I have noticed that since June 12 the return results have been going down over 15 days now. So I'm asking here.......

When a project comes to its end why not move the time for a result to three days or under and just close it out?? Or is it a case to see just how long you can keep a dead horse alive?? That's just dumb...

If it's it's time is up it's up reset the time requirements and kiss it good by you have a server aborted option just pull them all resend as if they were a beta test....top priority short lived.

Seems silly to just allow it to drag on when you seem to have the resources to change it.....this would also end your NO REPLY responses that just cause it's death to linger

I agree not all are no reply's and a good number require quorums, however server aborting them and handing them back out with beta test requirements to validated on current net connected computers would remove all the people that just loaded the programs just to see (no reply's) and all but cure buffer times.

Still the buffer reset default is a full day, 12 hours I think would be cutting out way to many resources. Even at that setting the work units to run at beta priority brings them next no matter what the buffer setting is. So the buffer times wouldn't matter in any case, if the server doesn't get it back in 24 hours the old copy gets aborted and sent back out. Even CEP2 with a large work unit and a full day of Netflix almost never requires over 24 hours.

I'm just thinking of how it must feel for the project heads to have to sit and wait for those few last results to trickle in.....that and why such a thing was never even considered. I only know if I was in charge it's not the only thing I'd change. What I'm sure of is they (the wcg team) are mostly winging it and they seem to not care what we think as long as we think well of them.

maybe resends should be with shorter deadline...like we had resend from a project 8xx, 9xx in FA@h...just not give 20days, make it a half - 10d deadline...

& program that any resends of those resends have a half of that deadline...so if 10d deadline doesn't compute - a 5d deadline will be resent! etc.

I don't agree.
Let us assume that a workunit was sent out to two recipients with 10 day deadline.
One of the recipients aborts the workunit or returns an error within a few hours.
The workunit is resend with a 3 day deadline and eventually causes panik mode on the new/third recipients client. In the end normal workunits on the third recipients client end up in panik mode too.
A short deadline for resends is ok if the two origin recipients return wu with error.

I do wonder if somewhere there'a group of scientists watching a digital beaker at 99% with a drop hanging from the tap but not yet falling...