You are here:
Abstract
Traceability focuses on the interrelation of information through all steps of a software process chain. It is required to integrate domain knowledge and knowledge derivable from software development processes and their artefacts. Traceability enables numerous software engineering activities such as change impact analysis, compliance verification of code, regression test selection, and requirements validation. Controlling the impact of environment changes on software systems implies the need of tracing business entities and rules to software artefacts. This goal requires support by conceptual descriptions of the application domainand appropriate traceability between application entities and software elements.
Mission
This work package will deliver a seamless approach of traceability from business entities and rules down to models and code by developing means to extract trace information between ontological entities and/or software elements. Furthermore, the work package intends to exploit given knowledge in order to infer traceability information that may be used to guide and stabilise the software development and maintenance process and to make software more easily adaptable to a changing environment.
Use Scenarios
Employing traceability, entities on higher abstraction levels of the model driven development chain, i.e. the abstract business part, requirements, and the CIM are connected to all kinds of models, from PIMs to PSMs, and down to concrete code. Then, given any modifications of the requirements from the business side, this horizontal traceability information helps to determine the impact of these changes to design, code, and test cases. Furthermore, unneeded functionality, and thus unneeded complexity, of the system can be detected.
Given one stage of a model driven development process, many different submodels form a big overall description, which has to be assured to be consistent and complete with respect to a vast number of conditions. The sheer size of such vertical traceability information makes an automated derivation of coherence information and an automatic reasoning-based approach to consistency checking necessary. With the aid of ontology technology these properties can be made explicit and operable.
Expected Results
The conceptual integration of ontologies and models implies the need for a dense interconnection of all artefacts in the development chain, typically called traceability. Originally, the term traceability refers to the ability to describe and follow the life of a requirement, in both a forward and a backward direction, thus linking requirements back to stake-holders' rationales and forward to corresponding design artefacts, code, and test cases. Here, the goal of MOST is to generalise traceability in a much broader sense: Business entities and business rules laid down in some processable format must be connected to domain knowledge and process knowledge and to all model fragments used through the model driven development chain.
Usually traceability information is explicitly introduced into software engineering artefacts using requirements engineering tools. There are also approaches to compute such information automatically using information retrieval methods or querying. Model driven approaches to software development introduce additional knowledge about dependencies inherent in the development process, for model-transformations imply such information. The automatic derivation of traceability information and its automatic maintenance during model driven development is essential for large systems and an important research topic. There are new challenges such as keeping models consistently synchronised, models and generated code in consistent states, and there are problems with mixing manual updates with automatic processes.
In MOST, the goal is to derive traceability information based on the analysis of model transformations. Process ontologies are used to infer additional linkage information between the MDA stages (horizontal traceability). Furthermore, domain ontologies support the integration of the submodels of a big overall model (vertical traceability). The traceability information is used to check all kinds of consistencies and supplies the overall view needed for guiding the developer and to avoid fragmentation.
Applications
Traceability helps to understand the many associations and dependencies that exist among software artefacts. Traceability information can be used for many different purposes, e.g. requirements management and validation, project monitoring and control, project costing, resource management, impact analysis, change control, contract compliance, development of appropriate verification procedures, quality assurance, resource monitoring, test plan development, and regression test selection. Depending on the purpose, there are different kinds of traceability relationships that may be taken into account.
Relationships to other Work Packages
This work package is heavily interdependent with WP1. In order to allow for seamless modelling using the MOST approach, the traceability concept must be integrated with the joint metamodel combining ontology languages and software engineering modelling languages. Furthermore, the derivation of traceability information relies on efficient querying and transformation mechanisms, whose conception is also part of WP1.
There will also be a strong interrelation with WP2 and WP6. In particular, WP2 will develop fundamental concepts for the guidance of the software process where the traceability technology of this work package can be applied. Furthermore, one of the SAP case studies will apply the traceability technology in a realworld SAP sefng. The work package will precisely define, how, when and for what purposes traceability information should/can be added, how it can be kept up-to-date, and how traceability information can be used to give guidance.