[Orgmode] Feasibility investigation: org-mode paper

[Torsten Wagner]

Hi all,

A few days ago I had this silly idea:

1. I create a template page, a kind of form, which assist me in writing 
directly in org-mode syntax. Different things could be predefined and 
others might be filled out.

2. I train a OCR resp. ICR to recognize this template with a hight 
accuracy. This might be easier then it sounds esp. since org-mode use a 
lot of keywords and key characters.

3. I write some lines of lisp code which allow (from within emacs) to 
scan the template by a paper scanner (or maybe use another digital 
picture source), perform some image processing, parse it through the ocr 
program, parse it through some org-mode releated parser and open the 
result in a emacs buffer.

This would allow in principle to write an org-mode file on real paper 
and get it into org-mode automagically.
Some of you might ask why?! Well, first of all "because we can" (its is 
still all plain text). Secondly, some of us might find this really 
handy. Think of this kind of boring meetings... hours of hours with 
nothing then a pencil and a paper in front of you. You know you will 
have to write a report, schedule, outline, etc. You know you will do 
this with org-mode and in front of your mind's eye you could already see 
the nice clear outline in org-mode style... however, you are stuck in 
that meeting for another hour... jotting down point for point and 
finally rush back to the office to type down all this. Others might find 
this interesting for conference meetings. There is normally enough time 
to write down clear notes clear and accurate ... thus why not directly 
in org-mode style.

I played with this idea (on a Linux system) and different tools come up 
to be handy.
Imagemagick can be used to improve the image quality for ocr and works 
perfectly in the background. I created a very first simple template with 
LaTeX. Scanning can be done easily from command line as long as the 
scanner is supported under Linux.

The real problem is the OCR or ICR. It seems there is a lack of really 
good OCR programs and handwriting recognition is even more poor.
I would be happy if someone can point me to some solution here.

As for the org-mode integration, I would wait and make it working first 
before I start to put my head into lisp-code.

In the meanwhile, I thought if handwriting is not supported well enough, 
there might be a kind of hand-writing assisted org-mode.
Printing out an existing org-mode file (maybe after a special form of 
export). Make certain annotations during the meeting e.g., stroke a 
line, stroke the keyword TODO and add DONE at the top, etc. and run this 
throw the process to update the real file. This would limit recognition 
to single keywords.

As I said it is just a silly idea, maybe not really useful but I thought 
it would be at least nice to make a "proof-of-principle". Finally, with 
the increase of ebook readers with touch screens, digital pens, tablets 
and netbooks with touchscreens, this might become even more interesting.

Happy to hear any opinion.

Torsten