Abstract

The aim of this work package is to adopt scalable reasoning technology to support the process of ontology-driven software development (ODSD). With models consisting of both software modelling and ontological entities, one must provide ways of querying over such joint descripBons. Since queries in standard ontology languages can be very costly, reasoning on large hybrid models require software engineers to be guided in how to tailor their queries; how to transform a given query into a nearly equivalent query in a less expressive language where reasoning can be optimized, or how to change the representation of the underlying concepts so that this becomes possible. Ontologies should be analyzed and suggestions for improvement of the complexity of queries and reasoning processes should be given. The work package will provide tools for transformation among different ontology languages and for language transformation guidance.

Mission

The aim of the work package is further developed into the following objectives:

1. The integration will address modelling languages. A unified view of metamodels must be provided to be able to query and transform integrated models.
2. Ontologies and models will be integrated so that ontologies can be used in models, and vice versa. MOST will enable integrated model chains, transformations between models either consisting of ontologies or system models.
   
3. MOST must develop syntactic and / or semantic transformations / approximations between languages in the hierarchy.
4. MOST must provide guidance for language transformations / approximations. Guidance information and supporting tools should be provided for explaining the consequences of the transformations / approximation for which translators were developed.

Use Scenarios

The MOST objective of seamless integration of ontologies and software artefacts and processes promises increased efficiency and productivity by easy reuse of domain models (WP1). For the first time, ontologies will leverage the reliability and quality of the resulting software by developer guidance (WP2) throughout the processes. Additional quality can be checked and guaranteed for complex systems by enabling efficient checking of constraints over large software models using tailored ontology languages and reasoners (WP3).

WP3 Reasoning Technology can be used to automatically infer the relevance of certain process steps or tasks in the context of the chosen solution approach, thus avoiding time-consuming but unnecessary steps.

Further details to WP3 Mission Objectives

1. […]. There are variations with polynomial time complexity, such as the systems QuOnto and ONTOSEARCH2 for DL-Lite reasoning, and the systems CEL and REL for EL+ reasoning, and there are variations for OWL-DL and OWL-Lite that are in complexity classes beyond NP such as FaCT++, KAON2, Pellet, or Racer for OWL reasoning (KAON2 also for SHIQ with DL-safe rules).
2. […]. On the one hand, this could provide useful information on how to make use of existing languages and related scalable services (if any); on the other hand, it is necessary to provide users and developers with information and guidance as to what languages they should use in which situations.
3. […]. Translating between languages can be used for trading off between the expressiveness and scalability of the different languages. More importantly, fine grained transformations and approximations could allow users who use expressive languages for modelling can still enjoy the scalable services of tractable fragments.

Reasoning Technology in the big picture of MOST

• WP3 Reasoning Technology will provide practical reasoning opportunities over the process of ontology-driven software development (ODSD). Transformation / approxiamtion technology will be provided for instances of languages in a developed and well-defined language hierarchy, to provide for appropriate and reasonable query complexity. This will contribute to the MOST innovation on ontology-based guidance for tailoring of the reasoning on integrated models, so as to allow for managing the logical structures to be queried and reasoned upon, exploiting the technology that is most appropriate wrt. expressiveness and scalability for the given task.
• Languages for software models (like UML) and ontologies (like OWL) have different elements and properties and, hence, different metamodels. MOST will contribute to the smooth integration (in WP1) between the two worlds. WP3 Reasoning Technology provides support for checking consistency and query the combined UML and OWL models in a practical manner at the metamodelling layer.
• One of the main objectives of the MOST project is to create an ontology-driven software process guidance system (in WP2) that is able to support software engineers in their daily work. The vision is a kind of workflow engine guiding the software developers through the software processes. WP3 Reasoning Technology will make use of the semantics of software artefacts (from WP1) and traceability information (from WP4) to infer appropriate next steps at a stage in a running development process. The transformation / approximation guidance tool to be developed in WP3 will be one of the key components of the ontology-driven software process guidance system.
• In the case study "Guidance of the Software Development Process" from SAP, results from WP3 Reasoning Technology (together with results from WP1 and WP2) will be used to improve the efficiency of the development process as a whole. We will do so by formalising SAP's internal software development process PIL, the Product Innovation Lifecycle, in an ontology that captures the inter-dependencies between the process steps, the produced artefacts and their nonfunctional properties. Reasoning can then be used to automatically infer the relevance of certain process steps or tasks in the context of the chosen solution approach, thus avoiding timeconsuming but unnecessary steps.