The State of Information Architecture

A large part of what we're doing at Pinyadda revolves around information architecture - the way content is organized on any given website. In many ways, the internet resembles the largest library ever created, harboring a collection of books, articles, photos, songs and videos in unfathomable quantity. Yet unlike the modern library, where every item is classified and sorted in a universal manner, the internet library is made up of millions of walled repositories, each with its own unique method of classifying content.

Classification is one of the most important parts of granting identity to any person or object - if I don't know what you are, it's nearly impossible for me to determine who you are. We classify things in our everyday life in a subconscious manner; it's how we make sense of the world around us. But if you've ever tried unsuccessfully to find a certain article on a certain webpage, you've experienced firsthand the limitations of the way things are classified on the internet. Was that article in the Culture section, or the Arts section? Maybe it was Entertainment? Or was it Local?

We have a few solutions to this problem, the first and most obvious being search, and most of the time it does the trick. You know what you're looking for, you type it in the box, and viola, up pops the article. Sometimes. Other times it's there, but it's buried 30 items deep in the results, and still other times it's not there at all. And most importantly, search requires that you knew what you were looking for in the first place.

What we're moving towards is a system of tags, like the ones at the bottom of this post, that allow us to label items instead of classifying them. This approach gives us a far more flexible way to categorize information while allowing it to remain organized and findable. By creating semantic relationships between tags, we can begin to infer implicit hierarchy instead of explicitly stating it. If all items tagged with "internet" are also tagged with "technology," but not all items tagged "technology" are also tagged "internet," we can begin to infer that the internet is commonly thought of as a subcategory of the Technology. But unlike defining a strict hierarchy that asserts Internet to be a subcategory of Technology and nothing else, inferring this relationship allows us to make reasonable assumptions without ruling anything out. Tags let us keep our options open.

But there's a catch. In order to infer useful relationships between tags, we need to place certain limits on their scope. A thousand tags loosely related to each other is far less useful than ten tags strongly related to one or two others. In the same way, a few tags used many times is far more useful than many tags used a few times. In order for us to identify patterns that make it easier for us to find information, we need to define the scope of our available tags. In short, too few tags limit different users' ability to find information, but too many render the system nearly useless . It's a balancing act, and a delicate one.

At Pinyadda, we feel strongly about creating a system that makes as few assumptions about individual users as possible. Everyone is different, and everyone's information consumption preferences are different. It's our job to make sure that we do as much to help you find the stuff you're looking for without getting in the way. The current state of information architecture on the web inhibits everyone's ability to get the most out of it, and it's a shame. The content you want is out there, somewhere. We hope we can bring a little method to the madness of finding it.