Duplex printing chaos

[Helge Blischke]

A few hours of testing with various printers from different vendors
showed up at least three different variants of how duplex and tumble
are interpreted:

    duplex tumble    media is turned around an axis
    ----------------------------------------------------------------
(1) true   false     parallel to user coordinate system's Y axis
    true   true      parallel to user coordinate system's X axis

(2) true   false     the long edge of the physical media
    true   true      the short edge of the physical media

(3) true   false     parallel to the leading edge of the media
    true   true      perpendicular to the leading edge of the media

That means,
(1) above conforms to the PostScript specification
(PLRM 3rd ed., page416);
(2) above conforms to what RFC 2911 tells about the sides attribute,
(3) is sort of weird to me (though I think I understand how it is
technically achieved).

Are there any plans to force a device independent handling of the
sides attribute, e.g. by specifying which of the above interpretations
holds in the printer's PPD and tweaking either the attributes sent
to the printer or the feature statements put into the PS setup?

Helge