In this comment on the PR for SPEC-1418, I pointed out that Bucket::find() is just a wrapper/proxy for Collection::find() on the fs.files collection:

IIRC, Bucket::find() is just a proxy for a find() operation on the fs.files collection and opens a cursor on the raw documents (nothing to do with opening download streams).

I don't recall why we defined GridFSFindOptions separately instead of supporting FindOptions in the CRUD spec, but the underlying operation should be the same. GridFSFindOptions was also never updated to support options such as collation. Was that intentional or just an oversight?

I might propose removing this and just creating a new SPEC ticket to consolidate and just reference FindOptions and the CRUD spec unless there's good reason to keep them separate.

I don't see any reason why we need to distinguish between find options for the GridFS or general CRUD API, unless some drivers actually implement these as separate code paths.