Originally posted by Pentaxor that's exactly what and why R&D is there for. to make things work where nobody else think that certain things can be possible. they had to start from somewhere or something that is perceived impossible.
EDIT: Before I blabber on, here is Sigma's take on it:
http://forums.dpreview.com/forums/readflat.asp?forum=1036&message=34594332&q=OS+SR&qf=m (well, by way of a forum post)...
While R&D is indeed there to make things that are impossible possible, the work would have to be done on the Pentax SR, because it would need to respond properly to the OS of the Sigma lens, by doing the fine compensation the OS system still cannot do.
This is certainly an interesting proposal, in that Pentax cameras could realize that there is an OS lens attached, and have a pre-programmed pattern on how the SR should react to each motion, knowing that the OS will cover most of the movement.
It is possible that Pentax has done this. I highly doubt it, but it is possible.
The other way around, as I thought was suggested, would not be feasible. In such a coupled system, the lens stabilization would always have to do the brute adjustment, with the SR doing the fine adjustment.
Well, let's wait and see until someone does a test with it in real life. Since 24-70 OS and 18-250 OS lenses are already out, someone probably tested it already.