| Author: | Tuomas J. Lukka |
|---|---|
| Last-Modified: | 2003-08-05 |
| Revision: | 1.1 |
| Status: | Current |
The transform_concat--tjl PEG left one type of concatenation out, which will make things slightly more efficient. Concatenating the inverse of one coordinate system to another can be done in a single operation instead of two.
This is a useful operation in, e.g., the View2D stuff in FenPDF.
The OpenGL libvob already implements invert as
public int invert(int f) {
return concatInverse(0, f);
}
This peg proposes to add concatInverse into the official VobScene definition.
Add into VobScene:
public int concatInverseCS(int into, Object key, int other) ;
and VobCoorder:
/** Create a new transformation that is the concatenation * of an existing transformation and the inverse of another * transformation. * If we look at the transformations as x' = f(x) and x' = g(x) then * the result of this operation is a transformation h, * for which h(x) = f(g^-1(x)) always. */ public int concatInverse(int f, int g) ;