Tim Bray thinks about digitally-signed financial news in Atom. Then, Bill de hÓra sharpens the thought in a way that I agree with (sign it, don't make the integrity checking reliant on hop-by-hop secure transport). But Bill worries about bit rot and signatures: "...though I suspect aggregator chains will result in altered signed data while leaving the signature as-is, a kind of syndication entropy that will take a few years to clean out."
I propose an easy fix: an new atom:link relation of "signed-original". This is to prevent overloading "self", although you could use "self" if you squint a bit. The content at the given URI might or might not match the content you have in your hand, but a computer can tell you what the differences are and maybe even guess if the changes are semantically significant. For that matter, "signed-original" could even be used when the Atom entry itself was not signed.
There's no need to start moving this through the standards process until we have a bunch of Atom readers who can validate signatures and, more importantly, give useful messages when a signature is invalid. Once that is there, signing Atom for financial news seems like a slam dunk application.