FWIW, since the comment was a reply to my message above:
It provided value by answering my question concerning serious downsides to providing optional post-TTL last-known-good caching within a DNS resolver. The answer is implicit in that a major DNS resolver provides exactly this functionality.
Don't want to say much more due to it being my job, and I don't want to give away too much.
EDIT: https://www.google.com/patents/US8583801