During the lifetime of the project, the INSPIRE consortium will develop a large and innovative software tool, and at the same time validate and use this tool in genuine field BPR projects. The essential principle behind the INSPIRE workplan approach will be the notion of rapid and iterative prototyping--that is, a procedure will be adopted in which the developer partners will release, from an early stage, successive incomplete but running versions of the toolset, and then respond quickly to user feedback with further releases; the lifecycle of the project will see a constant exchange of feedback and comments from the end-users to the development partners, and a corresponding succession of software releases. The workprogramme is divided into workpackages reflecting major components of the project effort. Every task in every workpackage will have or will contribute directly to a deliverable of the appropriate form (normally document or software), and the completion (or significant part thereof) of every workpackage is marked by a Milestone.
The INSPIRE tool is composed of a large number of separate, though integrated software modules. Two of the three software partners in the INSPIRE development team have considerable experience and expertise in industrial decision support and business information systems which will ensure that although the software effort in INSPIRE will be considerable, this experience and know-how will ensure that a functioning system is developed quickly, thereby allowing the End-User partners a working piece of software with which to begin field trials, and meaning that the development team are available to focus on the advanced features of the INSPIRE tool from the outset. In order to ensure the effective and coordinated development and integration of such a large tool, the consortium will appoint a Technical Coordinator from TXT. This role will be to lead and manage the software development effort. The various modules will be linked into a functioning whole at TXT's business premises.
The components of the system are described below.
This component forms the most significant part of the HCI, and supplies the main graphic development environment for the INSPIRE software tool. This module will support the diagrammatic display of the various representation of process models, at the same time as allowing the user the full range of windows-based diagram drawing functionality as well as intelligent auto-formatting features to ensure the best possible display appearance. The overall design of the graphical display functionality will take place in the specifications tasks (see workpackage descriptions below)--but it is envisaged that IDEF-type diagrams will form the basis for these views. Note that these formalisms are not designed for intuitive accessibility by non-expert personnel--this latter kind of view will be overlaid on these formal representations (see Human Resource and Iconic Process Description Views below).
Direct editing will be possible in this module (drawing, cutting, pasting etc.), and connections and coherence between views will be enforced automatically, with advice to the user where required.
These features will be overlaid onto the formal process description views described under the previous heading. The purpose is to allow icons to (possibly semi-transparently) stand in place of boxes or groups of boxes, in order to render the formal descriptions more intuitively acceptable. The Human Resource/Skills information will be superposed onto the formal views in the same way, allowing non-technical users the ability to relate the process diagrams with actual skills requirements and availability.
This module forms the ``engine room'' of the INSPIRE tool, and comprises one of its most innovative and powerful features. The PRM is responsible for the abstract representation of business processes in a form which will support the generation of the various concrete representation schemes which the tool will offer, make explicit the connections between these schemes, and support the more advanced simulation and Artificial Intelligence features. For this purpose, the PRM will deploy an activity-based Intelligent Agents architecture (see e.g. [Zanker 99]). Such representation schemes are powerful, flexible, and generic enough to support the design goals of the PRM Module. Detection of errors in the processes will be an advanced feature, offering immediate guidance on diagram completeness, consistency etc, but also on the coherent connections between the various process and performance views. Certain properties in particular will be highlighted, e.g. loops in chains of intentional dependencies, conflicting goals, opposing tasks, resources required in different activities.
It is important to note that the PRM will represent and manipulate all aspects of the BPR effort, including performance representation and evaluation, both qualitative and quantitative, in a centralised, formal model. The representations in the PRM will also be required to integrate tightly with the Natural Language front-end, and with the Open Model Repository.
The OMR is the database environment in which the business models developed in the tool are stored. The access library will allow the tool to load and save entire projects/versions, or else individual features of process models, as required, thereby supporting the smooth and friendly development of BPR activities. The OMR and its access library will also be deployed to serve as a link between the INSPIRE tool, and other tools and environments. In this sense, the OMR will be a mirror to the PRM, constituting an abstract process meta-model. The resulting coherence and flexibility of design will support the seamless and powerful integration between INSPIRE and external IT tools. In this connection, as demonstration of this feature, the consortium will develop export links form the INSPIRE tool to an Executive Information System (possibly Forest and Trees), and will investigate the possibility of links to external Workflow Engines and production simulation tools (e.g. WITNESS). Note that such data export could not contain sufficient information to drive a WITNESS simulation directly (more detail is required for a simulation package to run), but would nonetheless supply the basic parameters for such simulation.
The Simulation Module will handle process static evaluation and dynamic simulation of processes, and will provide both quantitative and qualitative performance analysis as well as process animation. To achieve this, the module will supply disambiguation functionality and automated generation of dynamic and workflow simulation from static models. Once again the mechanism to accomplish this begins with the rich abstract model in the PRM, which will have the knowledge and flexibility to generate these features. The simulation module will support a full range of quantitative Performance Indicator criteria built-in by the system (including product cost, process lead time, resource consumption etc.). In addition, the user will be allowed the freedom to define customised indictors either from scratch or by mathematical combination of existing indicators. These features are for the use of consultants and experienced BPR personnel, and will not be aimed at non-technical people.
An interface for the creation of fuzzy and imprecise qualitative performance indicators will be implemented using a fuzzy logic engine, built along the lines of an intelligent system shell (e.g. CLIPS), allowing the user to ``tweak'' and experiment with the indicator's parameters: so for example, certain activities can be associated with certain imprecise effects (e.g. ``quality'') in varying degree. Moreover, these indicators will be able to interact with each other (so for example a certain activity or group of activities might score highly on (e.g.) ``ease of set-up'' but less highly on ``flexible responsiveness to customer''. In this way, competing possible processes may be compared with each other according to an unlimited set of indicators. The setting-up of imprecise and qualitative performance indicators will be a specialist's task, but the result is intended to be immediately comprehensible to non-specialists.
The automatically generated dynamic simulations will be enriched with icons (related both to resources (documents, materials etc.) and to the process itself ( stop-go signs, clocks/hourglass etc.). Moreover, the dynamic simulation will use animation. The kind of thing envisaged here is that parts of the diagrams may change colour, flash, or give messages, and simple moving animation (not unlike the emptying of a Windows Recycled Bin) will be included. The direct purpose here is to bring the workflow representation ``to life'' in a way which renders the simulation obvious to non-specialists.
This module will serve two roles;
First, it will allow a non-specialist user to model business processes with (semi) natural language, with natural language feedback, and without being expert in any of the standard, formal process representation formalisms, and it will support the very fast ``real-time'' entry of process models into the tool, for example, actually during BPR team meetings. The mechanism for this will be ``point and click'' menu-driven: the user will be prompted to select the kind of process they wish to model from pre-compiled menus, given options as to how to link their models together etc. At the same time, as the user builds a representation through this interface, their commands will be relayed back to them in natural language, allowing them to validate that what they enter is correct. When the user is satisfied, the machine will automatically create/modify the underlying process model, and the graphical representation features will be formatted and generated automatically. Of course the user is free to adjust and manipulate the models ``by hand'' at a later time.
Second, it will provide (semi-) natural language support for the user to enter business rules for use in the Diagnosis and Explanation Module adjunct to the simulation module (see below). In a similar fashion to entering process descriptions, the user will be prompted to link the rules as they are entered to the relevant activities.
This will be attached to the Simulation Module, and will provide coherence checking facilities and expert advice on processes, as well as explanation of the reasoning performed in simulation and analysis (both quantitative and qualitative). This module is also responsible for fielding ``why'' and ``how'' questions where business rules have been associated with activities. To achieve this, the DEM will incorporate a simple Natural Language Generator, working with a canned corpus of text.
The IPM comprises a case-based implementation plan generator, linked to a repository of BPR studies and best practice, an engine for configuring precisely the human requirements for successful BPR take-up (skills/knowledge/personnel/training needs), and a generator of process objectives monitoring criteria with which the performance of the newly adopted process can be realistically monitored. This module is responsible for supporting the embedding of BPR modeling and design into a real-life industrial setting. For a given business setting, it will supply a template for implementation to ensure that the implementation plan is determinate and complete, critical dependencies are in place, and realistic milestones set up.
This is the database in which are held the various features of the INSPIRE tool which are developed and accumulated over time (icons, business re-engineering best-practice cases, business rules). As mentioned before, the ultimate goal is that the power of the tool grows as more and more information is stored.