Document engineering of complex software specifications

Abstract

The research presented in this thesis aims at document engineering of complex specifications, of which the UML Superstructure Specification (version 2.1) is our initial target. Document engineering deals with principles, tools and processes that improve our ability to create, manage, and maintain documents [40]. Our motivation is that such specifications are dense and intricate to use, and tend to have complicated structures with lots of repetitive, or 'boilerplate' material. End users cannot use them efficiently because of the general complexity of the document. Our objective and main contribution in this thesis is therefore to create an approach that allowed us to re-engineer PDF-based documents, and to illustrate how to make more usable versions of electronic documents such as specifications, conference proceedings, technical books, etc so that end-users to have a better experience with them. The first step was to extract the logical structure of the document. Our initial assumption was that, many key concepts of a document are expressed in this structure, which includes the headings of the chapters, sections, subsections, etc. We demonstrated this by analyzing some data, and created a special-purpose parser to generate a well-formed XML document with various types of tags. In the next phase, we created a user interface for end users by generating a multi-layer HTML version of the document to facilitate document browsing, navigating, and concept exploration. Although our targeted document was the UML Superstructure Specification, we chose a general approach for most phases of our work including format conversions, logical structure extraction, text extraction, hypertext generation, etc. Therefore, by minor adjustments we can process other complex documents to gain our mentioned goals. We also established the major infrastructure for a new document engineering framework.