OzWeb supports hypercode and other cross-linked formal and informal documentation, including text, images, audio/video, etc., that may be distributed among multiple repositories geographically dispersed across the Internet and/or organizational intranet. Project-specific data and services are added on top of World Wide Web entities. Multiple users are grouped together into collaborative teams. OzWeb provides software developers with:
Project artifacts are collected in an object-oriented referential hyperbase. OzWeb objects represent external Web pages or native database entities. Each object is an instance of a particular class, with multiple inheritance, primitive and file attributes, and composite and link relationships. Relationships impose two-level hypertext independent of the underlying media, which may have its own embedded HTML links. The same content can be used in different ways by different applications.
Commonly used search engines often return references to thousands of entities, most of which are irrelevant to the task at hand. While this may be very useful for finding previously unknown documents, it is often desirable to restrict search to project-related entities. OzWeb text-matching search is performed over only those entities represented in the hyperbase, optionally a selected subset of those entities, using a standard search engine available in the public domain. In addition, a sophisticated query language supports associative and navigational search over hyperbase objects.
Depending on the specifics of the process definition, OzWeb enforces task prerequisites and implications, including constraints on when artifacts can be viewed or updated; automates invocation of tasks at the appropriate time; notifies appropriate supervisors under specified conditions; and collects metrics and maintains a complete audit trail of who did what when with which data.
It is extremely tedious for human users to enter large numbers of links between objects, but automation based solely on queries or pattern matching, e.g., triggered by changes, can result in a dense mass of links that is virtually useless. The process semantics and enactment capabilities make it possible for OzWeb to infer semantic dependencies among entities, e.g., the inputs and outputs of a task, and automatically introduce the proper links in the hyperbase.
Multiple users may access the same document at the same time, and could potentially overwrite each others’ changes or view partially updated documents. Human work products are at stake, it is usually unacceptable to lose more than a few minutes work due to a system crash or network problem. OzWeb employs the process definition and separate concurrency and recovery policies to decide when to enforce the conventional ACID (atomicity, consistency, isolation, durability) properties, and when they can be relaxed to increase parallelism and enable sharing of partially completed work among team members.
Project participants may use a variety of COTS and GOTS tools that run on different platforms, and it is not always feasible to make every relevant kind of machine and every software package directly available on the user’s desk. OzWeb supports remote launching of tools as well as redirection of X Windows and Windows NT/95 GUI’s to the user’s screen. Peer tool servers communicate with each other across the wide area network to determine the best place to launch a tool, and automatically set up the local environment and invoke the tool on behalf of the user; the default is the user’s own machine.
Publications.
Gail E. Kaiser, Stephen E. Dossick, Wenyu Jiang, Jack Jingshuang Yang and Sonny Xi Ye, WWW-based Collaboration Environments with Distributed Tool Services, to appear in World Wide Web, Baltzer Science Publishers.Gail E. Kaiser, Stephen E. Dossick, Wenyu Jiang and Jack Jingshuang Yang, An Architecture for WWW-based Hypercode Environments, in 1997 International Conference on Software Engineering: Pulling Together, May 1997, pp. 3-12.
Availability.
OzWeb has been retired