Why the rush with returning work

[UBT - JohnR]

One of my computers has been sent 2 long running workunits. One of which is http://www.rnaworld.de/rnaworld/workuni ... id=4980864
The work was sent to me on the 18th March, but the original wu has until the 4th April to be returned.
3 further copies have now been sent. My computer has already done 24 hours work and has another 3 days work remaining to complete.
If any of the other 4 computers return before me then all my work will be lost as the "Redundant result" status will be sent to the remaining computers.
Why not wait for the first 2 computers to return/time out/or error before sending any further work, and then only send 1 copy at a time, perhaps with a shorter deadline? I could loose 3 days work in this one instance. My cache has been set to only 0.1 days to try to avoid this on short wu's.

[yoyo]

If you finish this wu, than you will also get credit for it. Even if somebody will finish it earlier. It will only be canceled on your computer if you havn't started this wu.
I sent this wu multiple time, because I must finish this archive and I'm not sure if realy each one who got this wu will also finish the wu. Often user just get away from the project and the Boinc server has to wait until the deadline to recognize it.
yoyo

[UBT - JohnR]

Thanks for the quick reply, now I understand that you have a short time scale for the archive, I can appreciate how the work is sent.