From: Peck Dickens <firstname.lastname@example.org>
After finally having about of hour of spare time to bang on gocr, I finally achieved joy and got it to process a few docs for me. Please notate the results were pretty damn good.
Assuming you get it to compile on your system here is what you do:
1.) Scan the document using SANE. You now have an image file name: yourfile.pnm
2.) At the $, type in "convert yourfile.pnm yourfile.pbm"
3.) At the $, type in "gocr -v 1 yourfile.pbm >newfile1.txt 2>newfile2.log
4.) once the above command finishes running, newfile1.txt is the file containing the ocr output. The file, newfile2.log is a text file that contains verbose output from gocr of exactly what went on during the conversion from image to text.
gocr appears to work pretty good... Amazingly high accuracy considering how young the code is...