The evolving information superhighway infrastructure links:

* Producers of information, including computer software, books, movies, music, etc.
* Producers of physical goods, including all manufactured goods now sold through catalogues and those in which computer technology can simplify product complexity and reduce asset specificity.
* Electronic retailers, differentiated into speciality retailers like Blockbuster or multi-product retailers like Lands' End, Sears, Macy's, etc.
* Electronic markets, expanded by market makers to include the travel and financial industries and specialty niches (e.g., shirts, personal computer software, or baseball cards).
* Physical distribution networks, simplified to move from the manufacturer to the consumer directly, or coordinated by electronic retailer or market maker transactions. The future delivery system might resemble the process that catalogue vendors use now, mostly via Federal Express or United Parcel Service. When next-day delivery is satisfactory, such companies can provide the desired service. When faster delivery is required, such as when ordering the week's groceries, variants like picking them up at the supermarket depot may emerge.
* Electronic channels, i.e., cable, telephone, cellular and electric utility industries that can provide access to the home. Although a choice of electronic channel is still limited, market dynamics are unfolding rapidly (e.g., wireless channels).
* The market choice box, where a vast amount of electronic commerce will be channeled and controlled. This is a server that manages the configuration of workstations, telephones and television sets in the home, and provides a telecommunication interface to those channels that directly reach the home (see Figure 3). The box will present a choice of markets and other activities such as entertainment, shopping, surfing the Internet, etc. It is anyone's guess how the primary graphical user interface (GUI) will be designed, although it is known that several corporations are working toward such devices. It is important not to bias the consumer to one choice over another, as the initial airline reservation system did for flight choice. Gilder (1993) suggested that the ideal GUI may be a newspaper format, an interface designed for human interaction that has indeed passed the test of time. For example, we read a headline, skip to sports, back to the financial page, advertisements catch our attention and we study them in greater detail in order to gain more product information. The primary GUI may also be adaptable to our life styles and it is conceivable that each class of market choice may require a GUI to help explore its potential. An example of such a possible market choice box may be General Magic's Telescript user interface. Such an interface would let the consumer put an interactive agent into the Travel Store on Main Street that would purchase flight tickets and act as a pseudo-electronic market potentially by-passing APOLLO, SABRE, and the travel agents. It is reasonable to speculate that user interface owners, such as General Magic, would like to appropriate a portion of the resulting value chain and market-maker savings rather than share them with the consumer. This scenario demonstrates easily how the market choice box and standards associated with market choice (labeling of catalogue items, marketing of client/server software for products such as Telescript) can all affect the openness of information and market access.