Issues of importance in the design of an ideal group-based systems are discussed by Mandviwalla and Olfman [45]. (A group, in this case, is defined as a combination of people, environment, and tasks). There is a need to support a variety of tasks. In the case of collaborative authoring, writing , reviewing, revising, communicating with co-authors, the system should not be group dependent; it should support the needs of a variety of user groups. It should also support group development, that is changes in membership, commitment changes, organizational relationships formed, and organization cultural changes. Also, the system should support changes in group roles and context. Group context includes: work methods, membership, and culture. This is especially noteworthy given the changes that occur in writing group membership and the fact that much discussion about organization, content, and relationships occur during the writing of the document and in sub- groups. Further, reasons for working in a group include maintaining relationships. A group-based system should provide some mechanism for recording group memory, that is a record of past activity. Again, changes in group membership and revision management require this option to maintain continuity. The ideal system should support multiple interaction modes, such as: face-to-face, verbal, written, email. The system would need to facilitate changes between modes. Support of social structures is an important aspect of the system. Social structures include the use of time, stress, commitment, and relationships.
Engelbart and Lehtman [19] identify the elements necessary for a collaborative system and they are applicable to a collaborative authoring system. One element is support for collaborative dialogue. This requires computer tools for composition, reviewing, referencing, modifying, transmitting, indexing, and storing messages. Also, included is automatic message delivery. Another element is document development, production, and control. This relates to composing, reviewing, and modifying document drafts. The tools needed for this element are mechanisms for controlling changes and version distribution. Further, the ability to authenticate submissions and comments. This relates to giving credit for contributions and ideas. A facility to remind those who do not respond to requests to comment, modify or approve an particular document is required under this element. Providing tools for management and access of previous versions of documents also applies to this element. Another facet of a collaborative system is the ability to manage items created externally, such as reports, notes, bibliography. Such items can be maintained by with a database system. The fourth element is the capability to develop and maintain a community handbook. This handbook would provide complete, uniform information on the current status of the group. It should include items such as principles, goals, goal status, techniques, standards, glossary of relevant terms, and hypotheses. A means of providing training to work group--geared toward its special training needs. (Such a system feature would address the situation where new members need to be acclimated to the work of the group.) Another aspect are tools for local and distributed meetings. Project management tools are also an important part of a collaborative system. These tools, such as the previously described dialogue support, document and handbook support, would include plans, commitments, schedules, and specifications. (This relates to the motivational needs as well as the changes in commitment issue.) Another requirement is that the support tools be available to individual members and that those members can customize (parts of) the system.
Engelbart and Lehtman [19] also indicate assumptions that are applicable to a successful computer-supported collaborative work system. Assumptions that relate directly to augmenting collaborative authoring (namely reducing cognitive overload) are presented. One assumption is common style of communication, for example command structure and menu format across different parts of the system. The user, in this case does not have to re-learn skills, but can enhance existing knowledge of the system. The system should support user with a variety of skill levels. (This supports the changes in group members and the ability of new members to begin contributing quickly.) As users learn more they are rewarded with the ability to customize their version of the system. The ability to move easily between various portions of the system is another aspect of a successful collaboration system. Also, the ability to add new features easily supports the notion that the system should evolve with the groups .
Posner and Baecker [60] investigated the collaborative writing process in different projects, aiming to evaluate the existing computer supported collaborative writing systems. Roles of participants, writing-related activities, document control methods and writing strategies were identified. Group members can have writer, consultant, editor, reviewer or equal-work roles. Major activities include brainstorming, planning, writing, and editing and commenting. Four document control methods were used: centralized, relay, independent, and shared methods. In centralized control method, one team member controls the document throughout the project. The control may be passed or relayed between team members and the only one person takes controls at a time. Alternatively, each individual can work on a separate part of the document and control only one's own part. In the shared method, several team members can have equal access and control on the document simultaneously. The writing strategies employed can be classified into five categories. The first strategy is the sinble writer strategy in which only one person writes the document based on discussions among other team members. In the scribe strategy, one writer is responsible to record all the thoughts and decisions of other group members. This strategy is typically used during the early phases of a project. When there are joint projects, the document is divided into parts each of which is assigned to individual members. This is called the separate writers strategy. Another strategy which is based on group effort is the joint writing strategy in which multiple individuals help writing a document. When the joint writing strategy is used, the document control method is critical. Lastly, the consulted strategy can be combined with other writing strategies, described previously. In this case, a writer asks for help from a consultant who could guide the whole process. This empirical evidents suggest that a system that supports collaborative writing activities should support different kinds of collaborative writing strategies.