Hide menu

Integrational Software Engineering



Uniform XML Composition of Documents with Semantics

Project leader: Peter Fritzson and Uwe Assmann

PhD students: Adrian Pop and Robert Kaminski.

Experts predict that every entity found on the web (XML data, software such as applets or 3-tier applications) will be coalesced to active documents (active components, docware). Active documents consist of XML tags with both static and dynamic semantics. Active documents are structured documents in the spirit of OpenDoc and XML but have a life of their own. They will be able to migrate, they will be shipped everywhere on the web, they will be embedded into other active documents, etc.

Ontology languages, such as DAML (www.daml.org), provide rule based mechanisms to specify class hierarchies and constraints on the classes. On the other hand, in static semantics of programming languages, Natural Semantics (Kahn 1987) is one of the most advanced, but still easy to use, specification technologies. It has been successfully used for such complex semantic tasks as compiler frontends, e.g. using the RML tool (Pettersson 1999), and additionally allows for the specification of dynamic semantics, i.e. interpreters.

For active documents, specification languages for both static and dynamic semantics will be indispensable. Hence, this work applies Natural Semantics. Since the technology will be integrated with existing languages, such as RDF and DAML, Natural Semantics will be one of the first l anguages for static and dynamic XML semantics.

During the initial phase of this project in RISE1, a preliminary study was done concerning the feasibility of compiling DAML and OWL to efficient code based on the RML Natural Semantics implementation mechanisms.
Recently we have focused more on industrial product models and their XML representation and composition. Such XML documents need to be composed, but also have formal properties amenable for certain semantic checking. Many product model attributes and constraints can also be expressed in an object-oriented equation-based language such as Modelica. Therefore, a ModelicaXML representation is being developed, together with a technology for composition and transformation of Modelica models using ModelicaXML, and database storage (ModelicaDB).

Experience from large-scale usage of Natural Semantics (e.g., the large Modelica specifications/documents in Natural Semantics) has clarified the need for better tool support, which led to the development of a prototype RML debugger being able to handle large specifications.
A preliminary prototype and techniques for restructuring general XML documents, DocumentRestructurer, are being investigated.

One experience so far from RISE is that developing large-scale industrial semantics specifications needs improved tool and language support. To address these problems, and to investigate the earlier mentioned composition issues, we are also doing ongoing work on the following software subsystems:

  • RML / MetaModelica: Compiler and Debugger for Natural Semantics represented as the Relational Meta Language (RML).
    The system is extended with better support for large specifications of industrial relevance including named patterns, and more efficient debugging. Moreover, a frontend for Meta-Model specifications in MetaModelica is developed to support integrated system modeling.
  • Eclipse plugin for RML/MetaModelica, to enable browsing and debugging or large specifications within the Eclipse development framework.
  • Modelica-DB, a database representation of product models imported from XML representations.
  • Document Restructurer, a document restructurer that also checks semantic constraints and includes document type-specific operations.

Page responsible: Webmaster
Last updated: 2012-05-07