Design Principles for

Internet Community Information Gateways:

MARINFO - A Case Study for a Maritime Information Infrastructure

Rainer Müller, Claudia Niederée, Joachim W. Schmidt

Software Systems Institute

Technical University Hamburg-Harburg

{ra.mueller,c.niederee, j.w.schmidt}@tu-harburg.de

Abstract

The goal of the MARINFO project and system is to create the major information infrastructure for the maritime community - the maritime information gateway. MARINFO achieves this goal by fully exploiting modern information and communication technology as well as recent insights into community-oriented gateway design.

For our purposes, a community is defined by a common sphere of activities with commonalities defined by

-well-documented Resources on which the activities are based,

-shared Interaction patterns by which the community cooperates,

-domain Categories which characterize the community’s field of interest, and

-high demands for Homogeneity in function presentation and data visualization.

The MARINFO system follows these four principles - the RICH approach to gateway design - and provides enriched information and communication services which exploit

-typed interfaces to digital resources,

-internet-based secure interactions,

-extensive domain descriptions based on categories, and

-template-oriented visualization for overall presentation homogeneity.

Digital representation of resources, internet-based interaction and template-oriented visualization are the technical means by which systems and services are realized, customized and personalized and by which a high degree of extensibility, automation and overall efficiency is achieved.

Categories allow for the extensive and coherent structuring of domains along their essential dimensions - conceptual, organizational, temporal etc. - and provide unifying design guidelines for the entirety of system interfaces and services.

From a design point of view MARINFO is a case study in Internet community information gateways, from an implementation point of view it is an exercise in component-oriented software construction.

1Introduction and Overview

The MARINFO project has developed an Internet-based Information Infrastructure for the maritime community which, beyond that, aims at generalizing networked content and know-how management for a wider range of industrial communities. MARINFO is intended to become the prime site for the maritime community - the maritime information gateway - by fully exploiting modern information and communication technology and by deeply involving the community into the design and maintenance of the gateway’s functionality and content.

An information gateway provides “one-stop-shopping place” for a wide range of services and information resources in which the members of a community are interested. The goals of a well designed information gateway for the community are

-easy and efficient accessibility for the consumer to relevant high-quality information and services,

-improved communication between the distributed and heterogeneous members of the community, initialization and fostering of effective and successful cooperation between partners of the community,

-facilitated information exchange between community members,

-aimed contacts between information consumers and information providers as well as between future cooperation partners.

In addition to these consumer-centered goals there are additional goals from the perspective of the operator of the information gateway, e.g., the effective maintenance of the gateway and its content combined with a flexible quality control for the content which ensures the topicality, correctness, and attractiveness of the gateway.

The fulfillment of the described goals strongly depends on the design of the gateway. This applies to the range of offered services as well as to the strategies and technologies used for the realization of the gateway. In the MARINFO project, design principles for this purpose are examined for a concrete community, the maritime industries. The identified design principles are discussed in the next sections.

The MARINFO consortium is a balanced mix of technology providers and content providers that represent a broad spectrum of experience in all required aspects of information and communication technology as well as in the multiple facets of the application domain, its resources and its community.

The paper is structured as follows. Chapter 2 gives a resume of our R&D experience in community-oriented information gateway design and discusses design issues along the three dimensions of domain modeling, resource improvement and community interaction support. The overall structure of the MARINFO server and its services is discussed in chapter 3 along with selected examples for the three design dimensions mentioned above. Chapter 4 sketches the MARINFO overall software architecture, its interfaces and its major system components. Chapter 5 gives an overview of related work on information gateways. The paper concludes by summarizing our current view on commercializing issues and on related future activities.

2On Community Information Gateways: Overall Motivation and Principles

The prime motivation behind the MARINFO project and system is to create a major value-adding information infrastructure for the maritimecommunity - the maritime information gateway. MARINFO achieves its goal by fully exploiting modern information and communication technology as well as recent insights into community-oriented gateway design.

For our purposes, a community is defined by a common sphere of activities with commonalities defined by

-well-documented Resources on which the activities are based,

-shared Interaction patterns by which the community cooperates,

-domain Categories which characterize the community’s field of interest, and

-high demands for Homogeneity in function presentation and data visualization.

The MARINFO system follows these four principles - the RICH approach to gateway design (see figure 2-1) - and provides enriched information and communication services which exploit

-typed interfaces to digital resources,

-internet-based secure interactions,

-extensive domain descriptions based on categories, and

-template-oriented visualization for overall presentation homogeneity.


Digital representation of resources, internet-based interaction and template-oriented visualization are the technical means by which systems and services are realized, customized and personalized and by which a high degree of extensibility, automation and overall efficiency is achieved. Categories allow for the extensive and coherent structuring of domains along their essential dimensions - conceptual, organizational, temporal etc. - and provide unifying design guidelines for the entirety of system interfaces and services.

Figure 2-1: The RICH Approach to Information Gateway Design

2.1Gateway Resources

MARINFO provides digital resources which essentially fall into three categories. Since MARINFO is understood, first of all, as an Internet gateway which unifies and concentrates access to an extensive amount of valuable content on sites relevant to the maritime community, it concentrates on information re-use and enrichment in the sense of “syndication”. Therefore, many of the resources reachable through MARINFO are external and are owned and maintained by third parties. For a selected subset of these resources, MARINFO provides added-value through augmented data or improved services. An example is Maritime CORDIS, a service based on the CORDIS database of the European Community and tailored to the needs of the maritime community. Finally, there are resources such as the maritime calendar of events or the expert profiles which are internal to MARINFO, i.e., created and maintained exclusively by MARINFO.

2.2User Interactions

A second dimension along which a community can be characterized are the interaction patterns which it establishes to pursue its major information and communication goals. For individual information requests MARINFO provides a balanced mix between site navigation, database querying and full-text search. Open customer-to-customer interaction can be established and maintained by an interactive service for expert profile definition and evaluation. The dimension of group communication is represented by a CSCW functionality (Computer Supported Cooperative Work) for controlled document access and distribution ([DemDubJar+98]).

2.3Domain Categories

At the heart of the MARINFO design is a sophisticated maritime keyword and classification scheme used all over the MARINFO system services to

-categorize information objects, e.g., on projects, users or conferences,

-support category-based search, e.g., for publications and events,

-model expert profiles and rank profile matches.

Experienced members of the maritime community developed the underlying classification scheme by integrating existing schemes and by evaluating ongoing developments. The use of common classification schemes across a wide variety of services contributes to a shared vocabulary for the area, guides system design (information structures, query and search functions, visual user interfaces etc.) and, finally, improves the acceptance of the system. Many of these MARINFO experiences generalize to Information Gateways for other communities.

2.4Template Homogeneity

The creation of the MARINFO web pages is based on a template-oriented approach that ensures high presentation and visualization homogeneity across the entire information gateway.

The use of templates also supports easy maintenance of information gateways. The modification of a template automatically changes the visualization of all web pages that used this template (after the next page generation).

Figure 2-1 gives summarizes the RICH approach to information gateway design.

2.5On Gateway Overall Quality

The overall quality of a gateway is determined by the communitysatisfaction which, in turn, results from the degree to which gateway usability and gateway content support the community’s sphere of activity.

Therefore, gateway quality is essentially determined by two quality dimensions:

-gateway usability: functionality, responsiveness, efficiency, …

-gateway content: correctness, completeness, timeliness, …

The RICH gateway design principles, which concentrate on resource descriptions, interaction patterns, domain categories, and homogeneity templates, approach as their prime design goal the enrichment of gateway usability and content.

2.5.1Gateway Usability

The individual MARINFO services are designed according to the RICH design principles. For usability reasons the services are grouped, at the top level, into four broad service clusters:

-marDATA offers content from essential community resources (e.g., technical and scientific publications, glossaries, maritime projects, maritime events, ...) as well as information on major players in the field (e.g., professional institutions, companies, funding agencies, organizations of the European Commission, ...);

-marWORK provides workflow support for group- and goal-oriented document management, distribution and control.Examples are research and project documents handled by the members of maritime projects and circulated across networks;

-marEXPERT supports matchmaking and communication between those members of the maritime community who have specific profiles of expertise and those who require such expertise. The specific skills as well as the roles of experts are registered thus enabling a finely grained modeling and a model-based matching of demands.

-marNEWS publishes topical information on maritime news and events of all kind; this service cluster is characterized by its highly dynamic content which is made available by a professional press agency.

2.5.2Gateway Content

While the MARINFO service clusters primarily provide access to the MARINFO information resources, two other interaction patterns are intended for MARINFO content improvement and resource maintenance: for cooperative content evolution MARINFO supports the model of “give-and-take-loops”, for flexible quality management and improvement MARINFO follows the principle of a “person-in-the-loop”.

A “give-and-take-loop” is characterized by a productive interaction between content providers and content consumers resulting in a user-driven update and evolution of the MARINFO resources. “Give-and-take-loops” are implemented by highly interactive services which allow users to search for information, to add new information objects and to update his objects. It is the responsibility of the “person-in-the-loop ” to control the quality of the user-provided information entered through “give-and-take-loops” thus assuring the high quality of the MARINFO content. Technically this principle is supported by distinguishing between two versions of the underlying databases, a “cold version” still to be checked and a “hot version” which is already quality-assured and can be published.

From a design point of view MARINFO is a case study in Internet community information gateways, from an implementation point of view it is an exercise in component-oriented software construction.

3The MARINFO Gateway Design

Information services are available for nearly all communities through the scope and versatility of the Internet. However, the general availability of information is no longer the key problem but the concrete availability of relevant high-quality information at the right time and in the right place. Information gateways improve the concrete information availability for a community sharing a common area of interest.

The role of a gateway is primarily that of an information broker focusing on the collection, structuring, and annotation of existing services and information for the community. In contrast to a primary information provider, a gateway does not mainly produce information on its own but augments existing services and information with added value.

In addition to information brokering, an information gateway also plays an important role in community building, extension, and consolidation.

3.1MARINFO Resource Collections

As discussed earlier, the idea of an information gateway is not information production but the provision of a common marketplace for a wide variety of information relevant to the community. A gateway acts as an information broker collecting, integrating and augmenting existing resources such as link collections, databases, and services as well as providing improved access to them.

A balanced mixture of

-resource syndication, relying on existing services,

-value adding resource improvement, and

-the production of genuine resources

promises to be an attractive service palette that is versatile as well as of high quality and in the same time efficiently sustainable.

3.1.1MARINFO Resource Syndication

The exploitation of existing services is an explicit goal of information gateways. The simple linking of a service from the information gateway, i.e., resource syndication, already adds value to the linked service in the sense of the “one-stop-shopping-place”: The resource becomes more easily accessible and is set into the context of other related resources.

3.1.2MARINFO Resource Enrichment

Added value is achieved by a gateway if existing resources are not only gathered but if the profit from using them via the gateway is increased by additional services and information. Examples of value adding information are comments, annotations, and ratings but also the aforementioned community-oriented classification of resources according to a common classification scheme. Value adding services are, for example, retrieval supports that span several independent resources and facilitate or improve access to existing resources enabled by cooperation with the service providers. The Maritime CORDIS database is an example of value adding resource improvement in the MARINFO project.

3.1.3MARINFO Genuine-owned Resources

The development of an information gateway is characterized by a intensive discussion of this kind of service coupled with cooperation between different parts of the community for which the gateway is developed. This often leads to the discovery of new service and resource types that are useful but not yet existing in the community. Such resources may be set up as part of the information gateway realization. In the MARINFO project a calendar of events and an expert brokering service have been set up genuine-owned resources

3.2MARINFO Interaction Patterns

The operation and use of an information gateway is determined by the interaction between the gateway, its resources, and its community. Developing adequate patterns for this interaction is, therefore, an important challenge for successful gateway operation. In the MARINFO project four such interaction patterns have been identified which are discussed below.

3.2.1User Access Patterns: Navigation – Query – Search

Three basic access patterns may be observed in an information gateway:

-navigation following the hyperlinks offered by the gateway,

-the use of specific, form-based query interfaces exploiting the service-specific query options,

-full-text retrieval using a search engine. The value of this access pattern is improved if the search space covers different resource collections and may be adapted flexibly.

In the MARINFO gateway all three access pattern are exploited. An elaborate structuring of the gateway enables quick access to the most important resources via navigation. Service specific query interfaces are, for example, used to mediate between experts and community members searching for expertise in the marEXPERT service.

The rationale behind the MARINFO search engine, a full-text retrieval engine for the MARINFO gateway, is to enable a user to find all information objects on the MARINFO server, including static pages as well as pages dynamically generated from MARINFO databases, that contain certain keywords, like “laser welding”, or a combination of keywords with a single point query. Hence, a user does not have to browse the MARINFO pages or query each MARINFO information service. In addition, a user should be able to retrieve all pages of important maritime web sites that matches the query.

Searching other web sites can also be done by querying common search engines like AltaVista, Yahoo, Excite, etc. The problem with world-wide operating search engines is that a simple query (e.g., maritime industry) may return a very large result set (e.g., 356000 pages found). Browsing through the result set is extremely time consuming and most of the results are not of interest to the user. Therefore, MARINFO contains a search engine that is restricted to all MARINFO pages and databases as well as pages stored on important maritime web servers. The latter can be extended dynamically.

3.2.2Resource Evolution Patterns: Give-and-Take-Loops

Give-and-take-loops are used as a paradigm for the cooperation between information providers and information consumers. This paradigm weakens the distinction between information consumers and producers by fostering a combination of both roles. Consumers of the information gateway also appear as producers of information and vice versa. The resulting cooperative information exchange produces an information gateway whose content is maintained by the targeted community itself. The quality of the services, such as the amount of information that can be searched, thus depends on the degree of community involvement and activity. This means the gateway operator, in this case MARINFO, provides only the information infrastructure and change services (insert, update, etc.) and the community provides the contents.