As shown in the figure above, the IMPACT argumentation toolbox being developed will consist of: 

1. An argument reconstruction tool, which uses a library of argumentation schemes to support the manual reconstruction of arguments from natural language texts. Moreover, an extension of existing XML formats for weblogs, such as RSS or Atom, will be developed which would enable future weblogs to mark up the structure of arguments in articles in such a way as to enable the arguments to be automatically aggregated, analysed and visualized, without human intervention. The extension may be based on the argument elements of the Legal Knowledge Interchange Format (LKIF) developed previously by the partners in the ESTRELLA project (IST-2004-027655).
2. A policy modelling and analysis tool, based on computational models of argumentation about alternative courses of actions given the goals and values of multiple stakeholders, and our prior research on knowledge representation languages for concepts ("ontologies"), defeasible generalizations ("rules") and precedent cases, including the Legal Knowledge Interchange Format (LKIF) we developed in the ESTRELLA project (IST-2004-027655). This tool will also include a graphical user-interface for engaging in a dialogue with an inference engine to simulate and analyse the consequences of a proposed policy in particular cases, including hypothetical cases. The tool will be able to produce a clear visualization of its reasoning, so that citizens can understand not only the effects the policy would have on their own lives, but also understand why. The tool will enable citizens and other actors in the policy development process to compare the effects of different policy proposals.
3. A structured consultation tool, based on prior research on the PARMENIDES system developed by the University of Liverpool, a member of our consortium. This tool will be an advanced, intelligent polling and survey tool, based on the state-of-the-art of computational models of argumentation. The questions in the surveys are generated automatically from a model of argumentation schemes, together with a model of the issues and arguments which have been put forward previously in the ongoing consultation. The tool will dramatically increase the signal to noise ratio in online discussions, without unduly restricting the substantive arguments which can be made, by helping citizens to apply a model of rational argument, using only argumentation schemes which are accepted by the community in which the consultation is taking place. The tool also enables arguments to be more easily tracked, mapped and visualized, since it obviates the need to manually reconstruct arguments from natural language texts.
4. An argument analysis, tracking and visualization tool, based on the state-of-the-art in the field of computational models of argument and argument mapping methods. The analysis features of this tool will enable citizens to identify which argumentation schemes have been applied, to reveal implicit premises helpful for asking critical questions, to determine whether some statement at issue is acceptable given the arguments which have been put forward, according the user's choice of proof standard. The tracking features of this tool will enable users to register their interest in particular issues and request and receive notification whenever new arguments have been put forward which affect these interests. The visualization feature of this tool provides a variety of interactive, graphical views onto argument graphs, at different levels of abstraction and detail. This will enable citizens to appreciate the complexity of the policy issues in their entirety, to zoom in on the issues which are of interest to them and then be in a position to articulate a reasoned contribution to the consultation.

These tools will be designed and implemented so as to be usable together, as part of an integrated argumentation toolbox, which can be installed upon and used with a variety of existing popular Open Source content management systems (CMS), such as Alfresco, Drupal, and Plone or the Joomla system used by Gov2Demoss project. The underlying content management system is responsible for providing identity, user and group management, authentication services and assuring security, as well as the usual document repository, communications, search and notification services. To facilitate the integration of the argumentation toolbox with a variety of CMSs, an abstract Applications Programmer Interface (API) for the services required by the argumentation toolbox from a CMS will be defined, making use of existing or emerging relevant standards, such as the OASIS Content Management Interoperability Services (CMIS) proposal, whenever feasible. Porting the argumentation toolbox to a new CMS will be a matter of implementing just this API. Style sheets will be used to enable the graphical design of the argumentation toolbox to be customized and adapted to the design of the web application in which they are embedded and used. Drawing 1 shows how the argumentation toolbox will be integrated on top of existing content management systems.

The IMPACT argumentation toolbox will be designed and implemented as advanced Rich Internet Applications (RIAs) or “widgets”, so as to enable them to be used in the widest variety of contexts, increasing accessibility and helping to overcome the digital divide, including within web pages, as applets, as standalone desktop applications, televisions, via desktop boxes, and, where appropriate, as applications on mobile devices such as smart phones. Moreover RIA technologies enable a new generation of dynamic, easy-to-use and “immersive” user interfaces, which set a new standard for attractiveness and usability, helping to facilitate increased participation in consultations. There are a variety of competing frameworks for building RIAs, including Adobe's Flash, Oracle's JavaFX and Microsoft's Silverlight. The alternatives will be evaluated against detailed requirements for the argumentation toolbox acquired in the first phase of the project. It may be that different systems are preferable for each tool, in which cases steps must be taken to assure the toolbox as a whole is well integrated.