The one feature that would just MAKE my PKM would be if Logseq could basically do for epubs what it can do for PDFs.
I spend hours every day reading epubs, highlighting them, adding notes. It's almost all in KOReader, but it ends up trapped there.
When I highlight and annotate PDFs in Logseq, they become connected with allllll my other notes. I even got a system running for scanning paper I receive to PDF, adding an OCR layer, and importing to Logseq.
But I spend something like 200x the amount of time reading epubs as PDFs and I haven't found any local/FOSS tool that can bridge this gap.
Web Annotations would in theory work better for epubs than they do for pdfs.
Making web annotations in PDFs is either coordinate based (page + rectangles) or text bases (quoted text). The quoted text in PDFs is error-prone because PDF is a layout format. Text quotations are more precise in epubs.
The base url for such annotations should be content-addressable storage, i.e. a hash instead of a plain url.
I spend hours every day reading epubs, highlighting them, adding notes. It's almost all in KOReader, but it ends up trapped there.
When I highlight and annotate PDFs in Logseq, they become connected with allllll my other notes. I even got a system running for scanning paper I receive to PDF, adding an OCR layer, and importing to Logseq.
But I spend something like 200x the amount of time reading epubs as PDFs and I haven't found any local/FOSS tool that can bridge this gap.