What about the idea of requiring new Experimental documents to include text
that indicates when the experiment is to be considered completed absent new
work on it? Essentially, the document declares a date by which the experiment
is considered concluded, and code points automatically deprecated, and the
document itself goes to Historic status, unless some other document action
updates the deadline or moves the work to the Standards Track.
Keeping track of such dates to trigger these actions by default becomes a new
burden for the Secretariat, I imagine. I don't know what that cost would be
compared to the "cost" of experiments currently languishing in this unresolved
state.
To Eliot's point, work that is resumed much later could always restore document
and code point status and declare new drop-dead dates when it becomes
interesting again.
-MSK