This is by far the best analysis of reading of electronic text that I have seen. He provides an excellent insight as to what is, what might be, and what is not important for on-screen reading.
He does not limit text to just fiction/non-fiction categories, but instead discusses: WHY it is read professional/personal reasons, to learn or not, out of interest/need, etc. WHAT type of information it contains technical or non-technical, subject matter, general or specific, textual or graphic, etc. HOW it is read serially or non-serially, once or repeatedly, browsed or studied in depth, etc.
His book suggests to me that text should alter its format to the meet the users - Why, What, and How. Possible examples: switch to all caps when searching for words or phrases, turn off hyperlink indicators for linear reading, ...
He points out that there have been many studies on editing text, but few on reading text. A good fraction of the book deals with on-screen reading.
Screen reading was better with: high resolution characters, increased space between lines (leading), proportional font, limiting the number of characters on a line, and not splitting a sentence across a page boundary.
He indicates that users preferred on-screen reading over paper reading for some tasks when the screens had enough improvements.
Screen reading might be improved with: landmarks/navigation, serif fonts, full left/right justification, ...
Screen reading was no different than paper reading for: orientation of the media, flicker rate, screen dynamics, and visual angles (< 36 degrees).