Steve Goschnick

UTILIZING WEB SERVICES IN AN AGENT BASED TRANSACTION MODEL

The Web services paradigm is an opportunity for a universal programmatic interface to the Internet, one that could parallel the web-browser, human user-centered interface, in scope and adoption. However, there remains a great challenge in being able to reliably perform transactions in the loosely coupled Web services environment, hosted by independent Web service providers, running on heterogeneous systems. The ACID properties in traditional transaction models cannot be guaranteed in this environment, due to unreliable communication, the uncertain duration of individual services and the decentralized nature of server management. We draw lessons from the long history of research into distributed transactional systems. The recent WS-Coordinate protocol draws upon some but not all of those lessons. Unlike WS-Coordinate, we are interested in being able to incorporate independent take-it or leave-it Web services, into orchestrated systems — not just those that agree at the outset, to provide a particular service, with a mutually agreed quality-of-service regime. To address these issues, we call upon the flexibility of the agent-oriented paradigm, and propose an agent-based transactional model (ABT) that orchestrates multiple, often independent Web services, into new robust services. We assert that the use of multi-agent technology to manage and orchestrate transactions is the right choice in the Web services environment — the choice most-likely to propel the Web services paradigm to levels of adoption that approach those of the web-browser interface.

Design and development of Melbourne IT Creator/sup TM/-a system for authoring and management of online education

This paper presents a case study in the symbiotic use of new Internet based technologies and an SQL server to develop a software tool in a new category of generic software: a system for authoring and delivery of Web-centric learning. In the design and implementation of the system, the developers drew upon the latest available languages and platforms, aiming for a high benchmark in this new software genre: Java for cross platform authoring tools; JavaScript and HTML V4 for scripting and markup; applets, video and other media types as object components; IIS (Microsofts Internet Information Server technology) to deliver dynamically constricted HTML markup. Behind the interfaces and business rules is robust SQL server technology, which is taking on an expanded role in proliferating Web based information systems. In the latter half of the paper problems and solutions are discussed, including the use of metadata and XML (the eXtensible Markup Language) as part of the solutions.

An agent-based digital self in a 24x7 Web services world: architecture and implementation

Intelligent agents have the potential to play an important role in meeting the challenge to provide tools and interfaces for people to productively interact with the Internet 24 hours a day, seven days a week (24x7). This paper describes a novel agent-oriented architecture called ShadowBoard with a novel user interface called ShadowFace. Both the architecture and interface draw on concepts from cognitive psychology. The architecture has been embedded in a software framework called ShadowSpaces, and has been used to develop a prototype lab manager for booking services and providing advice to potential users of the Interaction Design lab at the University of Melbourne. An overview of the architecture, interface, framework and implementation are all presented.

From Task to Agent-Oriented Meta-models, and Back Again

In the research discussed here, in addition to extracting meta-models from numerous existing Agent architectures and frameworks, we looked at several Task meta-models, with the aim of creating a more comprehensive Agent meta-model with respect to the analysis, design and development of computer games. From the agent-oriented perspective gained by examining the resultant extensive agent meta-model --- named ShaMAN --- we then revisit the Task Analysis research domain, and consider what benefits Task Analysis and Modelling may draw from the Agent-oriented paradigm.

Enacting an Agent-Based Digital Self in a 24x7 Web Services World

As broadband access to the Internet becomes pervasive, the need for a 24 hours a day, seven days a week (24x7) interface within the client devices, requires a level of sophistication that implies agent technology. From this situation we identified the need for a user-proxy++, something we have termed the Digital Self that acts for the user gathering appropriate information and knowledge, representing and acting for them when they are off-line. With these notions in mind we set about defining an agent architecture, sufficiently complex to deal with the myriad aspects of the life of a busy time-poor modern user, and we arrived at the Shadowboard architecture. For the theory, for the model of mind, we drew upon the Psychology of Subselves, a modern strain of Analytical Psychology. For the computation engine we drew upon Constraint Logic Programming. For the hundreds of sources of sub-agency and external intelligence needed to enact a Digital Self within the 24x7 Internet environment, we drew upon the Web Services paradigm. This paper presents the theory, the architecture and the implementation of a prototype of the Shadowboard agent system.

A Composite Task Meta-model as a Reference Model

In this paper we develop a comprehensive composite meta-model from Task Analysis models called the Reference Task Meta-model (ReTaMeta model) for the purpose of comparing numerous Agent-Oriented meta-models. The reference model needed to be derived from a field independent of the Agent-oriented paradigm, yet based on Psychology. To arrive at the ReTaMeta model we first extracted the meta-models from several well-known cognitive task models including GOMS, GOMSL, TKS, GTA and also the CTT and Diane+H Task Modeling notations for fine grain task detail, and then combined their respective concepts in a complementary and comprehensive way.

TANDEM - a design method for integrating web services into multi-agent systems

This paper introduces a new design method for multi-agent systems (MAS) that incorporate logic programming. The DigitalFriend is an example of a MAS with a built-in logic language interpreter, which is aimed at having the end-user as developer. It uses small fragments of logic as dynamic glue, bringing together numerous sub-agents that may exist as Web services. However, application developers need to devise well-formed predicates and logic rules - and there lies the need for a specific design method. The well-established techniques of Task Analysis and entity relation Normalization are drawn together into our new design method (TANDEM), introduced here with an example application - the movie-cinema problem.

Automating the TANDEM Design Method in End-User Programming Environments

Mashups are newly envisaged applications, made up from local information sources and processes, Web services and other distributed resources, bound together technically in some way. Interactive Development Environments IDE used to build mashups are becoming more accessible to end-user programmers. Design methods that end-users may apply to a given problem addressed by a mashup, are much less prevalent. This paper describes an end-user-friendly design method called TANDEM and demonstrates the use of it in detail, by way of an example: the design of a mashup of services that solves the so-called movie-cinema problem. An implementation of the newly designed movie-cinema app is then built within the DigitalFriend, an end-user programmer IDE. Furthermore, a significant part of the TANDEM design method, is then automated within the development tool itself. This automation removes the most skilled task required by TANDEM of the end-user: the automation of the process of Data Normalization. The automation applies data normalization to the initial model of components and data sources that feed into the mashup. The presentation here relies on some understanding of Data Normalization, so a simple example is presented. After this demonstrated example of the method and the implementation, the paper discusses the applicability of a model achievable by end-users using TANDEM coupled with the automated normalization process built into the IDE, versus, using a top-down model by an experienced information analyst. In conclusion, the TANDEM method combined with the automation as demonstrated, does empower an end-user to a significant degree in achieving a workable mashup or distributed application. And furthermore, the TANDEM method does have broader applicability to designing a broad class of logic programs, complementing the use of collected patterns in logic programs.

ShaMAN: An Agent Meta-model for Computer Games

In this paper, we detail recent research on agent meta-models. In par-ticular, we introduce a new agent meta-model called ShaMAN, created with a specific focus on computer game development using agent systems. ShaMAN was derived by applying the concept of Normalisation from Information Analy-sis, against a superset of agent meta-model concepts from the meta-models in-vestigated. A number of features are identified, including human-agent localesand socialworlds, that might be usefully added to a generic AO meta-model.