On Tue, Sep 8, 2026, at 04:11, Timo Gerke wrote: > Hi Richard, > > It looks like Richard Clayton aimed to say: > > > please stop digging > > To me it seems you don't like architectural debates. > > The inbound edge gateway (MX host) is the TLS termination > endpoint for the domain. Once the application-layer byte-stream > is decrypted at this boundary, the non-deterministic parsing > complexity introduced by polymorphic structures like the schema's > "oneOf"—as noted by Steve—remains an active resource exhaustion > vector regardless of the underlying platform.
We just don't like architectural debates with somebody who comes in here arguing that a truck needs to have the same aerodynamic properties as a bullet train, and when asked "but what about the tray capacity and cornering radius" keeps saying "BUT IT MUST BE AERODYNAMIC". Until you address my questions, I have no interest in your continued arguments about the moral equivalent of high speed airflow dynamics - because it just doesn't apply to the real world needs of this protocol. Provide a demonstration of the kind of hardware and software you expect this to run on, and a real world resource exhaustion proof of the current JSON schema, and I will engage with you again. Until then, I'm done with responding your line of argumentation because I don't think you're engaging with the actual needs of DKIM2 users in the actual world. Bron. -- Bron Gondwana, CEO, Fastmail Pty Ltd / Fastmail US LLC [email protected]
_______________________________________________ Ietf-dkim mailing list -- [email protected] To unsubscribe send an email to [email protected]
