I think I should focus on PhD on what I originally planned: The ‘switching between views’ more than the views themselves or data extraction. I will of course still focus on how to augment a student’s ability to do their Literature Review, but as a deliberate act of the student adding citations to their document and more manually pointing out the connections which are useful to them.
I am very concerned about how to extract metadata from academic PDFs in order to allow for a semi-automated information flow for the users citation documents and I think it would be foolhardy to work to build extracting systems and what I have found online does not seem maintained or robust.
I will support people and organisations, documents and the general category of concepts. There will be many issues but the key issue is how to fluidly switch between the basic word processing view and the dynamic view(s) which will be spawnable from different spark points (clicking on glossary terms in the document and pinch out etc.). The research question then becomes:
Based on the work in the field of symbol manipulation, hypertext and spatial hypertext in particular,
and novel interaction design and testing,
this project will investigate optimal ways for an academic user to change the view of their (primarily textual) documents to augment the process of doing a Literature Review,
with a clear design objective that the user should feel that the information is tangible and has a shape they can grasp,
in order to generate richly a richly connected knowledge space and then to linearise the multidimensional knowledge into a traditional academic thesis which will support the student’s aim of demonstrating their thesis examiner that they have done a thorough job of examining the knowledge space, while preserving the full interactivity for the reader who wishes to access them.
Now that I have the basics of the dynamic view in Author I think I have a space to start experimenting and testing based on what I have learnt in my own literature review on how to most effectively manage the view shifting process.