I just checked over the paper I wrote (1994)...
and the answers given there differ from one I
wrote off the top of my head and included in the
previous message (= 1/0 1/0) should be true in
the affine model I recommend. I think the project
model would be nil..
Anyone interested should look at
the paper. I also distilled it to pdf.
http://www.cs.berkeley.edu/~fateman/papers/extrat.pdf
RJF