A very nice project but... How many really want to scan to text handwritten text? Results will be messy anyway and typically today handwritten text is not more than few pages, far quicker to retype or even dictate than correcting OCR.
BTW personally I use Xournal++ to add text/images to pdfs, typically where I have some crappy low importance pdf-form, not a real one, and I do not want to invest time in a nice LaTeX + cart.el (Emacs artist mode wrapper to get coordinate of any form clicking with the mouse on them [1]). I still have to do with some scanned documents but originally printed from a computer not handwritten.
Handwritten text recognition might be very welcome to scan and index old public archives, witch is damn complex since there are countless of style of cursive, but it's still a very needed thing to merge the old paper world to the digital one not to loose history.