The 1.4.2 API description for Graphics2D has:
"Some Graphics2D objects can be used to capture rendering operations for storage into a graphics metafile for playback on a concrete device of unknown physical resolution at a later time. Since the resolution might not be known when the rendering operations are captured, the Graphics2D Transform is set up to transform user coordinates to a virtual device space that approximates the expected resolution of the target device. Further transformations might need to be applied at playback time if the estimate is incorrect."
I can find no further information about using this facility. I would like to be able to output to a virtual GraphicsDevice implementing PDF output.
Out of the blue, I was directed to the SVG Generator, which does for Batik what I have in mind for alt-design. Talk about serendipity! We have been discussing an integration of Batik and FOP under an XML Graphics umbrella, a discussion driven particularly by Jeremias.
It has seemed to me for a while now that using the 2D API as a basis for rendering offers the possibility of a clean, or at least well documented and understood, interface between layout and rendering. In that case of alt-design, it is also the basis for manipulating the layout.
I think the approach may offer benefits to HEAD. Providing a common interface between layout and rendering for PDF and PS would be beneficial for all, and would bring pluggable layout closer to realisation. The SVG renderer would virtually be in place already.
Obviously, I would love to be able to output alt-design's layout to PDF without having to build a new interface mechanism.
I believe that the same approach could be used with the structure renderers. With the current approach to RTF, it seems to me that a number of sacrifices have to be made. The FO input cannot be fully realised with a complete resolution of the properties, which in turn relies on layout. (Old argument, I know.)
Peter -- Peter B. West <http://www.powerup.com.au/~pbwest/resume.html>