I've been reading some academic papers recently and taking notes in org-mode.

What I find is that a lot of information expressed in solid paragraphs becomes a lot easier to parse once you convert them into an appropriate list, tree or graph structure. A surprising amount of the text I've been reading is lists that have some relation ("x is used in x_1,...","x consists of x_1,..."). This ties in to authors deciding to shake things up by adding unnecessary complexity to these simple lists. This kills the reading comprehension.

For example

    K. Merton developed in the 1940s his nor-
    mative framework for the conduct of sci-
    ence based on universalism, communalism,
    disinterestedness, and organized skepticism

becomes

     K. Merton developed in 1940s a formative framework of scientific conduct based on:  
     - universalism : <definition of universalism> 
     - communalism 
     - disinterest : <note on disinterest> 
     - organized skepticism : <example of organized skepticism>

A nice bonus is that that allows me to annotate specific items in these lists very easily, and visually match up items to notes quickly. (Note how that statement itself is a list due to being joined by an "and").

I've been thinking similar things. Also, diagrams are underrated.

Maybe "reading" is a low tech invention that we just don't have the patience for anymore. Instead we want to "explore" a theory or network of statements or social graph.

It's pleasant to click around in Wikipedia and get a sense for how different topics relate, but reading the (often turgid and boring) prose seems like work.

I have to admit, diagrams are a bit harder to do in text, especially interconnected graphs. That said, I could imagine something that uses triples to define a rendered graph, similar to how RDF for the semantic web works. Say

     #BEGIN_GRAPH
     'Paul Graham' founded YC 
     'Sam Altman' leads YC 
     YC funded AirBnB
     YC funded Reddit
     AirBnB <> Accommodation
     #END_GRAPH

Leads to a graph like

      [Paul Graham]----founded----->[YC]------funded----->[AirBnB]<------>[Accommodation]
      [Sam Altman]-----leads------->[  ]            `---->[Reddit]
      

Now that could be prettier, but I am too lazy to copy and paste box drawing characters. With a simplified form like that, all the data is in the form of subject predicate object sentences, which makes for easy parsing. The data could also be fed into a diagram making application like yED for pretty graphical output.

Great idea. With the flexibility of org-mode, you could even turn those notes in spaced repetition flashcards using org-drill.