For those who haven't tried it out, here's what the PediaPress output
looks like (after it's done compiling the book, it'll give you a
That specific book is a good example of the problems that we've always
had with PDF generation by way of LaTeX, such as complex tables. Also
note the intermittent appearance of unsupported tags in the output.
As far as I know, the renderer they use is still partially
proprietary. I'm not sure if it would still be seen as valuable to
open source fully, given that LaTeX is indeed probably a technical
dead-end for these kinds of conversions, and given that the codebase
is very old.
If you're mainly using English Wikipedia, you might be under the
mistaken impression that the book creator is hidden from view. But it
is in fact still linked from the sidebar of many of the largest
Wikipedias, including French, Arabic, Spanish, Italian, Dutch,
Japanese, Polish, Portuguese, Russian, Swedish, Ukrainian, and
Vietnamese. A link on every page - that's quite a bit of exposure!
I agree it's a fair question what should happen to it: removal,
replacement, or repair. In general, I do think there's a strategic
case to be made for a more user-friendly way to create custom
collections and share/export them in multiple formats (and to point
people towards Kiwix and the ZIM format, which are indeed awesome for
educational and offline use cases), and it'd be great to see direct
collaborations with the OpenZIM/Kiwix community on this as Emmanuel