On Sun, Aug 17, 2008 at 12:41 PM, darklama <darklama(a)gmail.com> wrote:
Having the FSF involved in writing a book might be a
good start to
clarifying requirements Wikibooks and other projects need to follow
to be compliant. However I am hoping for something more. Like
a took to automatically generate the necessary attributions for
inclusion in a book in a way that makes it clear who did what under
what licensing terms. This involves needing to answer the question
of what's not enough and what's too much information?
Ah, the ever wished-for technical solution. This is a problem not just
for Wikibooks but for all WMF projects. There just plain isn't an easy
automated way to extract information about the contribution history
into an easy-to-print format. GFDL does require we attribute our
authors, and at the moment we (the royal "we" for all WMF
contributors) are hoping that our difficult-to-parse history pages
satisfy this requirement.
Getting attribution information for authors, both authors of content
and images, included into our distributable versions (PDF and Print
versions) is a must, for the GFDL or any other copyleft license that
we use. This is a feature that we should demand be included in the PDF
extension that is being tested right now. I agree with darklama that
we can't really move forward on any issue until this most basic
requirement is satisfied from a technical perspective.
One possible solution to this could be to take the
same attitude towards
media as some would like to take towards dual licensing books. Require
that at a minimum all media must be licensed under the GFDL or allow
relicensing under the GFDL only.
Here I think is my biggest objection, and I think I've been unable to
describe it properly heretofore. Dual licenses are typically either-or
situations, options are presented and the reuser may select one, the
other, or both at their discretion. What we would have to do on
Wikibooks to allow dual-licensing of books is not only to specify what
licenses were the available options, but also we must mandate
additional terms: That the entire multi-licensing scheme must be
preserved on Wikibooks for compatibility, that reusers may select
either-or license (except reusers on Wikibooks itself).
Imagine that we have a book that is GFDL+CC-BY-SA-3.0. Content reusers
on a different website take our book, make modifications to it, and
release the book under the GFDL only. Now, those changes cannot be
folded back into our book because the contributions do not allow
CC-BY-SA-3.0. We then lose access to changes made to our content
downstream, the exact situation that copyleft licenses try to prevent.
In the case we have now our content is always GFDL, downstream
derivatives are always GFDL, users and reusers always have to use the
GFDL, everything works seemlessly.
Another solution is to just have a tool to generate
the attributions
for all media used in a book. Question though is what is needed to
attribute media used in a book? The media filenames aren't
necessarily going to be included with every use of the media within
the book to make it easy to associate license with media. Does this mean
in order to acknowledge the contributors of a media and the license used
that the media would need to be a literal part of the attribution? Does
every media that uses the same license still need a separate
attribution? Would separate attributions be needed if the book and all
media used in the book used the same license? Would this still be a
problem if the book and all the media in the book used the same license?
Great questions, all. I submit for the consideration of all discussion
participants now some of the work I've attempted to do on the
[[Control Systems]] book. Take a look at this page, for reference:
http://en.wikibooks.org/wiki/Control_Systems/Licensing
This is the most that I've seen in any book to try and satisfy the
differing license requirements of our books and our image media, and
it still isn't enough. For most licenses, we would need to include
attribution lists not just for the pages in the text, but also the
images. Some other licenses also require that we include the text of
those licenses with their respective images in the distributed copy of
the book.
What we really need is something to standardize and automate this
process. Some kind of MediaWiki-based technical solution would be
ideal for this, but some kind of extension or JavaScript or whatever
would be fine too.
--Andrew Whitworth