Exception: could not get assignment

[bruce]

Stanford is aware of the recent shortage of projects designed to run on higher CPU count machines. They're gradually making more projects available but some projects will not scale up proportionally as the CPU count goes up and people complain about that, too, so some restrictions are required.

Yes, people have been known to abuse the points system but in some other projects, they've also abused the scientific results. Another well-known Distribute Computing site has decided that every piece of work done by one donor must be duplicated by someone else before the result can be approved. FAH has been successful in getting good scientific results while also avoiding discarding more than half the work our donors produce.

Not-FAH Example: What is 2+2?
Donor #1: Four
(Reissue the WU and wait until somebody else reports an answer)
Donor #2 Three
(Bad hardware or attempted fraud or some other error detected)
(Reissue WU again.)
Donor #3: Four
(OK, discard 2/3 of the work and accept "Four" as the right answer.)

Same example, alternate results
Donor #1: Four
(Reissue the WU and wait until somebody else reports an answer)
Donor #2 Four
(OK, discard 1/2 of the work and accept "Four" as the right answer.)

FAH example:
What is 2+2?
Donor #1: Four.
(Perform both open and non-open result validation steps and accept or reject the result.
Reissue the WU only if a failure is detected.)