> FWIW, the definition of "Historic" is in RFC 2026,
That is *a* definition of "Historic", but it has been modified informally since RFC 2026 has been written. The IESG is quite aware of this problem and has steadfastly refused to update 2026 to bring it up to the current understanding of "Historic". Doing so would open many large cans of worms.
Proposal: ignore the "historic" label altogether and simply go to "should not be used in new applications and protocols". That way, you can word it as strongly as you want. Such a document can be a BCP, which has the same effect as a standard.