picturephoning writes about Fujitsu's "new" technique for embedding a barcode into an otherwise normal looking image. It has, however, already been done. At least one example, Dataglyph , was developed at least as early as 1995 (see citeseer ) and maybe earlier, I haven't tracked it all the way down yet. These techniques all amount to essentially the same thing, which is steganography . The main disadvantage of this technique for the kind of applications that Fujitsu, PARC, and others imagine is a simple matter of UE, user experience . If the users can't see the message, how do they know it's there?