML.

Other common verbs are [1]

HTTP also defines that after this first line there may come any number of headers, i.e. key/value pairs, containing meta information about the request.

In our example these are the hostname that was used in the URL, and the (simplified) name of the browser that was used to send the request. Both of these actually are mostly for information, and not a core part of the request.

An HTTP Response

Now, let’s have a look at the response.

Again, HTTP defines that on the first line there need to be three bits of information: The HTTP version number, a status code number, and a status description. And again, we can just ignore the version number.

The status code is a way for the responding server to tell our browser what type of response this is. Here are some common examples [2]:

The status code indicates the type of response, and most of the time, for most GET requests, it will be 200 OK.

On the following lines, again, there will be some headers: key/value pairs that contain meta information about the response. In our case there is a header that says the content type of the body is text/html (which just means it is HTML).

Now, when there’s a body (“main content”) on a request (sometimes requests come with a body, too) or response (as in our example), then it needs to be separated from the header with an extra, blank line. This tells the browser that we’re done with the headers, and the rest of the message is the response body (the actual content).


To summarize:

There is, of course, a lot more to say about HTTP, but these are the essentials of how computers talk to each other when they speak HTTP, as browsers and webservers do. And these also are the basics of what you need to know in order to build a simple web application, receive requests, and send back responses.

If you are interested in more details, here is a presentation that Konstantin Haase, maintainer of Sinatra, gave for Ruby Monstas in early 2015. Also, the Wikipedia page about HTTP is an interesting read.


[1] There’s a lot more to be said about HTTP verbs than this, and some developers might argue that these explanations are wrong. Let’s say these verbs mean these things in applications like Rails applications. The HTTP standard defines these verbs in more abstract, and generic terms, while Rails has its own, more practical definition of these HTTP verbs.

E.g. POST actually means more “do whatever you think is right” in terms of the HTTP specification, while in the context of Rails it means “create this thing”.

For us, in practice, for the time being, it is enough to remember that GET means “get”, POST means “create”, PUT means “update”, and DELETE means “delete”.

[2] There are lots of other status codes that servers can respond with, here’s a list. Many of them are very rarely used, and maybe the funniest one is 418 I'm a teapot.