http://www.theatlantic.com/magazine/archive/MichaelApproved/...
The page canonical meta tag shows the current article url http://www.theatlantic.com/magazine/archive/2010/07/the-poli...
The previous article shows /2010/06 http://www.theatlantic.com/magazine/archive/2010/06/the-poli...
Maybe HN should use canonical meta tag instead of just url to check for duplicates?
Therefore, I suggest the opposite: automatic reposting - but only for those "great" articles, in the "HN library" (membership being defined by a special-case vote - i.e. vote for this article to be in the library). These would be reposted on a 6 month cycle, and include the old comments. Biannual resurrection. And to be marked as "library", so you can skip if you want.
But this won't happen, because the existing "bug" already does it (if people vote it up, it was a reasonable repost; the present repost checker act as a low barrier). And pg advocates iteration only for things needed.
It posts articles that make it into the top 5 of the site. Scroll through the past tweets to see what you've missed.
_Resubmissions are allowed after a year or so._
[beyond paraphrasing, completely reconstituting the thought, still meaningful however.]
This solves all the problems in your first paragraph. Automatic re-posting is almost a guarantee if it's that good, so that's paragraph two handled.