http://escatter11.fullerton.edu/nfs/workunit.php?wuid=7736835
http://escatter11.fullerton.edu/nfs/workunit.php?wuid=7736939
http://escatter11.fullerton.edu/nfs/workunit.php?wuid=7736250

I'd sent a PM to Greg about this when it was only one workunit, but now it's happened twice more. It seems that once a workunit is reissued as a new task, the person with the original workunit is able to submit theirs and get credit, even when it's three days after their deadline. My result is considered a duplicate even though I'm well within the deadline assigned to me.

I didn't think it was supposed to work this way - seems like some sloppy programming somewhere. Am I right to be upset over this?

Quite honestly I really don't know the details of how BOINC handles recycling of workunits. I just know they eventually get done and the results arrive in my data directory. I have not changed BOINC's default behavior.

''Redundant result'' means the server cancelled your task before you had a chance to start it as it was no longer needed. So you have lost nothing. If you look at the other host in question in your examples they submitted this task late and the 2nd person got the credit.

---

Hmm, the client state does indeed say "Cancelled by server". That hadn't crossed my mind - I thought my computer had done the computations only to have them ignored. So it's not really a problem. I'll chalk up the accepting of late results as one of the idiosyncrasies of BOINC.