Add support for trickle up messages that automatically extend the deadline like how RNA@home does. This would be useful for when you send out a giant work unit by mistake so that the user does not have to worry about deadlines that cannot be met. Also, those trickles should contain all of the work up to the point of the trickle so that if the work unit crashes, the work that managed to be tricked up and is validated could allow the creation of a work unit that does not repeat the already validated work but contains the remainder of the work.

Add support for trickle up messages that automatically extend the deadline like how RNA@home does. This would be useful for when you send out a giant work unit by mistake so that the user does not have to worry about deadlines that cannot be met. Also, those trickles should contain all of the work up to the point of the trickle so that if the work unit crashes, the work that managed to be tricked up and is validated could allow the creation of a work unit that does not repeat the already validated work but contains the remainder of the work.

Sounds interesting. I will have to look into how much work this would be. The easiest solution might be to just not create long running WUs. Recall, this project has been running for several years and these problematic WUs only occurred several months ago (due to a bug in the WU generator).

Climateprediction.net already does trickled up work. That way, work that is already trickled up before a crash (either the program crashes or it detects a disallowed condition in the climate model like the oceans or atmosphere boiling away and therefore decides to terminate early) can be saved. Unlike RNA@home, its trickle model does not extend the deadline because a very generous deadline is already given.