In working on the dynamic view’s citation analysis space I had a hard time thinking about exactly what useful views should be possible, apart from chunking and labelling and timeline. I found Advanced Visual Analytics Methods for Literature Analysis which had some interesting ideas but the main point, which is quite obvious, is that it’s easier to to do ‘distance reading’ if you have access to the contents of the documents. For example, in this mockup, which doesn’t use documents, just headings but let’s pretend they are documents, the user has performed a search (cmd-f and type, then ‘enter’) for the keyword ‘Hypertext’ and it appears on the screen with lines connecting it to where that text appears in the other nodes (which could be documents/citations). This can of course only happen if the system knows what the nodes contain.
Therefore we are adding a text field in Author for the user to paste any/all the contents of the document to enable this analysis but that text will not be readable or extractable by a user, to hopefully allay copyright fears.