A few additional corrections / clarifications:
1) Our partnership with PediaPress has not displaced comparable
community efforts, nor did PediaPress "offer money" and therefore
received attention that other efforts did not get. Most of the
community-based efforts at that time, including the ones Robert is
referring to, were of an entirely different nature: manually
collecting content from projects and creating reasonable-looking PDF
files, then selling them through a print-on-demand publisher like Lulu
(obviously a completely commercial/proprietary operation). There were
a few barely functional PDF exporters, but nothing coming close to the
It's true that a 2006 community-driven effort to publish Wikijunior
content had incorrectly identified "Wikimedia Foundation" as the
authors of the content. That mistake was corrected; as far as I can
tell, the same volume is still available at: http://www.lulu.com/product/paperback/wikijunior-big-cats/1875136
You can review some of the relevant threads here: http://lists.wikimedia.org/pipermail/foundation-l/2006-July/thread.html
Volker from PediaPress first introduced himself at the same time, but
nothing much happened between WMF and PediaPress until January 2007,
when we contacted them about working together, which ultimately
resulted in the current business relationship.
2) In terms of open source code, as I explained, PediaPress has
contributed a full alternative parser implementation, a complete PDF
export implementation, a complete tool for assembling and managing
article collections, etc. These are all very important and valuable
contributions. By last count, the PDF feature is used to create about
85,000 PDF files per day, keeping two dedicated servers busy. We'd be
happy to integrate open source LaTeX support if someone provided it,
and we'd be happy to consider paying for implementing it if enough
people found it useful.
3) We've in the past explored various other partnerships with
publishers resulting in commercial use of Wikipedia content. One of
the most elaborate such partnerships was the Bertelsmann "Wikipedia in
one volume", based on the German Wikipedia (using edited lead sections
as mini-articles). Trademark use for this book was negotiated by
Wikimedia Germany with approval by WMF. The book was a commercial
failure. See http://www.amazon.de/Das-WIKIPEDIA-Lexikon-einem-Band/dp/3577091029/
for information about this book.
In general, we've concluded that most such commercial partnerships are
a) Commercial publishers are not comfortable with freely licensed
content, and try to find ways to lock it in;
b) Most such partnerships would be poorly scalable one-offs;
c) Both the revenue potential and the mission benefit are relatively small.
We like the PediaPress model, because:
a) it's fully consistent with the intent of free content licensing;
b) it allows people to create their own customized experience in any
c) it can scale flexibly to accommodate demand.
That doesn't preclude other models from being potentially viable. Even
the PediaPress model allows for more carefully curated content (using
collections pointing to specific versions of pages that have been
reviewed for book export), and of course it would be great to see more
community efforts to vet, collect and publish content.
Such efforts don't require our permission where no trademark use is
involved. If trademark use is involved, then we'd have to consider
such requests on a case-by-case basis, but we'd certainly consider
them. (There's a big difference between claiming "authorship" of WMF,
or labeling a book "Wikibooks: Physics" -- the former is factually
incorrect and never acceptable, the latter is a potentially
permissible trademark use.) I'd argue that working with PediaPress on
this would be advisable: They have an existing 10% revenue sharing
agreement with WMF, and the existing toolchain allows for export to
multiple formats using entirely open source tools. But alternatives
are always worth looking into.
4) There's a big difference between something like
Special:Booksources, and something like the book creator tool. The
former links to separate and independent services (commercial or not),
the latter operates commercially on Wikimedia project content.
Services that integrate and use our content commercially should at
minimum be vetted by WMF, to establish fair and reasonable parameters
and to ensure compliance.
There's actually an example of a commercial printing operation that's
been entirely developed by individual community members: the
WikiPoster service running on the French Wikipedia. To see it in
operation, click on any image in the French Wikipedia and click
"Obtenir un poster de cette image": http://fr.wikipedia.org/wiki/Fichier:Pirates_of_the_Caribbean.jpg
I have very little information about this service -- neither whether
they are meeting their promises of donations (I could ask accounting
to examine our records, but we receive no explicit reports), nor
whether any of their software is open source, nor whether it's
functioning correctly and delivering a quality experience. Wikimedia
France may have more detailed numbers. In general, I don't think this
is a good way to develop commercial relationships, but since it's
supported by the French community in this case, and we have no plans
to provide the same service in some other fashion, we've not asked to
We'd be more than happy to consider alternative print-on-demand
services. So far, we've had two commercial contacts related to the
print-on-demand tool. One, an alternative printer servicing a specific
geography, was happy to work directly with PediaPress instead of
developing their own wiki-to-print infrastructure. The other was too
small to be relevant.
5) I've given you the scale of sales (about 8,000 per year at current
numbers). As "premium" as the placement in the sidebar may seem, the
reality is that the "Print/export" link is collapsed by default,
doesn't obviously relate in any way to ordering printed books, and
provides access to a number of related services and functions. You
have to expand the sidebar, activate the book creator, assemble your
own book and go through the ordering process to get a printed book.
The fact that this still leads to thousands of sales is testament both
to our huge readership and the actual interest in such functionality.
I understand the desire to apply fair and neutral principles to such
partnerships. Fairness and neutrality, to me, would indicate that
another for-profit publisher would either have to provide some serious
added value (in the form of product or code), or pay a higher
commission to WMF, to make up for the lack of any investment in the
underlying technology. But given the sales numbers and our reluctance
to advertise the tool in any prominent fashion, it's not surprising
that major corporations aren't exactly beating our doors down.
In general, commercial partnerships we've entered that relate to code
or services provided through Wikimedia projects are very careful and
conservative, and tend to focus on mission value more than revenue.
PediaPress falls into this category: the relatively tiny amount of
revenue we're getting from it is not the reason we're engaged in the
relationship. Rather, supporting community efforts to print content
had long been something we wanted to do, and this was a scalable way
to do it. And yes, having a third party responsible for printing and
sales is a good idea from a legal perspective.
External partnerships such as wikipedia.orange.fr , where Wikipedia
content is commercially used (using the Wikipedia trademark) with our
explicit permission, are significantly less conservative and more
revenue oriented. We've generally placed less focus on these types of
business relationships in recent months, as part of our general shift
towards the "many small gifts" revenue model as the central pillar of
our revenue strategy.
Deputy Director, Wikimedia Foundation
Support Free Knowledge: http://wikimediafoundation.org/wiki/Donate
foundation-l mailing list