I just finished scanning my old bullet journals so I can put the physical copies away in storage and free up some space.

I used an old Epson Perfection 1650 flatbed scanner for the project. The scanning software I used is called simple-scan. Setting it to scan multiple pages allowed me to scan a whole journal in one batch without having to manually start each scan.

After the batch scan, I would save it as a pdf and run it through OCRmyPDF (which uses tesseract), to embed the text in the PDF and make it searchable. To my surprise, the tool had absolutely no recognition of my cursive handwriting. However, my print handwriting seemed to be perfectly recognized and converted to text.

According to a stackoverflow answer, tesseract can be trained to recognize handwriting, but accuracy tops out at 90% (not good enough in my opinion).

I've barely looked back at my old bujos since I finished them, but through the process of scanning them, I've remembered some of my old journaling habits and started doing them again. I think the ability to search my old bujos on the computer would be very useful. For that reason, I have started using all print handwriting in my bujo, even though it feels slower and more clunky.

I might switch back to cursive if LLMs get good enough to read my handwriting in the future.