Problem with faked-system-time option

Robert J. Hansen rjh at
Thu Jun 16 02:29:49 CEST 2011

On 6/15/11 8:11 PM, Jerome Baum wrote:
> Actually, no. It's still evidence. It's just not evidence that says
> "this document existed on the 10th of July". It's evidence that says
> "someone who was at any point in time in possession of this key, and
> who has the ability to manipulate the timestamp (or was in possession
> at the actual time listed in the signature), is claiming that the
> document existed on the 10th of July".

You're moving the goalposts: you're changing the terms of what the
statement is meant to prove.  By moving the statement from "A is true"
to "Tom claims A is true", you've moved it back to something both
parties stipulate -- presumably, Tom and Jerry both will agree that Tom
claims A is true.  Etc., etc.

Simply put, if both parties agree that "X, and X indicates the truth of
A, therefore A," then X is evidence for A.  As soon as one party objects
to either the fact (X) or the deduction (the inferential rule), then it
becomes simply a claim to be evaluated by a jury.  Claiming that "well,
it becomes evidence, just evidence for a different thing," is the sort
of thing that, while true, kind of misses the point.

> What tricks would those be?

At this point, you need to ask a contract lawyer.  I know enough
contract lawyers to be certain they are very on top of their game, and
they've already given thought to the subject of repudiating electronic
documents -- but I haven't been fascinated enough to go to law school,

> Oh, and I am pretty sure that a timestamp created in accordance with
> federal electronic signature laws would be considered reliable and be
> very difficult to dispute.

As I am given to understand, the major problem with creating documents
in perfect accordance with all extant law and precedent is just how
difficult it is to keep track of everything that's applicable.  It
sounds nice and simple to say "create something in accordance with
signature laws and applicable precedent," but my understanding is that's
a highly nontrivial task.

More information about the Gnupg-users mailing list