[OAI-implementers] Handling set membership changes

[Tim Brody]

See:
http://www.openarchives.org/pipermail/oai-implementers/2003-October/001045.html

And:
http://www.openarchives.org/pipermail/oai-implementers/2003-October/001046.html

Its a bug with no elegant solution. As I said in those threads I think 
the only practical solution is to create new items & flag the old as 
deleted.

All the best,
Tim Brody

Chris Wilper wrote:

>Hi all,
>
>So, I was looking at what it would take to get Fedora's OAI provider interface to support sets for selective harvesting and I realized that communicating set membership changes to harvesters could be trouble.
>
>Is there a sane way to deal with this?  The only thing I can think of is to just  treat the set membership change as a record modification, but this leaves certain harvesters in the dark.  They will believe the record still exists in its prior set forever and hasn't been modified forever, unless:
>
>  They are smart enough to recognize a setSpec change in a header, and
>     a) They are lucky enough be selectively harvesting 
>        on one of the new sets the record is in, or
>     b) They only do incremental (not set-selective) harvesting.
>
>-or-  
>
>  They periodically dump everything and do a full harvest.
>
>If I know up front that set membership is apt to change, should I:
>
>  1) Not support sets in my provider.
>
>  2) Support sets, but have a policy that if someone is harvesting
>     from me, they agree to either:
>       i)  periodically dump everything and do a full harvest, or
>       ii) never do set-selective harvesting
>
>  3) Something else
>
>Any recommendations or opinions on this?
>
>Thanks,
>Chris
>_______________________________________________
>OAI-implementers mailing list
>List information, archives, preferences and to unsubscribe:
>http://oaisrv.nsdl.cornell.edu/mailman/listinfo/oai-implementers
>
>
>  
>