I recently unearthed an interesting IA fossil: a review of the first edition of Information Architecture for the World Wide Web, the venerable polar bear book. The post isn’t dated, but it seems contemporary to the book’s publication. (I.e., late 1990s.) The author, Tom Wilson, seems to have liked the book. He concludes:
This is a useful introduction to a complex area of work in the field of Web site design, and it fills a gap. Some of the same ground is covered by general works on the subject and in more detail by …, but the beginner will find this introduction invaluable — and perhaps it will stimulate other students at other departments of information studies to look for a career in the field. Before they do, however, they had best be sure that the teaching of classification and indexing in their school is up to it!
A few thoughts.
First, the page is typical of its time: it features a minimalistic table-based layout, only one visual element, no ads, text set in default OS fonts (Times New Roman and Verdana), blue underlined hyperlinks, and the obligatory XHTML 1.0 badge. It’s obviously an artifact from a time before broadband. As such, it feels simultaneously ancient and crude but also more respectful of my time and attention than most current web pages. The fact it can still be read as intended in a modern web browser is a testament to the resilience of simple HTML.
Second, the review reminded me of what it was like to design for the web in the late 1990s when many of the practices and norms we take for granted hadn’t yet been codified. For example, the page doesn’t feature a clear publication name. Today, we’re used to web pages prominently displaying the site’s name and primary navigation structure at the top of the page. However, this page only has a (minimal!) navigation bar set in a column on the left side of the page. The critical upper-left corner, where modern web users expect a logo, is blank. (Perhaps this page was set in a frame?)
Websites in the late 1990s were also simpler in other ways. Back then, much of the web consisted of static HTML pages, such as this review, many of them hand-coded. Sites were also smaller than today — but still larger than many other publication types. Organizing web content was a novel challenge and more of a top-down activity than today — i.e., it was possible to define a single high-level categorization scheme to serve as the primary means for users to find stuff. (Note there’s no search box on this page. Again, perhaps there’s a missing frame here — but local search was rare at the time.)
This was the context in which the polar bear book appeared. Wilson grapples with the (then unfamiliar) key phrase in the book’s title:
Information architecture, of course, is a computer science buzz-word and useful in selling the book to the intended audience — it is likely to have more impact than Classfication and indexing as tools for the organization of Web sites!
(Emphasis in the original.)
The notion that you could “architect” a website for improved findability and usability was exhilarating back then. Applying lessons from library and information science to these new organizational challenges was a significant contribution.
Third, implicit in the review — and in the zeitgeist of the time — is the notion that websites are publications as opposed to applications, a distinction clarified by Jesse James Garrett’s Elements of User Experience diagram (pdf). When Wilson writes “the organization of Web sites,” I expect he’s mainly thinking of collections of static content-heavy web pages, which were the norm in the 1990s.
In contrast, today’s web is highly dynamic. Consider Twitter: each tweet is a content item mainly consisting of text, not unlike early web pages. Individual tweets are also addressable, again, not unlike individual web pages. But that’s where the similarities end. You experience tweets in a context that is utterly unlike an early website: they flow by in a fast-moving stream that looks different for each system user. (You could argue that many people experience Twitter via apps, so it isn’t a website; I’d disagree.) Twitter users can also arbitrarily categorize tweets by including hashtags, which produce an emergent (i.e., bottom-up) structure at a scale unimaginable in the 1990s.
But the fact that it’s dynamic and personalized and used by hundreds of millions of authors doesn’t imply a system like Twitter has no need for architecture. On the contrary, the modern web requires more thoughtful IA than that of the late 1990s. Those hashtags won’t be of much use if the system doesn’t provide effective means for users to interact with them, and someone must design the algorithms that determine the order in which content items appear. (At this point, it’s clear that optimizing solely for engagement is a corrosive strategy.)
And, of course, the web is more central to society today than in the 1990s. The structure of online systems determines our wellbeing in ways unimaginable then. Consider the state of our political discourse or misinformation surrounding vaccines. The web is the central medium in both cases. It behooves us to structure web-based systems more mindfully, looking not just to drive engagement but also (or even primarily) to build understanding and knowledge.
Information architecture — and the polar bear book itself — has evolved since 1998. Today, much of my work focuses on dynamic systems instead of static publications, and I suspect the same is true for many of my colleagues. But the core principles of architecting information for findability and understandability are as relevant today as they were in 1998. As with so many things, much of our job isn’t about inventing new means for structuring information but applying timeless principles to the needs of the time.