A narrative-based collaborative writing tool for constructing coherent technical documents
One important feature of an effective document that makes it easy to read and understand is known as coherence. Technical documents produced collaboratively are often incoherent due to a lack of group consensus and misaligned contributions by the individual authors. However, current document planning techniques and writing tools do not provide explicit support for improving coherence. The goal of this research, therefore, is to develop and evaluate a new technique and tool that helps teams of authors to structure coherent technical documents. The coherence of a document can be attributed to the story (or narrative) it conveys to the reader. If this story is consistent and coherent, the same can be said about the document. A discourse theory such as Rhetorical Structure Theory (RST) that has been developed by linguists helps further to analyse and improve a narrative. RST explains the coherence of a text by virtue of relationships (such as “paragraph A justifies paragraph B”) between parts of the text. This research has combined the ideas from these parallel strands of research to develop a new document planning technique called narrative-based writing. The method involves writing down an explicit précis of the story (called a document narrative or DN) and then analysing it using RST. The DN and RST analysis are then used to structure the eventual document. To extend the usability of narrative-based writing to geographically-dispersed authors, I have designed and implemented a collaborative tool that allows co-authors to edit, analyse and review DNs. The thorough design for the tool uses a combination of three models (conceptual, business process and functional) culminating in a set of functions that enable collaborative narrative-based writing. This dissertation discusses how, in the future, these functions could be incorporated in existing collaborative writing tools. Implementing this tool, albeit in its current prototypic state, has been invaluable in understanding the complexities of modelling and manipulating DNs and RST structures. Initial investigations using the new technique and tool have been positive, encouraging me to continue the research and evaluation in this field.