As mentioned earlier, information architecture is broadly defined as a model (or the practice of building the model) for an information space, or a set of information that is organized and managed together. That model describes the rules for how that information should be maintained, interlinked, accessed and presented. To describe an IA more accurately, though, you need to use a few information architecture concepts.
Rosenfeld and Morville are authors of a book titled "Information Architecture for the World Wide Web." The cover of the book, like other O'Reilly publications, features a drawing of an animal. Due to its popularity in the IA community, the book is known primarily in reference to that animal as the polar bear book. Many IA professionals reference this book and its authors as authorities on defining and describing information architecture.
In the 2002 edition, the authors state that information architecture consists of three conceptual circles: content, users and context [source: Rosenfeld and Morville]. Content may include text, numerical data, images and videos. Users are the target audience for the information, with IA including the audience's experience and how audience members look for information (information-seeking behavior). Context encompasses goals and resources such as technology, company culture and politics.
During his presentation on information architecture at the University of North Carolina at Chapel Hill, Gary Marchionini, professor and dean of the university's School of Information and Library Science, defines a concept for the basic unit of information in an IA as a package [source: Marchionini]. He describes a package as a paragraph of text, an image, a video or some similar piece of data. Part of information architecture involves a plan for what packages can consist of and how they're managed and accessed within a larger information system. For convenience, this article will use Marchionini's term in the same way.
Another concept in information architecture is attributes. Attributes are the details we use to describe something, such as describing a person by height, weight or gender. Information can have attributes that describe its packages or the objects within that package. These could be physical attributes such as the number of characters in a paragraph, or abstract attributes such as the appropriate context for displaying that paragraph. For an information architecture to work efficiently, attributes should be applied consistently to packages and their contents. The IA might also include logic for how attributes relate to each other.