From my experience, and the posts above, I think that there are two use cases:
1) Someone posts a "plate" that has a bunch of separate parts, in a single file for ease of printing. The desire is to be able to separate the parts. The STL format is hierarchical, and I think this use case is splitting apart the top-level objects. This would be quite useful, quite often, as many people share multi-part objects packed into plates. The value of splitting one file into many is that the parts can be rearranged to fill a different sized print bed, or printed individually, or printed with distinct profiles. This tool looks like it does this, and it's open source (MIT license, so can be used in commercial software):
https://github.com/fsantini/python-stlsplitter . It'd be quite convenient to have this capability built into S3D, and since there's no UI associated with it (it just runs on a file and makes a bunch of files) it'd be fairly clean to add to the GUI.
2) Chop up a large object into smaller pieces, so that they can be printed on a smaller printer and glued together. PuzzleCut is an OpenSCAD script
http://www.thingiverse.com/thing:35834 that does this, with the nice capability that it cuts them with a "jigsaw" pattern, so the pieces slot back together, which is stronger than gluing together flat surfaces.
Both would be nice, but if I had to pick, personally I'd use (1) more often. The second case is more complex, and it might be better to leave in a separate tool like PuzzleCut, so that people can play around with settings, etc., without adding complexity to S3D's GUI.