Am i wrong or is the fact that workunits are being processed at a rate that frequent delays in delivery of workunits are a good sign?
To me it means that enough people are donating that this is actually a problem rather than not enough people and a surpluss of work units.
I know i know, its probably due to people ditching less-valued units and what not which we obviously DONT want; however, it stands to reason as the F@H community grows that this lack of resources is, at its core, a good thing!
Ditching less-valued units may be important, or not. I can't really tell. I do know that acedemic research goes in cycles. Projects finish, researchers write/edit papers. new areas of study are proposed, and new projects are started. If all of the researcher are in different phases of that cycle, there will always be new projects coming along to fill in for the ones that are completed but if the new projects don't come on-line when the old projects are completed, there may be a shortage.
Also, servers do have problems that need fixing. If there are a number of servers that can issue WUs to your particular client, you never notice if one is off-line, but if there are few servers with work conforming to the criteria set by your client, you may experience a shortage when others do not.
The recent ATI shortage is an instructive example. WUs for the ATI GPU client generally have come from two servers. One server has been out of work (possibly related to my first paragraph). The other server had one or two other types of problems so the clients didn't get work, even though there were projects that needed to be processed. As far as I can tell, there was no CherryPicking (ditching WUs) going on. Repairs to the server were completed (during a holiday weekend, by the way) and assignments have been flowing freely.
