Taskoriented Information Modelling (TIM) Proposal
The aim of this proposal is to standardize such a formal language using Taskoriented Information Modelling (TIM) as a starting point. TIM has been used for 10 years to produce task models in various fields (machine, process and software modelling). It proves easy to learn, its expressions translate straightforwardly into natural language, and its syntax is defined by a short set of XML-compliant rules and patterns.
TIM builds on an incremental ontology, where a reduced set of categories [objects, operations and facets] generate descriptions [states, properties] and actions. These, in turn, populate the categories describing tasks (preconditions, input, output and side-effects). Tasks are organized in a hierarchy. Tasks exhibit types (standard, alternative, critical, conditional, optional, cyclic, generic), which may constrain their position in the hierarchy (critical tasks are always first-level tasks, alternative and optional ones are always sub-tasks), or their structure (generic tasks are always split into alternative or optional sub-tasks), or their content (only last-level tasks are enhanced with a list of actions, termed "procedure"). Conditional tasks host preconditions, critical tasks a "trigger", cyclic tasks a cycle-stop condition. Some type-combinations are allowed (conditional + optional, for example), while other ones are not (for instance, conditional + critical). The formal language expressing this ontology is a predicate-argument type syntax. Every entity in the model has a numerical identifier allowing reuse. A reduced set of symbols (O: object, S: state, P: property, A: action, T: task) combined with the relative position of an element in the model (which is a frame-based structure), define which descriptions act as the Input, the Output, the Side-effect, etc.
|sitemap||Copyright 1971-2012 01 COMMUNICATIONS INC. ALL RIGHTS RESERVED. - Powered by DNAOS||contact|