Dave:

The PDF archive format can preserve a block of bytes representing the original document plus the signature dictionary which contains information about the hashing algorithm, the key and signature values in a manner that they can be preserved and tested in the future. The exact mechanism is very complex and includes callback to test as the file is being written out to disk to ensure no tampering occurred between the time it was signed and the persistence to disk as well as other safeguards.

The PDF itself could be signed again thus making two certification events per document. Multiple signatures on a document have extra complexity as you first have to certify documents. It basically works on a function v(function v(function v()))... Basis. The second signature or certification event includes the bytes used by the first set of signatures.

Based on currently acceptable algorithms and historic CPU breakthroughs, I would suspect that what people use today for Dsig is not what will be acceptable tomorrow for things like certifying documents.

I have a set of PDF slides on the PDF signature mechanism if anyone wants to understand this in more detail.

Duane

On 30/07/08 9:41 AM, "Dave Pawson" <dave...@gmail.com> wrote: