architect or librarian?

architect or librarian?

In Release 2.0, Esther Dyson argues that large organizations are at a natural disadvantage in using the Web. Committees and corporate reviews slow their reactions, delaying even the smallest and most urgent changes. Levels of hierarchy insulate decision-makers from direct contact with the Web audience. Competing interests make it difficult to create focused Web sites, and internal struggles make it easy and tempting to design sites to appeal chiefly to corporate offices, not customers.

Information Architecture for the World Wide Web is a handbook for addressing precisely these problems, the challenges that confront large organizations on the Web. The authors describe themselves as consultants "for some of the largest companies in the world", and their perspective on Web development is organizational and procedural. Much of the book explores the process of leading corporate committees toward a consensus on Web site design. Note these chapter headings:

The authors give tips on running meetings, on handling interdepartmental disputes, and on managing teams of writers, programmers, and designers. Indeed, they assume that the detail work -- writing, programming, graphic design -- will all be performed by specialists working under their direction. These specialists must, they argue, be firmly directed and controlled.

The book's strongest chapter, on Labeling Systems, is almost pure library science.

Rosenfeld and Morville are librarians by training, and the "information architect" they envision is really a librarian. The information architect is concerned with cataloging, signage, and search, with making sure that each item in the inventory is stored in the proper place, and that it is described in the proper way. The book's strongest chapter, on Labeling Systems, is almost pure library science. Haphazard categories, they argue, are inefficient and distracting. On many sites, some sections are labeled by content, others by ownership, others by intended audience:

bad better
Faculty Skunkworks
Office for Technology
K12 PDN Projects Web Page
Digital Libraries Project
The New Media Center
Arts & Humanities
Places & Staff

Consistent labeling requires care and attention. Labels should reflect the readers' needs, not the company organization chart. Labels should be grammatically consistent; if some labels are nouns and others are verbs, readers will be confused. Labels should be drawn from a controlled vocabulary; haphazard mixing of synonyms simply misleads readers and confuses search engines.

Though the authors' background as librarians is their strength, it also weakens the book. Cataloging and searching are what Rosenfeld and Morville do best; once the reader gets to the "right" place, the authors lose interest. Navigation is regarded to search engines and indexes. The book's skepticism toward "cool" Web design has drawn many accolades, but misses a key point: Web sites are strongest when everything -- graphic design, organization, writing, and hypertext structure -- serves the same end.

Beyond the supervisory power of the job title, Rosenfeld and Morville aren't very interested in architecture.

It is odd that Rosenfeld and Morville seize the title of architect, because the central claim of the architect's profession is the very breadth of concern that Information Architecture lacks. Architects have always competed with craftsmen, construction firms, and engineers; what architects offer is an original and coherent vision that inspires and entire Web site or building. Beyond the supervisory power of the job title, Rosenfeld and Morville aren't very interested in architecture. The book is silent, for example, on spatial hypertext, and says nothing about the sudden prominence of pattern languages -- an architectural idea -- for understanding hypertext structures. On the few occasions they mention architecture, their similes are superficial and indifferently informed:

Some architectures disgust us. Ask someone who owns a house with a flat roof how they feel about its architecture.... Why do bad architectures happen so often? Because their architects generally don't live or work in the buildings they design. That hardly seems fair. The same is true for so many Web sites....Don't these Web sites' architects ever use their own sites?

Even the casual observer of architecture will see the flaws here. First, flat roofs are not always a bad idea, even in rainy climates. I grew up in a flat-roofed house in the Windy City -- an undistinguished, and beat-up example of early modernism. Most of my playmates lived in flat-roofed houses, too. Although I did wish my parents would let us play on the roof, the roof's other shortcomings didn't bother us terribly. Architects, even those who seem least user-centered, are notoriously prone to live and work in spaces they design. Frank Lloyd Wright, Walter Gropius, Mies van der Rohe -- all famously designed their own quarters. Wright's chairs may be uncomfortable, but he sat in them; Mies' windows may not open, but he sat beside them.

One legacy Rosenfeld and Morville do borrow from architectural rhetoric is the unfortunate habit of confounding error with idiocy. It's not enough to note that some aspects of a Web site might be improved; if a page contains too many links, or if it buries contact information too deeply, we must feel disgust. It's not enough to interview clients about their goals; we must explore what they love and hate about the Web. A main page with thirty links is horrible. Information Architecture deserves praise for avoiding the inordinate emphasis on visual design that dominates Siegel's Creating Killer Web Sites, but even Siegel, who often argues from other creator's design mistakes, is far more charitable.

There is little of architectural interest to be found in Information Architectures. Its rhetorical excesses are sometimes irritating. Like other recent guides to site design, it withholds data from actual experience in favor of casual speculation. The authors note, for example, that they pioneered a site that anticipated key features of Yahoo, yet instead of explaining why Yahoo overtook them, the jokingly speculate that they should have chosen a better acronym.

Lessons from library science are worth learning.

Still, at its core, this book has useful ideas and valuable insights. The chapter on Labels is particularly novel, innovative, and useful. All designers of large Web sites need to be familiar with the indexing and cataloging skills that Rosenfeld and Morville explain. Lessons from library science are worth learning, and the author's preference for the magisterial title "Information Architect" over the more familiar but still honorable title of "Librarian" should not blind us to the lessons we can learn from the traditions of the library.

Copyright © 2005 by Eastgate Systems, Inc. All Rights Reserved.

Eastgate Systems, Inc. 134 Main Street, Watertown MA 02472 USA.

Email: Phone: (617) 924-9044 (800) 562-1638