Michael Snow wrote: I would point out that none other than Richard Stallman himself has addressed this topic (I do need to see if I can find the source.... I heard him explain this at a conferece regarding the GFDL where I was in attendance). The problem with considering a paper copy as a non-opaque source is that somehow the content needs to be transcribed or scanned using OCR techniques to be edited again. Very few fonts are explicitly designed to be easily read with OCR technologies (not completely unknown, but not commonly used). Mainly the issue here is that the process of transcription... or even cleanup from a clean OCR scan is a very labor intensive process that is unnecessary in the era of digital technologies. Even with the transcription issues, you still have to put in the formatting information as well, not to mention that the error rate for scanning even a very clear document with a good font is very high. My experience with Distributed Proofreaders is that clean scans still have about a 1% error rate, messing things up like c and o, or mistaking punctuation like the comma or period. And these are just common problems that happen with a good OCR scan. In short, a printed paper version is an opaque format and requires the "source code" somewhere else. I'm not sure if a 2-D barcode that would hold the full contents of the document would qualify as a as a non-opaque format, but that is something to think about as well.... if you don't want to include something like a CD as an appendx or maintain a website with the data. But the raw text alone is not sufficient.