One question that’s sure to come up during this year’s information architecture summit is, “What is the role of IA in Web 2.0?” Some people have proposed that IA–creating meaningful structures of information–is obsolete. A key component of many Web 2.0 sites is that people can impose their own structures on information through tagging. In some people’s minds, this spells gloom and doom for our field.

Information architecture has long been associated with navigation–finding one’s way through space. In the early days, information architects were responsible for generating the list of categories along the left-hand side of many a web site. They thought about how all the information on a web site fit together, and where there were meaningful relationships. They collaborated with designers and others to shape the virtual space of a web site.

Frequently, information architecture was compared to fields like urban planning. Urban planners shape city space. They identify the goals of pedestrians and other people using the space and create designs to support those goals. This was a comfortable way to frame information architecture. If the web is a virtual space, it needs its own urban planners.

The metaphor WEB IS SPACE has been entrenched since the beginning. The linguistic evidence is overwhelming: The basic currency of the web is a site, which is located at an address. Even technical terms, like URL, participate in this metaphor. What do you think the L stands for? Location. Even words like browsing and searching derive from seeing the web as a physical location.

In discussing this with James last night, he pointed out that this trope was inevitable. Long before the web became popular, the same people who would make it what it is today were reading books like Neuromancer and Snowcrash, which envisioned a virtual space populated by projections of our real selves.

When the naysayers say there is no longer a place for information architecture in web development, they are still conceiving information architects as shapers of space. But as new frames for the web emerge, so too must new ways of understanding information architecture.

Web 2.0 is, in one sense, shorthand for the idea that some web sites challenge the WEB IS SPACE metaphor. What makes any of the sites profiled on TechCrunch special is that they don’t think of the web as virtual space.

Before getting to new a new metaphor for the web, I want to establish a framework:

Web Metaphor => Primary Concern => Role of Structure

In English: By framing the web with a particular metaphor, certain concepts are established as the main unit of currency, so to speak. Those concepts, in turn, force us to think about structure in a particular way. In the case of Web 1.0, the framework would play out like this:

WEB IS SPACE => Location => Navigation & Wayfinding

When we think of the web as physical space, the primary concern becomes location, and the role of structure with respect to location is navigation.

While I haven’t done a complete linguistic analysis of Web 2.0 writings, it’s obvious that there’s no clear-cut definition yet. I’ve started compiling links of definitions, but each person brings a slightly different perspective to this shift. In doing a linguistic analysis, what’s important is not just what words people use to describe the phenomenon, but how they reason about it. With WEB IS SPACE, we don’t just simply apply the same words to web and space, we actually conceive of the web as a physical place.

Initial analysis points to a more complex conception of the web, revolving around two central models. Tim O’Reilly addresses both of them in his seminal paper “What is Web 2.0: Design Patterns and Business Models for the Next Generation of Software.” WEB AS PLATFORM positions the web not as a virtual space, but as a foundation, much in the same way an operating system like Windows or Linux is a foundation for software.

WEB AS NETWORK is the other cognitive model we’ve been using to reframe the web. O’Reilly addresses this in the second section of the paper, though he refers to it as “collective intelligence”. This model makes salient the connections between web sites, through simple links or some more complicated mash-up.

So far, everything I’ve read about Web 2.0 conceives of the web in one of these two frameworks, or some mix of both. In a sense, these models have always existed, and O’Reilly points out that people have been building sites that exemplified these models all along. I won’t speculate about why this shift is happening, but it would appear that a little Gladwell-esqe Tipping Point action isn’t out of the question.

O’Reilly discusses the business implications of both these models in-depth. There has been a ton of discussion about these models, especially the best way to leverage them to make money. The chief concern of my endeavor is not how to make money, but whether there is a role for information architecture in this new way of thinking about the web. (Of course, for information architects, it’s very much about making money.)

To put one of these models in the framework I mentioned above:

WEB AS NETWORK => Participation => ???

It’s clear from the various writings about Web 2.0 that participation is the primary concern when framing the web as a network. The size of the network and the ability for the network to engage participants become very salient. Whereas a key metric for the WEB IS SPACE model is number of visitors (the more people who come to your location, the better), the WEB AS NETWORK model would spin this differently, looking at level of participation. With many Web 2.0 sites looking to appeal to “edge” communities, quantifying participation is more interesting because they are going after smaller and smaller populations.

WEB AS NETWORK services generally don’t limit participation, but they erect a structure to make it meaningful. Look at del.icio.us or Flickr–quintessential Web 2.0/networking sites–and it’s clear that there are highly structured experiences. The designers behind these sites made conscious decisions about what it means to participate–not to exclude people, but to make the participation meaningful. You can post a URL or a photo to either of these sites without tagging or describing them, and yet they’ll still be attached to a user account and have a date. These simple contributions won’t be as meaningful as one of these items posted with tags, but participation does require a minimum amount of information.

Wikipedia is another great example of WEB AS NETWORK. It thrives on participation, but that participation is only meaningful within the context of rules established by the site. Some of these rules are business rules, but some are structural. O’Reilly’s article includes a sidebar with the tantalizing title “The Architecture of Participation.” He’s talking about technical architecture, but perhaps some of his points relate to experiential architecture as well. In this excerpt, he’s actually talking about the open source software movement, and how it provides a good parallel to what’s happening on the web:

We could restate the first phrase like this:

When we frame the web as a network that values participation, there is a clear need for structure to lend meaning to participation. Information architects must lead the efforts to define that structure, and collaborate with graphic designers and interaction designers to make sure that structure is clearly communicated to users. The problem is, I don’t know what this is called. We have convenient terms for creating structures in space–navigation, wayfinding, etc.–but not for creating abstract structures in other “environments”. Therefore, I can’t fill in the rest of the model, at least with a clever word or phrase:

WEB AS NETWORK => Participation => Structuring Participation

The second model of the web, WEB AS PLATFORM, has been more difficult for me to nail down. In many ways, it resembles the WEB IS SPACE model, because platform, in this sense, refers to an environment in which software is constructed, like an operating system. Like Windows or OS X, a platform is more than a space, though. It’s a set of rules, even a culture, establishing boundaries around development.

This environment, in O’Reilly’s words, is “without an owner, tied together by a set of protocols, open standards and agreements for cooperation.” Del.icio.us provides a good example of this frame as well, exposing all their data in an easy-to-parse format. It appears the primary concern for WEB AS PLATFORM is the format of the data, and value is measured by how easy it is to appropriate the content for some other use. In this frame, highly valued sites are also those that make effective use of other people’s data (usually Google Maps).

The role of structure in this frame is also difficult to pin down, and information architects generally don’t play in the space of data architecture, though they may collaborate with those who do. Still, I wonder if can get more involved in designing APIs and data feeds, especially if they have a role in designing the underlying structure of participation. Expressed in the framework:

WEB AS PLATFORM => Data Sharing => Data Formats

Sites that exemplify these new frames are not without their critics. Scott Karp, who writes Publishing 2.0, indicates that liberal participation from a broad audience can lead to meaningless data: “If you have a random group of people act as a filter, you’re going to get a random result.” To put it in terms of this new frame, participation demands structure, because without it Karp’s “sea of information” just becomes more turbulent. (Isn’t it interesting that someone writing about the new age of publishing still refers to information in terms of navigable space? I don’t think we’ll ever lose that trope.)

Next steps: Like my work in deconstructing content management, these frames offer rich opportunities for further research. I’ll continue to collect linguistic evidence and try to flesh the models further. At this point, my original question has not been answered in full. That is, initial analysis indicates that IA may place some role in the web as it’s been reconceived through these models, but the precise nature of that role is still up for discussion. What I’m most interested in is elaborating on the framework (model => primary concern => role of structure) and further clarifying structuring activities in the contexts of these models.