secure data export

[Oleg Goldshmidt]

> The point of the additional file is to leave little room for anything else.
> Regarding the FAT place: Assuming the CD ends up on an infected machine, or
> falls into the wrong hands ( example: you want to make your client an offer
> on a CD, but you do not wish to give the client info about other offers you
> made, in this case the wrong hands are exactly the hands the CD goes to),
> the infected internal machine and the infected external machine agree on the
> interpretation of the extra space in the table sectors, and may communicate
> information through it.

Let's be clear about one thing. What is the primary concern:
preventing malware from spreading or preventing information from
leaking?

Depending on the answer some of the responses you've got may be more
relevant than others. E.g., I think that Shachar's comment about FAT
tables is correct in the context of malware propagation. If catching
steganographic messages is the point (as, e.g., I understood the
problem) then custom filesystem metadata is as good a channel as any.

I liked the idea of printing the stuff and OCRing it back, by the way.
A low tech / dead tree step in the middle is a good way to sterilize
bits. ;-)

-- 
Oleg Goldshmidt | pub at goldshmidt.org