Nathan, love the concept here and look forward to playing with the code. The notion of performing operations on remote documents accessed via stub documents reminded me of the conversation Chris Blatnick are having about activity tracking. I wanted to put a few ideas out there so you might keep alternate uses for this code in mind during development.



Chris' Application Activity Tracking OpenNTF project also utilizes stub documents to track read, edit, and delete activity within a given database. An enhancement to that would be the ability to place those stub entries into a central tracking database (a la OpenLog). Doing things with the "source" documents from inside the central activity log db would work much the same way as in Haystack.



One exception that jumps to mind is what to do with deleted source documents. It should be no biggie to handle an orphaned stub document, but not handling it might cause errors. Come to think of it, even a Haystack search could run into this problem in an active database with lots of churn.



There are probably other subtle differences between the two types of stub document scenarios, but I can't think of anything else at the moment. It's enough that you are aware of another potential use for the code.



-Kevin