In the brief history of interactive media, popular design practices have evolved through two generations, and can be said to be just entering a third. Roughly, these can be broken out as:

• 1st Generation: Product-Centered Design. During the birth of interactive media, information was typically organized around the corporation's own view of the product (or service). This view varied with the nature of the corporation and their own approach to communications: sometimes it represented the messaging "push" developed by a marketing department; occasionally it reflected the perspective of technical personnel who viewed the product in ways that were difficult for users to understand. In less fortunate cases, it was completely ad hoc. In any of these cases, product myopia frequently generated interfaces that left users lost and frustrated.
  
  Although not absolutely concurrent, product-centered design and experimental, often complex, interface devices frequently occurred together. Unhelpful and distracting animation, hidden navigation models, and lack of user-driven content conspired to create high dropout rates.
  
  
• 2nd Generation: Information Architecture. As awareness grew of the shortcomings of product-centered design, Information Architecture arose as a more objective approach to dealing with organization, particularly in content-intensive applications. The science of IA remains a critical discipline for working with large bodies of content, even into the 3rd generation of UI modeling.
  
  As a driving philosophy to UI design however, IA (as commonly practiced) has its own shortcomings. "Purely logical" taxonomies, while not reflective of the idiosyncrasies of an individual marketing or product manager, may also not reflect the perspective or goals of your users. IA seeks to distill and generalize, while simple empirical investigation may produce more meaningful results.
  
  I've seen this take a couple of common forms. One is the misguided belief that there is one "right" taxonomy for any body of information. In fact, personal, cultural, and temporal contexts can lead different individuals or groups to have very different ideas of how a given body of information might be organized. There is also a temptation to put too much stock in generalized behavioral models when it is not necessary to do so. One Information Architect, working on a site to help users find information about nightlife in a major US city, sought to incorporate human factors in this way. The proposed approach included reviews of cognitive psychology, computational modeling of human categorization ability, and neuroscience. Why not just work directly with real users to discover how they would look for the information?
  
  
• 3rd Generation: User-Centered Design. Although the principles and practices of user-centered design predate the Web and most other interactive media, they were not widely applied during the Internet explosion.
  
  User-centered design comprises a set of practices that keep the user's needs at the center of the design process. It involves in-person discovery of the nature of the users and their goals and tasks, design that responds to those, and very importantly, scientific validation that the design works, by testing prototypes with actual users. Iterating these practices appropriately throughout the design cycle is the proven best practice to minimize user failure or dropout.
  
  In the rush to market, usability milestones are a frequent target of schedule or budget compression. However, this is the by far the most time- and cost-effective point in the process to make the critical discoveries available through user-centered design practices. When these steps are bypassed, crucial flaws that prevent user success are frequently uncovered only after expensive development, sometimes creating much greater delays in time to market.

All three of these perspectives are actually valuable input into robust design process. It is only when considered as driving design philosophies that they can be viewed in this generational model. But it is interesting that just as in the process of designing a particular project, the simplest and seemingly obvious solution is not the first approach that presents itself.