someone already answered my question about polygonal model driving a subdiv model via wrap. It should work OK, just like that, but no guarantee. Thanks for that.
A further question though: if my character consists of separate pieces of geometry (face, clothes, etc), what then? Should I combine them first, without joining the vertices, and should I do the same with the subdiv model, or what would be the method? I'm also planning to use blend shapes for facial animation so I guess combining the pieces is not the way.
Thanks in advance!