[Foundation-l] Re: Copyright issues...walking on thin ice

Wed Aug 11 07:33:14 UTC 2004

Michael Snow wrote in part:

>Jimmy (Jimbo) Wales wrote:

>>It is perfectly fine to aggregate independent
>>works under clause 7 of the GNU FDL.

>Clause 7 of the GFDL is predicated on aggregating the GFDL content with
>"separate and independent documents or works". Are images actually
>separate and independent?

Yes they are, as I will show.

>As files in the Image namespace, it's easy enough to say images are
>separate and independent works from articles. But once the image is used
>in an article, it gets fuzzier. When I view an article with an image,
>the image shows up there with the rest of the article, and looks to the
>reader like it is part of the article. It does not look very "separate
>and independent", nor does the image give any indication that it may be
>under a different license than the GFDL. You may point out that in wiki
>syntax, the image is actually linked to rather than being incorporated
>directly into the article, but that argument does not instill much
>confidence in me. The law has been known to disregard such attempts at
>technical distinctions and look at things from the simpler practical
>perspective.

The law shouldn't pay attention to wiki syntax' that's quite irrelevant.
You don't download wiki syntax (unless you click on "Edit this page");
rather, you download the entire web page that the server presents to you.

So am I agreeing with you?  No!  Because the Wikimedia server does *not*
present to you a web page that combines text material with an image!
Look at the HTML (*that* is relevant, because *that* is what you download);
you'll see that the HTML source (arguably a single document in the FDL)
contains the text, but not the image.  The image is only referenced,
with something like "<IMG src = "foo" alt = "bar" title = "baz" />".
The file <foo> contains the image, which is downloaded separately.

It's not Wikimedia's fault if your silly web browser confuses the issue,
making you think that the text and the image comprise a single document.
If you used a sensible web browser, like lynx ^_^, then it would be clear.

To be sure, since we have a reasonable expectation that most browsers
present content in a way that is liable to cause this very confusion,
we ought to take steps to avoid this sort of confusion.  And we do!
We say "All text is available under the terms of the GNU Free
Documentation License" -- note the phrase "all text", not "all content" --
and then we give a link to [[Wikipedia:Copyrights]], which explains things
in more detail (especially in section 1.2, at least on [[w:en:]]).
Even you, although not convinced about the aggregate nature of the page,
knew that something was up -- because you followed that link.

>But online content is not that big of an issue really, because anyone
>who says we're infringing on their copyright has to give us a takedown
>notice first, and we can remove the offending image. The real problem is
>print. And once you get to print, I have a _very_ hard time buying any
>argument that the image which illustrates an article is somehow a
>separate and independent work from the article text. The one kind of
>print version for which I might entertain this argument is if the images
>are segregated as is done in many books, on separate glossier facing
>pages or in a batch of illustrations in the middle of the book. But in
>the routine print version, where the image is printed out on the same
>page as the article, they look like part of one document and I don't see
>how you can make much of a case that they're not. As a result, I think
>the article as a whole, _including_ associated images, is the smallest
>Document to which we can legitimately atomize the GFDL.

I think that you have a good point about the paper version.
When a paper version is printed, we will have to take more care.
Even so, we'll probably want to treat images as separate documents;
but we'll need to take additional steps to make this clear to the reader.
That's because printed paper intermingles text with images
in a way that HTML (our only current publishing method) does not.

Offhand, I would suggest printing a notice of copyright information
along the side of each image that appears in the paper version,
similar to (but more detailed than) the photo credits in magazines.
The FDL notice for the article would again have to specify text only.
But we can (and should!) think about this matter more carefully
when we start producing printed versions.

-- Toby