To enroll as tester / idea inserter go http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_alpha (posted this link for you before) and explain the problem and solution. Maybe amstrdj, knreed, uplinger, seippel and cleanenergy never discussed the case ** to include with Berkeley and failed to consider the pros and cons [e.g. user impact when 8 concurrent get saved at 8-16GB per interruption and reload]. Somehow the one big con could and will outweigh the wishes of the micro-manager all too over-concerned of a few hours progress reversal [notwithstanding the special notification sheet of why things are what they are].

P.S. The techs also explained in the case of another science, what is needed done at point of resume by the applications to ensure that a task remains integer. May not apply to all sciences [sizing up asteroids springs to mind], but I'd like to think that if we compute a compound docking, that all bits are pointing proper on resume and not one being off, the difference between a cure and a cancer.

** Actually they did, several times and surely more often then I know off.