The problem with printing ASCII armored data on a paper is that some characters are looking similarly and this is a prerequisite for OCR errors.
This script is changing some characters with cyrillic and greek characters.
Another method for reducing the likelihood of OCR errors is by using RFC1751 encoding.
It's recommended to use OCR-friendly fonts like OCR-A and OCR-B. The font DejaVu Sans Mono is also good.
You can install them on a Debian-based distribution easy:
$ sudo apt-get install fonts-ocr-a fonts-ocr-b fonts-dejavu
I don't like that "0" and "O" in the OCR-B font are looking too similar. In DejaVu Sans Mono they can be differantiated by the dot in the zero.
Fonts like "Courier" have too similar characters - "l" is looking like "1".
Translation table:
Q ⇨ Я
0 ⇨ θ
l ⇨ л
I ⇨ Д
G ⇨ Ж
$ enhance-ascii-armor.py --action=encode < test.txt.asc > test.txt.asc.enhanced
$ gedit test.txt.asc.enhanced
$ cat test.txt.asc
-----BEGIN PGP MESSAGE-----
jA0ECQMKKEnwbwZvywT/0kQBvTttT+fvKRQk0qz1MRUb3sLM8lisu90LCe8jOgQd
MT1oFmbh4k+1ViD0zqjUd76kMuSCEP+bxx92ykCe3laoIZRquA==
=C3bh
-----END PGP MESSAGE-----
$ cat test.txt.asc.enhanced
-----BEЖДN PЖP MESSAЖE-----
jAθECЯMKKEnwbwZvywT/θkЯBvTttT+fvKRЯkθqz1MRUb3sLM8лisu9θLCe8jOgЯd
MT1oFmbh4k+1ViDθzqjUd76kMuSCEP+bxx92ykCe3лaoДZRquA==
=C3bh
-----END PЖP MESSAЖE-----
$ enhance-ascii-armor.py --action=decode < test.txt.asc.enhanced > decoded.asc
$ diff test.txt.asc decoded.asc
$