So many of the difficulties in reading many papers (well, disciplines) boil down to adding links to things the first time you mention them. Good academic writers do this, bad ones don’t. The gist of this paper is that assuming all knowledge prior to your development is very limiting, not only because far fewer people can read it, but because every time you don’t introduce knowledge you also skip over part of a story. Very well, but I think you are correct that CS’ problem is pretty much isolated to the first point, because nearly all papers get to talk about real world applications of the research and that’s the story covered.

A big problem for CS papers, particularly in PL (programming language research), seems to be heavy reliance on assumed knowledge of Greek letters and notation in the very field-specific way you like to use them. People would understand your paper if only they had read your previous three, which alluded to what you might have meant by these Greek letters, but only by figuring out the two citations they each have in common. If you are bad at pronouncing Greek letters and it’s a PDF so you can’t copy them, you can’t even google what you see. Even if you could, it wouldn’t help. Notation is ten times harder to search for.

(I have never, ever had this problem reading a law paper, not even slightly, not even once.)

There’s an interesting demo here from Will Crichton about how to prepare better documents for conveying understanding in PL. He has a thing to show you the “read as” on hover. https://twitter.com/wcrichton/status/1442891297333800966 https://willcrichton.net/nota