home ....interface design .... web development .... movie reviews .... travel .... about peterme
ASIS&T 2001 Summit, Feb 3-4
Reflections and Projections Panel

Sections
1. 

Intro

2.  Trends
- Bottom-up
- It's The Content, Stupid
- Models in Design
- Working with Clients, Not At Them
3.  Future of Information Architecture
 

- The Spread of "Good IA"
- Data Analysis
- IA Playing Nice in the Sandbox
- Library Science Impacts Agency Information Architecture
- Data-Driven Information Architecture
- Professional Affiliations for Information Architecture
- Further Specialization

4.  Pain
  - Business Savvy
- Ill-defined Roles and Responsibilities
 
Trends
Bottom-Up
Spreading like wildfire throughout the community is the idea of developing systems from the bottom-up, designing bits and pieces to work together in myriad ways. The word "modularity" recurs in people's discussions, whether it was Karen McGrane (from Razorfish) at ASIS 2000 Summit, or Phil Oye (Sapient) and Alex Wright (Liquid Thinking) at this year's summit. Also at this summit, Bob Boiko discussed a "component" approach to content management systems. Out in the world, Jeffrey Veen promotes design patterns in his latest Web design book, and even the New York Times has published articles on self-organized sites.

The problem is clear--the systems we're tasked to design and work with are overwhelmingly complex. Whereas top-down solutions were feasible in the "early," more static, days of the Web, the dynamic database-driven sites of today require a flexible design architecture.

For design modularity, the analogy most often used is "Legos." A design Holy Grail is to create an interlocking set of interface modules that you can cobble together as needed to solve particular problems. We attempted this at Epinions (and partly succeeded), and it's become a goal at Walmart.com.

Less prevalent, but growing in awareness, is applying bottom-up strategies for managing information, particularly using people's actions as a way to sort content to ensure the most relevant and worthwhile documents are featured. Epinions, Plastic, Slashdot, and a few others use this model.

 
It's The Content, Stupid

The phrase "Content is king" has been applied to the Web since its public reception. This typically meant that the Web was a content-based medium like newspapers, magazine, or television--that people were using the Web to find out information.

A trend in recent thinking has been to acknowledge that content isn't the end, but the means. This is particularly true with transactional sites, like online stores. Content isn't important in and of itself, but it serves as impetus for action. Jared Spool addressed this in his keynote, "It's All About Content," discussing the findings of his extensive user researching. Peter Morville talked about it in his learner-centered design presentation, holding up REI and Wine.com as exemplars of actionable information. Research at Epinions showed that testing the interface was misguided--it was the content, the reviews, that people acted on, and we needed to figure out how to make it useful.

 
Models in Design

This trend is less well-established, but I think we're seeing the beginnings of a movement in the importance of developing conceptual models in design. Only one person discussed them at the Summit--Alex Wright showed user models and system models that were developed in the design of Spotlife.com. The premiere issue of Gain, a journal on experience design published by the AIGA, features a long article on Hugh Dubberly (.PDF) and how his design work has moved away from designing artifacts towards designing process and explanation models. (Hugh's work is also featured in Richard Saul Wurman's Information Anxiety 2.)

My experience is that such models are the unsung heroes of systems design. When we were redesigning the Epinions site, we started with a very bottom-up approach. We knew all the little problems we wanted to solve, the features we wanted to offer. But we didn't have a Big Picture view of the site experience--we were getting lost in the weeds. My manager (the VP of Product Development) encouraged me to "come up with the metaphor." Poking around a bit, I did some research into developing conceptual models, and came up with some drawings to model user behavior, and the user's "view" of the site. Though quite high-level and abstract, they were remarkably helpful in cohering our approach to the design.

 
Working With Clients, Not At Them
A refreshing trend that came up again and again was how agencies were moving towards a "one team" approach with their clients, working with them side-by-side on solving problems. When I worked at Studio Archetype, that company's graphic design roots definitely led to an "us-and-them" approach with clients, where we'd hear requirements, run back to the office, make stuff, present that to the client, get their feedback, iterate, etc. While that approach works fine for graphic design, the communication loop is abysmally slow when dealing with the complexities of building web sites. If you're agency continues to approach problems in that outmoded fashion, you'll produce sub-par work and take more time to do it.