>>So then the conclusion is; since the optimization may not work under certain conditions we should prevent it from working under all conditions?

>Certain conditions ? I think this could be replaced by many conditions. I really do think that only in a few cases rushmore would benefit from a tag on DELETED(). I've never seen a case where a tag on deleted() lead to a significant improvement of performance other than where tables contain more than acceptable amounts of deleted records.

Though an tag on DELETED() would help with situations like this:

BROWSE FOR DELETED() OR a Browse in a table over a WAN With lots of deleted records clustered together (like in the beginning of the file).

But of course these situations are not so common.... :-)

Regards, Walter