May I 'nudge' you in the direction to a thread just below yours: http://www.worldcommunitygrid.org/forums/wcg/viewthread?thread=21613

Did not know until nelsoc clarified, that some sending of tasks is dependent on previous results in the same distribution. Those folk better be quick as quorum 10 is going to take a wee bit longer this way.

Logically, the work is dynamically generated to optimize the diversity of structures returned. The goal is I think to acquire 120,000 minimum, so who knows will the distribution be reduced when enough confidence is developed from the data generated.

When 19 modern day quads would meet all running 8 hours, chances are the number of seeds generated are larger than required, so yes, the objective could be to reduce the number of the initial distribution. It could be that this goes across a whole batch/gene instead of just one set of 19.

Somewhere knreed discussed a bit of the future in upcoming clients and implement the ability to cut jobs short if the target is achieved. You get the "In the beginning we collected the data to measure...." All jobs seem still to be send out. Once you see jobs being withdrawn and getting a status of e.g. "Other", you know things have found their way into practice.

This is theorizing on basis of what I read, may have remembered right or confused.

WCG Admin: Please delete this thread at your convenience