[pemulis]

One of the biggest questions will be how transformative annotations are, and whether that will give adequate protection if users upload large chunks of copyrighted text. Google's use of thumbnails in image search wound up being covered by fair use[1] for this reason, and a judge recently ruled that book scanning was covered by fair use[2], as well. From the article:

> "The use to which the works in the HDL are put is transformative because the copies serve an entirely different purpose than the original works: the purpose is superior search capabilities rather than actual access to copyrighted material," wrote Judge Baer. "The search capabilities of the HDL have already given rise to new methods of academic inquiry such as text mining."

This might set precedent for a showdown over annotations that quote extensively from the source text, since online annotation systems also "[give] rise to new methods of academic inquiry."

The main point here is that laws about fair use are pretty vague, and new technologies always shake things up. We don't know what is legal and what isn't until it's put before a judge. In the case of Google's book scanning, the court case took seven years, which shows just how tricky this area of law can be.

[1] http://arstechnica.com/tech-policy/2007/05/google-v-perfect-...

[2] http://arstechnica.com/tech-policy/2012/10/court-rules-book-...

[001sky]

Over annotations that quote extensively from the source text

The page they have up with the Topic Blog post, however, is just a 100% reproduction with flyover notes. If You were to reprint the entire Harry Potter series, and add notes to the margins, it might seem similar. I believe that was already a subject of legal debate, though. Which might be precedent [see notes 1,2].

That being said, I don't per-se disagree with any of the issues being raised in your post. Also, I wouldn't rule out they will find a way to license content. The issue is just at what price? That's the spotify connundrum, IMHO.

_______

[1] http://en.wikipedia.org/wiki/Warner_Bros._and_J._K._Rowling_...

Judge Patterson said that reference materials were generally useful to the public but that in this case, Vander Ark went too far. He said that "while the Lexicon, in its current state, is not a fair use of the Harry Potter works, reference works that share the Lexicon's purpose of aiding readers of literature generally should be encouraged rather than stifled." He said he ruled in Ms. Rowling's favor because the "Lexicon appropriates too much of Rowling's creative work for its purposes as a reference guide."

[2] Another example, might be you-tube. Who has its "annotations" listed below each work of art. These are commonly referred to as "Comments". Some of the comments can now even be time-inserted into flyovers, more similar to the "flyover" type annotations currently up on the linked article. This is an example where there is both licensing and unlicensed content as a hybrid model.

[pemulis]

That's a great point about the Harry Potter Lexicon case. One thing that's interesting about that ruling is how narrow it was, and the implication that if the Lexicon contained more extensive commentary it would be legal.

Here's a question[1]. Let's say you took a work that was not in the public domain but has been the subject of extensive academic study, like A Perfect Day for Bananafish. Every line of that short story has been quoted in an academic paper at this point, and each quote was fair use. What would happen if you compiled all of those papers in a single website? That website would contain the full text of A Perfect Day for Bananafish, but in a fragmented form. Would it still be fair use?

Now let's say you pieced the quotes together like a jigsaw puzzle, and included the extensive commentary in flyover annotations. Is it still fair use? Probably not, would be my guess.

What if only half of the quotes were pieced together, and the rest was summarized? What if the annotations of the summaries included quotes that fill in the rest of the story? Where exactly is the line drawn?

I'm not arguing for the sake of arguing, I promise. I just find it interesting. It's like we're facing the metaphysics of copyright, sort of a Borges As IP Lawyer thing.

[1] Posed as a thought experiment, because I have no idea what the answer is.

[001sky]

That website would contain the full text of A Perfect Day for Bananafish, but in a fragmented form. Would it still be fair use?

This is a good question. The Academic purpose, however, would seem to be very useful as a criterion for Fair use (under point 1). The completeness of a compendium, however, may create its own issues (under point 4).[1]

The closest commercial example, might be sampling of music. Perhaps the early 80's, with the emergence of DJ culture etc. Small pieces of extant art, repurposed. This did not usually impact the original adversely. And the issue is not that no new value is being created -- it most certainly is -- its just that these things tend to end up in court. So, to avoid that, people settle or license.

I would be open to hearing other arguments or cases, but I think that we would need to start with something that has at its origin a commercial endeavor. Another example that comes to Mind is Shep Farley, the Obama Hope poster. This had a good amount of transformation on top of a sample of an image from a photograph. But ended up being settled out of court. See eg:

http://en.wikipedia.org/wiki/Shepard_Fairey#Legal%20issues%2...

_______

[1] Eg if it rendered away in full the need to buy at least 1 extant copy to study

[pemulis]

All good points. I want to make note of two other cases that seem relevant, both related to Youtube. In 2010, a judge in the US threw out Viacom's lawsuit against Youtube/Google, ruling that they were protected by the Digital Millenium Copyright Act when users posted copyrighted material[1]. In Lenz vs. Universal Music Corp.[2], a judge ruled that rightsholders have to consider if a piece of content is protected by fair use before filing DMCA takedown notices.

Let's say an annotation platform allowed users to post snippets of material as freely as Youtube, along with annotations and commentary. And let's use A Perfect Day for Bananafish again as the source of these snippets of text. Imagine the following sequence.

Alice posts the first paragraph of the story, along with extensive notes. Bob posts the second paragraph of the story, along with extensive notes. Carol adds an annotation to the first paragraph with a link to the second paragraph. Dave, Eve, and a few dozen other people jump in the mix, and pretty soon you have the entirety of A Perfect Day for Bananafish on the annotation site, and it's incredibly easy to read the paragraphs in full and in their original order. Maybe the annotation site even has an algorithm that detects related pieces of content, and creates the links automatically. (This is certainly the case for Youtube.)

Where should the takedown notice be directed? Each person in the chain only posted a small part of the work, and each post includes annotations and commentary that would seem to grant it fair use protections. Taken as a whole, though, the entire work has been recreated in full. Can the rightsholder demand that the website remove links users make between posts? If so, under what section of the DMCA? And what happens if the links are made algorithmically?

Does a rightsholder have to issue takedown notices for each piece of the chain individually? If the people who made the posts file appeals, does the rightsholder have to take each and every one of them to court in separate cases? And if they do, would they get demolished by counterclaims?

This isn't an idle question, when you think about the annotations for The Great Gatsby[3] that the founders of Rap Genius have been mentioning in their recent interviews. It seems like this sort of scenario is inevitable, and I don't know if copyright holders have any way of actually dealing with it.

[1] http://www.google.com/hostednews/afp/article/ALeqM5h_AfErLSM...

[2] http://en.wikipedia.org/wiki/Lenz_v._Universal_Music_Corp.

[3] http://rapgenius.com/search?q=the+great+gatsby

Edit: It's important to keep in mind that, in this scenario, all of the potentially infringing content is being posted by people with no commercial motive, which makes it different from the Shepard Fairey case. The annotation site itself is making money, but is protected by the DMCA under current understanding of the law.

Edit x2: I love how deeply annotated this discussion of annotations is. Quick, someone link to this thread from Rap Genius and close the loop!

[001sky]

[Just to play devil's advocate]

__________________

Each person in the chain only posted a small part of the work, and each post includes annotations and commentary that would seem to grant it fair use protections.

Is likely going to fail under (4). ie, it fails not at the annotation-of-a-single-sample-in-isolation stage, but in the pre-meditated-compendium-cum-reproduction for profit stage.

You Tube, for the most part is a different example. They have a reasonable expectation that their content is intentionally UGC (original) or short-form (sampled) or a montage (also sampled, with transformation). The are multiple cases of ex-ante reasonable fair-use. So they can claim safe harbour.

Right now, ex-ante, that would be a tough argument for Rap Genius (given Harry Potter example). Their protoype looks to a plain observer like re-purposing in full the material (Annotated Linked Article), as a base upon which to add XYZ. Just like the NYTimes has to license a photo to annotate a story (they cant claim the story is annotating the photo=fair use), or else Getty images would complain. etc.

[edit]

Also, I do think you need to start with the presumption of a for profit business. If Wikipedia (or its foundation, etc) did a non-commercial version, I think then the analysis would change, fairly. The combination of extant materials/prototypes/early versioning, and their VC backing, etc. seem to put them in a different game.