In general terms, task modelling is concerned with describing how the work is performed by one or more persons to achieve a given goal. However, such a loose definition also means that a primary concern in the planning of the model is the desired level of granularity, which can go from a general level down to specific hand movements on a control panel, for instance. A modelling language is intended to express the properties of the modelled entity in a way that covers all the aspects needed to fulfil the purpose of the development work. It is therefore useful to look at what problem domains a task modelling language is expected to cover and what the models needs to express to be helpful in that domain.
Task models are attempts at describing tasks, subdivisions of a sequence of activities, in such a way that they can be treated formally and provide a useful level of predictability for their intended purpose.
The MagigPED task modelling consists of two parts:
The first one is based on hierarchical task modelling (HTA), especially regarding the task hierarchy. Planning (which is also a part of many HTA schemes) is supported through the use of “time constraints” between the tasks: before, parallel, choice (or without any constraints it is unordered). This provides a high level of modelling capabilities, which typically encapsulates normative task knowledge. The formal structure of the models is based on the work of W3C.
The second part extends the modelling with a rule level, which will provide a more descriptive modelling of the actions close to the actual physical equipment, software UI, other agents and the controlled entity. This level of the modelling is based on GSM (Goal-State-Means) modelling, to form an overall modelling framework that connects the higher-level goals of the top level model with the actual state of the environment.
- UML Profile for Task&Rule Modelling
- Task Hierarchy Diagrams
- Task Modelling elements, like Tasks, Subtask-Connectors, Time-Constraints)
- Rule Diagrams
- Rule Modelling elements, like Rules, Motor Actions, Percepts, Speach-Actions, Memory-Stores and Memory-Retrieves
- Plug-ins to enhance the modelling experience (Editor, Validation, …)