dot Stop testing, start deploying your AI apps. See how with MIT Technology Review’s latest research.

Download now

6.1 Autocomplete

back to home

6.1 Autocomplete

In the web world, autocomplete is a method that allows us to quickly look up things that we want to find without searching. Generally, it works by taking the letters that we’re typing and finding all words that start with those letters. Some autocomplete tools will even let us type the beginning of a phrase and finish the phrase for us. As an example, autocomplete in Google’s search shows us that Betty White’s SNL appearance is still popular, even years later (which is no surprise—she’s a firecracker). It shows us the URLs we’ve recently visited and want to revisit when we type in the address bar, and it helps us remember login names. All of these functions and more are built to help us access information faster. Some of them, like Google’s search box, are backed by many terabytes of remote information. Others, like our browser history and login boxes, are backed by much smaller local databases. But they all get us what we want with less work.

We’ll build two different types of autocomplete in this section. The first uses lists to remember the most recent 100 contacts that a user has communicated with, trying to minimize memory use. Our second autocomplete offers better performance and scalability for larger lists, but uses more memory per list. They differ in their structure, the methods used, and the time it takes for the operations to complete. Let’s first start with an autocomplete for recent contacts.