Louis Rosenfeld Information Architecture

Louis Rosenfeld Information Architecture – August 23, 2001: Future Directions for IA

As a brand new field, we information architects have spent the past few years absorbed with such typically adolescent pursuits as self-definition and self-justification. I don't know whether this soul-searching is necessary or completely pointless, but I think it's time to move on. At a certain point introspection becomes tedium.

So let's begin looking forward instead of inward: what will we be doing in ten years that's completely different? Or will information architects continue to see their jobs as being about creating wire frames and blueprints?

Please God, I hope not.

Here are some areas that I think information architects can and should take on. Some are orphans that may have gone unnoticed and therefore unowned. Some are simply not well understood by most people working on web sites today, including most IAs. Each of these areas presents us with difficult and interesting challenges that will increasingly demand our attention over time. And they fit squarely within the scope of information architecture:

  • Distinguishing users' information needs: “User” is practically the third word out of every IA's mouth. But we rarely discuss the types of information needs that drive users each time they use a web site or other information system. These needs, ranging from known-item (“I know what I'm looking for and where to find it”) to open-ended (“I'm not exactly sure what I want, and am really just here to learn”) to research (“give me everything on topic X”), are as important as any other factor to developing an appropriate information architecture. I'll go into greater detail in a future Bloug entry, but every IA should read Marcia Bates' seminal article “The design of browsing and berrypicking techniques for the online search interface” (Online Review 13, 5 (October, 1989) pp. 407-424) as an excellent introduction to information needs and how they vary (and no, it doesn't seem to be available online).
  • Determining content granularity: We're drowning in semi-structured content while dreaming of content reuse. The answer seems to lie in componentizing our content to a level more finely grained than documents. But is there a practical model for developing content components that us IAs can use in the real world? The answers probably lie in the worlds of XML and object modeling. But XML schemas don't always integrate well with other means of pulling meaning out of content, such as descriptive metadata. And object modeling approaches can be very data-centric and often don't apply to semi-structured text.
  • Developing hybrid architectures: It's inevitable that, as information systems accommodate greater volumes of information and manual indexing costs go up and up, we'll need to increase our reliance upon automated tools to help us find our way. And there are more options available than ever before, ranging from auto-classification tools to popularity engines. But we don't know a lot about which tools work the best in which situations, or how to combine different tools with more conventional manual approaches (and, for that matter, with other automated tools). Developing hybrid architectures which effectively mix manual and automated approaches is perhaps the main task we'll find ourselves grappling with for some time to come.
  • Presenting search results better: Do the familiar textual lists that use a handful of different sort orders cut it? How about clustered results? Could we benefit from applying information visualization techniques? And what content components should we show per result? As 2000+ results become more and more common, there's gotta be a better way than what we have now…
  • Understanding and using metadata: It was no accident that we Argonauts were so obsessed by controlled vocabularies and thesauri. They're increasingly necessary as a way to provide consistent and accurate representation of knowledge domains to both humans who are browsing and automated tools that augment searching. Many see IA as the structuring of information, but when are we going to start focusing on unlocking the meaning contained within those structures?
  • Rolling out enterprise-wide architectures: Ultimately all of the above have to make sense in the real world. Yet despite our frustration with organizational politics, few IAs have focused on creating a coherent model for developing and maintaining architectures within a large enterprise. So how do we do it? How do we balance the desires of the enterprise to centralize its processes and communications and cut costs with a suite of inexpensive technologies that makes every podunk department a publisher? How do we combine carrots and sticks to create a truly successful enterprise-wide information architecture? The answers will vary depending upon each enterprise's unique culture, meaning we IAs had better start learning about ethnography and organizational behavior.

What's missing from this list? What doesn't belong? And most importantly, does this list mesh with your vision for the future of information architecture? Maybe we can get a discussion going on SIGIA-L; I'll try to start one up. Bombs away …

Leave a comment