As Eric quoted me in the Revision Control for all of CPAN proposal, "I'm looking forward to it."
Having the history of a distribution available publicly in a format which allows (or encourages) a history-like view is a great thing. That said, it's not an easy or simple problem to solve, but in my mind the benefits are not primarily offering public revision control to authors. (I can imagine offering read-only access.) The primary benefit is offering Backpan in a form much easier to use than juggling tarballs.
That wouldn't be as useful if all authors made their complete version control histories public, but first things first.