Photo: Dylan Love
For the maths-phobic out there, don’t let the word scare you. An “algorithm” is nothing more than a set of instructions, just like a recipe or how-to book.And the Internet relies on many, many algorithms in order to function properly.
When you type search terms into Google, it follows a very complex algorithm to determine which results to show you. When you buy something on Amazon with a credit card, it uses an algorithm to safely transmit your credit information. And just the simple act of navigating the Internet and loading a webpage requires the use of an algorithm.
From basic and essential to complex and useful, here are 11 of the many algorithms that make the Internet work.
This is a huge one. You probably use it several times a day without realising the complexity behind it. It's an algorithm that's constantly changing in order to bring you the best search results possible.
And it's VERY protected. It's the Internet's equivalent to the recipe for Coca-Cola.
Routing is the process of finding a path in a network to send traffic. A number of routing algorithms are out there with names like fuzzy routing, heuristic routing, and adaptive routing, but each one helps make network communication possible.
Routing is one of of the core technologies at the heart of the Internet -- no routing, no Internet.
Encryption is the process of turning information into unreadable nonsense characters. It's what makes Internet credit card transactions possible and it's at the heart of Internet security.
128-bit and 256-bit encryption algorithms are two of the most popular today. Prior to 1996, both were recognised as munitions by the US government -- they were literally illegal.
If you're the type with several hundred Facebook friends, your news feed would absolutely explode with numerous trivial updates from people throughout the day. Luckily Facebook has put some algorithms into effect that attempt to identify the most important updates and only show you those.
Sendmail is an email routing tool -- it's makes sure that your email gets where you want it to. It supports the Simple Mail Transfer Protocol (SMTP), which is used by almost every webmail service.
It actually depends on several algorithms to handle tasks like queue management and delivery processing.
Netflix already had a pretty good recommendation algorithm, but they wanted to make it even better. In 2009 they announced the Netflix Prize, a competition to see who could design a more successful algorithm that yielded higher customer ratings on recommended movies.
The team that won was able to improve ratings by 10% and they walked away with $1,000,000.
Hashes are essential to ensure that transmitted messages haven't been tampered with. The sender creates a 'hash' of a message and sends that along with the message itself. The person on the other end decrypts the message and the hash, produces another hash from the received message, and compares the two hashes.
If they're the same, then it's almost certain that the message hasn't been altered by a third party.
Intelligent advertisers will implement an algorithm to determine which ads are displayed on a page in order to make sure that they prove relevant and drive revenue.
Take Google's Adwords, for example -- its algorithm relies upon your search terms in order to show you ads related to your search.
Google's news algorithm looks at the volume of material that appears on a given topic and considers the number of other pages linking to it in order to determine if it should be displayed.
There are likely many more factors involved, but like Google's search algorithm, those are kept under wraps.
Going off of its nearly endless list of items customers have purchased together, Amazon can implement an algorithm to show you related products you might be interested in buying.
(It's so useful that I sometimes use it as a way to find new authors or bands.)
Pandora follows an algorithm to determine which songs to play for you after you tell it what you like. It pulls its data from the Music Genome Project, an effort to 'map' songs based on identifiable elements and group them with similar songs.
It worked so well that Pandora went public with a $3.2 billion market cap.