Ontology-Driven Management of Change
Final Report Abstract
Nearly every kind of workflow in society, jurisdiction, economy, politics, technology, education and science produces a huge amount of documents for documenting events, plans, knowledge and results which persist in multiple versions being continually adapted, rewritten and adapted to changing situations. Change management is a critical issue, its difficulty is aggravated by the fact that the documents are inter-related. Although most documents are digitally born nowadays, tool support offered by document management systems typically is too coarse grained and agnostic of the semantics of the documents. As a consequence, more often than not any change in a document requires manual re-checking of all possibly related documents. The OMoC project developed methods and tool support for a semantic management of change for informal, though technical sets of documents. These documents have a structured content and specific dependencies that can be used to localize changes and propagate their impacts along the dependencies in a fine-grained manner. Parameterized over document-type specific ontologies formalizing the semantically relevant document structures and dependencies, OMoC followed a model-based approach to change impact analysis for heterogeneous collections of documents and any kind of electronic artefact in general. It is based on the explicit semantics method which consists of juxtaposing the documents and their semantic content and dependencies using the document ontologies. The semantic content is computed from the documents by document-type specific analysis rules that establish the fine-grained juxtaposition between semantic entities and their syntactic origin in the documents. Upon change of the documents, the juxtaposition is maintained and enables mechanical processes to distinguish how the semantics has been changed and to exploit this to selectively propagate these changes throughout the other parts of the same document and parts in related documents by document-type specific impact analysis rules. Since only the document-type specific aspects of the semantics but not the entire semantics of the documents (typically formulated in natural language) is used, the impact analysis and propagation remains possibilistic, but on a fine-granular level. To consider only semantically relevant syntactical changes in the impact analysis, the difference analysis between two versions of a document is parameterized over document-type specific semantical equivalence classes of syntactical changes. The methods has been implemented in the GMoC tool and applied to analyse the impact of changes in informal documents such as lecture slides, mathematical wikis, documents written in the DocBook format, in formal documents like heterogeneous formal specifications and annotated C program code as well as in mixed informal and formal document collections like the documentations, specifications and program code artefacts typically occurring in software development processes. The evaluation confirmed the adequacy of the generic approach and model-based methods to provide model-based change impact analysis for heterogeneous collections of documents. Based on the results of the project, we identified the following issues for future research: The approach taken in OMoC allows to design and apply impact analysis, but no guarantee can be given to the user about the completeness of the impact analysis. This would require to develop formalisms to specify the requirements for impact analysis and proof support to prove that impact analysis actually meets the requirements. Furthermore, so far difference analysis uses generic descriptions of the changes made to documents; to improve the impact analysis it would be desirable to have document type-specific change descriptions. The first phase of the OMoC project provides the methods to design impact analysis, but how to integrate this with specific workflows in practice requires more research. Already at this stage the results of the OMoC project seem highly applicable in industry. The principal investigators consider industry projects with German software companies that will instantiate parts of the results in an industrial context.
Publications
-
Semantic management of heterogeneous documents. In Proceedings of the Mexican International Conference on Artificial Intelligence (MICAI-2009), number 5845 in LNAI, pages 1–14. Springer, 2009
Dieter Hutter
-
TNTBase: a versioned storage for XML. In Proceedings of Balisage: The Markup Conference 2009, Balisage Series on Markup Technologies. Mulberry Technologies, Inc., 2009
Vyacheslav Zholudev and Michael Kohlhase
-
Change Management on Semi-Structured Documents. PhD thesis, Jacobs University Bremen, 2010
Normen Müller
-
Dimensions of formality: A case study for MKM in software engineering. In Serge Autexier, Jacques Calmet, David Delahaye, Patrick D. F. Ion, Laurence Rideau, Renaud Rioboo, and Alan P. Sexton, editors, Intelligent Computer Mathematics, number 6167 in LNAI, pages 355–369. Springer Verlag, 2010
Andrea Kohlhase, Michael Kohlhase, and Christoph Lange
-
Semantics-based change impact analysis for heterogeneous collections of documents. In Michael Gormish and Rolf Ingold, editors, Proceedings of the 10th ACM symposium on Document engineering, DocEng ’10, pages 97–106, New York, NY, USA, 2010. ACM
Serge Autexier and Normen Müller
