[cups.development] Pseudo duplex mode for several printers

Till Kamppeter till.kamppeter at gmail.com
Thu Oct 18 13:37:40 PDT 2007

Probably it is more reliable to let the rastertospl2 driver do this work 
(you probably thought in delegating it to pstops).

Page reordering, scaling, and so on is currently done by the pstops 
filter. Unfortunately PostScript is not designed well enough to force 
every Postscript file to have easily distinguishable pages. Later the 
DSC comments were defined to have the possibility to make a 
well-structured Postscript where one really can distinguish the pages, 
but many applications produce a PostScript which is not DSC-conforming.

To solve this problem in the future, PostScript will be replaced by PDF 
as standard print job format (decided on the Printing Summit in April 
2006 in Atlanta). In PDF files you can always separate the pages.

So for now it is better to go to a later step of the filter chain, where 
the data is already in the CUPS raster format or even in the printer's 
native format. Here it is very well defined where one page ends and the 
next one starts, only disadvantage is that the data chunks representing 
the pages are much bigger than in the PostScript stage, so more disk 
space for temporary files is needed for the reordering. This would mean 
that you reorder the pages in rastertospl2. Also rotating pages by 180 
degrees is probably more reliable with bitmaps. You could stream the 
incoming data into two temporary files, one with the even and one with 
the odd pages, after that you stream out one file and then the other.


Aurélien Croc wrote:
> Hi,
> I'm the main developer of the SpliX CUPS driver (driver for QPDL 
> printers (Samsung, Xerox, Dell, Ricoh) and i'm working on the "pseudo 
> duplex" mode which is available in several samsung printers.
> This pseudo duplex mode prints the even inversed pages, waits the user 
> get the printed sheets, puts them in the printer and press a button, 
> then prints the odd pages.
> Unfortunatly, this options needs that the document struct must be 
> "altered": the inverted even pages must be placed before the odd pages..
> Finally the driver needs to know where is the first odd page to change 
> several bytes sent to the printer (to specify to the printer that the 
> next pages are odd pages).
> So, i'm wondering how CUPS can help my driver by giving to it pages in 
> this special order (and rotation) ?
> Is it a way to do that? Is it "impossible"?
> Regards,
> Aurélien
> -- 
> M. Aurélien Croc
> Projet AP²C - http://www.ap2c.org
> Clef Publique : http://www.ap2c.org/aurelien.gpg

More information about the cups-devel mailing list