Digital annotation is too formal

Computerized annotations are too literal and constraining in their data structure. For example, they represent annotation using precise textual ranges. But when I mark up a book, I’m often gesturing at a general area, not an exact range. Often there is no exact range that corresponds to what I’m referencing. To specify an exact range would be misleading (and distracting in presentation).

All this is very natural in the physical world. I’ll write in a general area and use arrows to point to various phrases. I can add stickies and make scratch marks and dog-ear pages. But digitally, we have the Tyranny of formality in interfaces.

This kind of strict structure also makes inter-textual commentary much more difficult. When bringing two texts in conversation with one another, it’s relatively rare that you want to make a point by referring to two precise textual ranges. You need to talk about broader strokes of argument, supported by “passages”. You might provide a pull-quote to illustrate a passage’s argument, but you’re really referring to the passage, not to the precise quote.

LiquidText and MarginNote are the interesting exceptions here.

Last updated 2023-11-20.