For a while already, we wanted to experiment with the Scribus Scripter API in a printed job and a catalogue for the Piet Zwart Institute’s Media design graduation show formed the perfect excuse.

Mute‘s Simon Worthington connected us with Identic, a digital printer based in Brussels. They can print as many different files as you want without additional costs, so we asked Michael to help us with a Python script (modifications by Nicolas, Ivan and Femke) which runs from inside Scribus, iterating through folders and files generating a different book cover each time.

Each of the participants in the exhibition has provided us with 500 elements ranging from a series of unique IP addresses, a few lines of Python that alternate letter sizes, a game icon split into 500 giant three pixel images, 500 lines of a Wikipedia entry on Berthold Brecht, stills from a performance video and titles generated in a different font each time. According to the design rules contained in the script, Scribus places, colors and sizes each element before automatically exporting the resulting cover as a PDF.

OK, we did not manage to auto-place svg documents, which we really really really regret… and our first tests produced 500 super heavy pdf’s that managed to crash the Adobe preflight checker at the printer.

After a bit of testing and trying with the nice people at Identic, we found that we were embedding a corrupt font in the autogenerated pdf and this caused the crazy file size. When exporting the PDF manually from Scribus, Scribus detected the problem and outlined the font automatically but the scripter-API obviously does not offer such sophisticated features. So, we simply deleted the troublemaker and the files Scribus produced after that, are much lighter and most of all: they print!

With the help of Pierre Marchand and other helpful people on the Scribus IRC channel, we surpressed our panic about the initial slowdown of the process, and after 24 hours of continuous tireless work (1440 minutes / 500 pdf’s = 2.88 minutes per pdf. Not so bad, as Craig Bradney pointed out on the mailinglist), Scribus produced 500 different pdf’s, ready to print.

We can’t wait to see the final result!

Scripts and material for testing: http://ospublish.constantvzw.org/sources/pzicatalogue

Read the discussion on the Scribus mailing list:
http://lists.scribus.info/pipermail/scribus/2008-June/029400.html