Chapter 1.
An Information Network Paradigm

"In the Information Age, the computer is an appliance and learning is a lifelong occupation."
Anonymous

It may now be possible to construct a "new information network paradigm" where computer networks supply not only connectivity and transportation, but also information services. Like ships and railroads in the nineteenth century, like automobiles and airlines in the twentieth century, information highways will drive economic wealth in the twenty-first century.

Driven by the popularity of the World Wide Web (Web), the Internet boomed in the 1990's, growing from 500,000 hosts to ten million. Major internet providers report growth in subscribers of fifteen percent per month. It is estimated that the number worldwide exceeds forty million. Future growth may be difficult to predict, but the potential is great and the benefits greater. Information networks will deliver information, network intelligence, services, and applications seemlessly and transparently to a large and diverse world-wide user base.

From mainframes, to minicomputers, to personal computers, this industry is poised to take another step forward. This thesis does not suggest that these legacy systems will be totally replaced by a new paradigm. The market for these systems will continue to grow. The almost universal acceptance of the Java virtual machine means that these legacy systems can interoperate in ways not previously possible. With the most to gain, IBM acted quickly to adopt the Java programming language--with its develop once, deploy anywhere advantage.

Hanafy Meleis(1) traces three phases in the development of computer networks and outlines the requirements for a new information network paradigm. "Over the past ten years information networking has evolved from centralized mainframe databases containing numbers and simple text, to distributed client-server databases of richer text and images. In the next decade, many multimedia information networks will be deployed to enable new commercial applications such as virtual corporations and distance learning. The guiding vision for the future information network is a system that lets users access all the information they need, when and where they need it."

To realise the promise of information networks, computer technology needs to be available to an ever increasing base of users. I agree with Dr. Meleis that user access and usability is a potential technical barrier to the acceptance of information networks. Intuitive, interoperable user interfaces are required along with generic tools for authoring and presenting multimedia information. Users should be able to obtain services without concern for architectural details or the location of information. Ubiquitous Software is a method to hide these complexities.

Meleis' three phases of network infrastructure evolution are:

1. Host-based networks: Characteristics: circuit-based, protocol-dependent, centralized, connection-oriented. Applications: Mission critical, constant use, low bandwidth, some quality of service (QoS).
2. Client-server networks: Characteristics: router-based, protocol-independent, distributed, connectionless. Applications: Departmental applications, sporadic use/bursty, medium bandwidth, few QoS constraints.
3. Information networks: Characteristics: switch-based, protocol-independent, collaborative, connection and connectionless. Applications: Collaborative applications, constant use/bursty, high bandwidth, variable QoS constraints.

Meleis makes note of three technology barriers that must be addressed to realize an information network.

1. Operations and management: A large and often unrecognized barrier to the exploitation of information networks is the lack of cost-effective solutions for operations and management. In 1994, US labor costs associated with the operation and management of private information networks was estimated to be more than $200 billion.(2) The information infrastructure of the future will provide not only connectivity but information services as well. A paradigm shift will be necessary for the operation and management of the future information network.
2. User access and usability: Lack of a common user interface could turn the vision of an information services network into a nightmare if users are unable or unwilling to use new services and products. Six issues need to be considered.
3. Digital video: This will by no means be routine or inexpensive. Requirements are: infrastructure capacity, interoperability, tools for creation and management.

This thesis proposes Ubiquitous Software (U_S) as one solution to the user access and usability (computer user interface, CUI) issues described by Meleis. He points to six areas that need to be addresed by software engineering.

1. Intuitive, interoperable user interfaces: Tomorrow's user interface must improve on GUIs, integrating multimedia search with text and database queries. It must also support all potential interface techniques, including keyboard, mouse, voice, and pen.
2. Intelligent agents: The huge amount of information available will require intelligent agents and filters that can learn user preferences, deal with multimedia data, and use standard methods to capture, represent, and communicate information.
3. Information fusion: For the information to be useful, content-retrieval technology must combine the results, in a meaningful way, from query servers running different search engines against very large multimedia collections.
4. Information preparation: Generic tools for authoring and presenting multimedia information should enable nonspecialist users to handle information from a variety of sources.
5. Authentication and security: The information network will require security and authentication techniques that are scalable and applicable to heterogeneous environments. Techniques for document authentication and copy protection must be further developed and applied to multimedia objects. All forms of payment, including subscriptions and pay-per-use, should be supported in a common framework.
6. Service portability and transparency: This heterogeneous, multisupplier network must hide its complexity from the user. The user should be able to obtain services without concern for the details of the networking environment, the geographic location of the information sources, or other participants.

The CUI issues addressed by U_S target points 1, 4, and 6 above. document-centric software offers an intuitive, interoperable user interface. A hypertext writing space makes possible nonspecialist authoring environments for presenting multimedia information. Object-oriented software built from vitual libraries of reusable components hide network complexity by combining idenity, behavior, and state. U_S is a paradigm built on top of the emerging information networks. The goal is ease of use similar to telephone networks.

U_S is pronounced with the letters U and S. The underscore is used to introduce a brief pause between letters so as not to be confused with US or us. Emphasis is placed on the U.

This thesis is organized as follows:

Chapter 2 provides rationale for another software engineering paradigm. document-centric software is proposed to provide an initiative CUI.

Chapter 3 lists the tools and technologies to construct a developer environment. U_S is objected-oriented. document-centric, component-based, and network-aware. IBM's OpenDoc and Microsoft's OLE are examples of competing document-centric container-based environments. Sun's Java virtual machine shows the importance of interoperability. And JavaBeans can build object oriented reusable components deliverable using a pervasive network.

Chapter 4 addresses related research. And to review the Network Computer (NCs) and its implied architecture-neutral software. The Java virtual machine is an important element of the first round of NCs and paves the way for building software that understands how to display itself across different platforms.

The information network paradigm makes possible a fourth technology of writing. Chapter 5 describes a writing space where readers are writers and authors are programmers. The World Wide Web servers as a protype authoring environment. U_S turns data networks into information networks and makes possible an authoring environment that is a new Writing Space.

In the Information Age, education may be the most important industry. Chapter 6 discusses the benefits already accruing from the popularity of the Internet and the World Wide Web. Virtual lesson plans (VLPs) and education brokerages are presented as examples of how U_S can be used to implement the supporting infrastructure of an information network. VLPs is an authoring and presentation environment.

This thesis also recognizes that a U_S model is not a general software engineering model; for example it is not well suited for real time applications like fly by wire or for systems where quality of service issues are paramount.