The "check for duplicates logic currently offers possible matches without regard for type, so that often a merge between a shortfiction title and a novel is offered. That is acceptable, because once in a while when a novel has been mis-entered as a short work (or vice versa) such a merge is the right thing to do. (This can happen when a large anthology is enterd from a secondary soruce and the entering editor incorrectly assumes that all contents are shortfiction.) It would be nice if there was some visual alert about mis-matched types, because often such a merge is a mistake.
However, it will also offer a match between fiction types and container types -- a collection may be matched with its title story; an omnibus may be matched with a novel of the same title; and now that support had been enabled, a chapterbook is quite likely to be matched with the corresponding work of shortfiction. it is pretty much never correct to merge a title record of a fiction type (NOVEL or SHORTFICTION) with a title of a container type (COLLECTION, ANTHOLOGY, OMNIBUS, or CHAPTERBOOK). Such matches should not be offered by the check for duplicates logic.
Anonymous
Clarified the summary.
As of r2009-73: Eliminated case sensitivity when running in "Exact" mode. The logic no longer tries to match SHORTFICTION records against COLLECTIONs, ANTHOLOGIES, OMNIBUSES or CHAPTERBOOKS. Further refinements under consideration.
Michael and I would like REVIEWs to be excluded from matching, due to people like P. Schuyler Miller.
edit/find_dups.py 1.9 submitted for consideration.
Reviews excluded from Find Duplicates in r2011-56.