Adam Back wrote: | Technical questions: If this is to include uuencoded or radix-64 mime | encoding, we might want to think about redundancy to allow error | correction. Perhaps we want that anyway to ensure that what we have | is 100% character-by-character perfect. Or perhaps not as it may | damage the legality aspects. They may start saying that you can only | export human readable stuff on paper, etc. Then we move on to `texto' | apparently human readable steganographically encoded paper based | remailer messages. The place we really want the redundancy is in the alphabet used, not in the data. Most OCR systems have clever algorithims to figure out that that blob after a 'q' is really a 'u'. To take advantage of this, you could encode everything in RFC1751(?, S/Key style) word lists. The expansion factor is extreme, so use gzip --best. Alternately, you could turn off context sensitivity on your scanner, and use an alphabet of abcdfgijknopqrstuvxyz (depending on your font--in lucida these are all pretty dissimilar, using a hueristic of 'more than one led bar' different.) With some experimentation, you might be able to expand that a bit with punctuation and numbers. Adam -- "It is seldom that liberty of any kind is lost all at once." -Hume