In Tiger (Mac OS X 10.4), Apple will release a technology called Core Data. You can learn a bit more here: http://developer.apple.com/macosx/tiger/index.html.

My belief is that Core Data and QuickLite do not fight for the same space. The two products deal with data in two *very different* ways. Core Data is more of a black box, where the internals are completely hidden from the developer. It completely manages data objects behind the scenes, and it provides, among other things, undo/redo support. All this is performed via Bindings, which is not for the faint of heart.

On the other hand, QuickLite is a whole lot more transparent. You write the SQL query to select the data (that's the only time you'll write queries) and QuickLite creates the cursor. There you can retrieve the data bits per row/column, the row as a whole unit (QuickLiteRow), use QuickEdit (add, edit and delete data directly to the cursor), in-cursor data matching, perform set operations between cursor. Currently there are 3 cache methods available: DoNotCacheData, CacheDataOnDemand, and CacheAllData. It has BLOB support, is multithread/multiprocess savvyy, and uses distributed notifications. You can read more here: http://www.webbotech.com/quickliteoverview.html

QuickLite has its shortcomings too, of course: does not manage object graphs (relations, dependencies), does not provide undo/redo, etc. Still, it's a solid tool to deal with data the way *you* want.

Started a project with quicklite, and switched over to CoreData. Both are good, and each provides a different model for accessing/managing an application store. Quicklite is designed as an object wrapper over sqlite, while CoreData is a more generic data model, hiding the specifics of which store type (xml, binary, sqlite) is being used. If you like an sql sort of model (select, update, insert - though Quicklite wraps these up nicely), Quicklite is the way to go.

After figuring out CoreData, I prefer it's approach/model - but the biggest benefit of CoreData is kvm support, quick schema changes, and migration of data from different datatypes. Short thought; I think they can co-exist.