[PPL-devel] Re: Question on shrink_bounding_box

[Roberto Bagnara]

Pedro Vasconcelos wrote:
> I'd to like to ask what is the exact operation done by the
> shrink_bounding_box() method. My interpretation of the documentation is
> that it would 1) intersect the provided bounding box with the polyhedron
> and 2) compute the bounding box of the result.
> 
> I've done some tests (with PPL 0.6.1) and this does not appear to be the
> functionality provided. If initially
> 
> P   = {x0==x1} 
> BB = [0;1]*[-inf;inf]
> 
> then shrink_bounding_box() leaves BB unchanged (rather [0;1]*[0;1] as I expected).
> What is the intended operation? 

Dear Pedro,

what the shrink_bounding_box() method does is the following:

1) the smallest bounding box containing the polyhedron is computed; then
2) the bounding box provided by the caller is (destructively) intersected
    with the one computed in step 1 by using only the methods of `Box'
    listed in the documentation.

Notice that there is no way shrink_bounding_box() can obtain information
from the `box' parameter.  This interface and semantics allow for
maximum efficiency, yet minimizing the requirements we impose on the
`box' argument.  Any suggestion about how to improve the documentation
is welcome.
Cheers,

     Roberto

-- 
Prof. Roberto Bagnara
Computer Science Group
Department of Mathematics, University of Parma, Italy
http://www.cs.unipr.it/~bagnara/
mailto:bagnara at cs.unipr.it