Copyright Registration and the Compendium
Robert Kasunic, Associate Register of Copyrights and Director of Registration Policy & Practice, U.S. Copyright Office
25th anniversary of Feist. Cornerstone of his approach to copyrightability. Feist isn’t a telephone book case; it’s the only post-1976 case focusing on originality and what constitutes a creative act of authorship. It also explicitly stated that originality is a constitutional requirement. Neither Congress nor courts can protect a work without sufficient originality, and Office can’t register. Court stated that the amount of creativity required is low; most works make the leap easily, but some works won’t. Alphabetizing is a type of creativity that won’t qualify. Also, time/effort/expense aren’t relevant to copyrightability. Distinction b/t creative authorship and discovery of facts.
Feist didn’t answer all questions—even questions about yellow pages: selection of categories enough? What if they’re garden-variety categories? If there are creative categories, registration ought to be allowed, but how creative can telephone directory categories be? The more creative they are, the less useful it will be. We received a claim in a medical x-ray as a photo: but how many patients desire creative x-rays? Impetus for creation is discovery or functionality; that doesn’t mean that some x-ray could be created creatively, but mostly they’re done to discover facts, just like the census taker in Feist.
Yellow pages are typically alphabetical, but raise issues about collective works as opposed to compilations of facts. An ad in a directory could be sufficiently creative in image or words to be copyrightable. But unless selection, coordination, and arrangements is sufficiently creative, they’d need to be registered separately rather than together as a collective work.
Feist’s principles have to be translated to all categories of works: a circular arrangement of diamonds on a ring is equivalent to alphabetizing in its typicality. Examiners decide more claims every day than federal courts adjudicate in a year—80 examiners and ½ million claims per year. We see difficult/novel claims on a regular basis—individual performance of actor in film; DNA sequences depicted in photos; social dances; end-zone routines; online criticisms of doctor claimed by the doctor; tweets; compilations of tweets; all the recorded notes of an electric piano; jingles; interior designs for apartments; floral arrangements; edible arrangements; cheerleader uniforms. Creators often don’t understand copyright law and often aren’t interested in learning the nuances.
Common area of confusion: work made for hire. Many people would love to be hired, but don’t understand this is a term of art.
Examiners need to explain their refusals—Atari court was very concerned with this. Allowances don’t have to be explained, but we now annotate certificates to show where we thought there was an important ambiguity because registrations are presumptive evidence of validity. 95% are registered easily. But some works in every category of authorship may not make the grade. This result is “neither unfair nor unfortunate.” It’s the means by which © advances the progress of science & useful arts. Examiners must remain diligent; it would be easy to be a rubber stamp and let the courts weed out 5%.
Subject matter issues: dances; movements (yoga poses); performance of actor copyrightable as dramatic work? Do the statutory categories serve as a limit on copyrightability? Feist: Court demonstrates careful reading of all the words in statute, to make clauses consistent w/other clauses. Significant example: June 22, 2012 Fed. Reg. statement of policy on registration of compilations. Step towards building on Feist beyond originality to the limits on the categories of copyrightable subject matter. Being a compilation isn’t enough: has to be a compilation of stuff within the subject matter of copyright. But is §102(a) limitative? Many have interpreted the word “include” in §102 to indicate that the list is just illustrative, but no court has sanctioned a new category of copyrightable subject matter. Intellectually honest assessment was required. Other empirical evidence: Congress clearly chose not to include some works, such as typeface or works of industrial design. Choreography was deliberately chosen; architectural works were added explicitly; if the categories were merely illustrative, then addition should not have been necessary. So there’s clearly ambiguity about Congress’ intent, given its own actions.
Legislative history: H. and Sen. Reports are identical on this issue: by using original works of authorship, rather than all writings of an author, Congress intended to avoid exhausting its constitutional power in the field. The use of “includes” was designed to preserve Congress’ power in this field. “There are unquestionably other areas of existing subject matter that this Congress doesn’t choose to protect.” §102(a) was worded to avoid courts’ struggling with the scope of particular categories, but not to authorize them to create new categories of authorship.
Helped resolve recurring issues about compilations of noncovered subject matter, such as yoga poses. Also shows that registering a catalog of garden sculptures registers the photos in the catalog, but not the sculptures themselves, which have separate creativity.
Compendium of practices: we strove to clarify many issues, such as that the creativity in a work must be perceptible in the deposit, rather than expressed in a creative explanation of the work’s creativity. Puts Office at odds with CDN v. Kapes and CCC (Red Book) case—courts stated that individual numbers that were the product of various creative processes were copyrightable. It’s not a fact, but a creative prediction or estimation. A number as a deposit reveals no selection, coordination, or arrangement in itself.
Some courts have ignored the Office’s role as filter and the statute by allowing suit upon mere application, not upon acceptance or rejection of the application. Short-sighted. Nimmer sees this as no-harm, no-foul because most applications are accepted. But Feist counsels otherwise, as the Compendium explains. Expediency is not an excuse for ignoring the plain language of the statute and the unambiguous legislative history. Any applicant in prospective/pending litigation can pay for expedited service in 5 days. We mediate a much higher percentage of applications than 5%. In 40% we either correspond for clarification of authorship, ownership, WFH, joint work, removal of ambiguous terms, resubmission on proper form, submission of deposit, payment of proper fee, limitation based on previously registered/published/public domain material, and exclusion of authorship not fully owned by the claimant. Thus even claims in which registrations ultimately issue are regularly changed by the examination. We increasingly specify the basis on which something was registered—serves as a check on overclaiming and abuse, as w/doctors seeking to register criticism of their professional services. The doctors claimed ownership via transfer agreement, appended to federally required HIPAA notices. Rejected such claims: apparent involuntary transfer—doesn’t need to be invalidated by a court; we can do it.
Courts that accept pre-registration claims may find infringement where the registration wouldn’t have been accepted, or would have been limited. Many claims are just abandoned when the Office asks questions about them. Genetically modified poinsettia: courts can benefit from Office’s analysis of legislative history and statute; so too with social dance and yoga. Also the monkey selfie.
Software: Copyrightability and Registration
Moderator: Michael Carroll, Professor of Law, American University Washington College of Law
Terry Illardi, Copyright Counsel, IBM: Older companies tend to register more heavily. IBM is 102 years old and has 40,000 registrations. Mixture of documents and software; register 500 software programs/year and 1000 documents/year. IT companies are more mixed than entertainment companies; one large firm hasn’t registered one work in years; Microsoft has 3000 registrations. Special practices for software source code; object code doesn’t mean anything to a human being. Most source code has confidential/trade secret info, so Office allows special registration procedures. Now we do “striping”; the idea is that you deposit a copy with parts taken out. Whether it’s effective as registration is a good question, but we’re happy to be allowed to do it by the Office. Now we get digital copies of source code and a program obscures pieces of the code for us. Imagine a page of code with diagonal stripes going across the page; you can see 63%.
Software is complicated by different versions: IBM lingo is the version, the release, and the mod. We don’t file for registration on every mod (third position in version 1.0.1), but do for most releases and all versions. When 5-10% of code has changed, it’s time for a new version. Other companies don’t. You could do it for availability of attorneys’ fees, etc., but IBM doesn’t bring a lot of lawsuits, so that’s not a compelling reason. Valuable to our litigation dep’t for defensive purposes—can find information that generally helps our case; we get sued more than we sue. Registration also provides proof of ownership, and where this is important is with certain foreign gov’ts that require proof that software you’re licensing is yours. The easiest way to show that is a bunch of registration certificates. Makes deals go through more quickly—pays for the practice. Similar experience w/ foreign tax authorities and transfer taxes. Also useful in divestitures, which happen many times a year; provides proof that what we’re divesting we actually own.
Chris Mohr, Vice President for Intellectual Property and General Counsel, SIIA: Obscuring source code is real priority for members. Would also really like to see modernization of the Office. Increased security for deposits in an upgraded system (we wish).
Carroll: TRIPS requires protection for computer programs as literary works under Berne. But what about the Internet of Things? Software is embedded in a thing. How should that work? Does it make sense to ask for first/last 25 pages where programming art has changed so much? What should 21st century deposit look like?
Illardi: We are in the age in which software is put together like Lego blocks: old pieces snapped into new configurations. Companies have found efficiencies in reusable components/libraries. Open source is also a source, following the rules of the licenses. There usually is a beginning of a program, nonetheless, so there’s still some vitality to the old rule; not clear how useful the last 25 pages is b/c that can be almost anything. But to the extent that you can compare to the deposit copy to prove infringement, even if the code shows up in a different position, that’s useful; not clear to him there’s a better alternative. Programs have 10s of thousands of level blocks—some a dozen lines, some 100,000 lines. Shouldn’t have to deposit/pay for each one. Maybe a rule about the biggest blocks of code?
Carroll: what about encrypting them in a hash?
Illardi: interesting, though might not allow comparison with copied works which could be copies w/o having the same hash?
Q: isn’t the purpose of © deposit also to disclose? If it’s obscured, how will it ever enter the public domain?
Illardi: only keep copies for a limited time anyway. So the Office’s procedures won’t help historians anyway.
Kasunic: for published works, we keep for 30 years; it had been much shorter.
Q: the deposit copy may not help with actual litigation, where you have to show they copied what you registered.
Carroll: software is iterative; the Office is clear that a registration for 1.0 will not cover a prior version unless it’s unpublished. If prior version is published, you have to be very clear about what you’re claiming as the registrable material. So how does the IBM process work?
Illardi: careful records kept internally of what the changes are. Document filled out by developers and reviewed by lawyers.
IBM’s focus has historically been on freedom of action; relationship with Dep’t of Justice meant that it had to give others access to its patents, and wanted access to others’ in return. Our focus today may be relatively more on enforcement, but only relatively.
Carroll: Looming question: will we have to reconfigure the practices and understanding for internet of things? Current definition of computer used by Act (which then tells us what software is) requires an output on a screen/display or printer, but not all IoT objects will do that.
Mohr: the Office had an inquiry on embedded software, and registration didn’t really come up at all in the comments he looked at. The answer may be that computer programs can be embedded in machines, but it’s a question that needs an answer more detailed than we can give right now.
Illardi: may need to rethink definition of computer; smartphone is a computer that matches the current definition, but many things won’t b/c they’re meant to work with something else. Fitbit is meant to have an output, but the output shows up on another machine.
Copyright at the Edges
Moderator: Peter Jaszi, Professor of Law, American University Washington College of Law
Rebecca Tushnet, Professor of Law, Georgetown Law Center
Copyright puzzles: A “useful article” is an object having an intrinsic utilitarian function that is not merely to portray the appearance of the article or to convey information. Is a jigsaw puzzle a useful article? Copyright Office guidance: “No matter how novel, distinctive, or aesthetically pleasing any clasps, motors, or other functioning parts of an article may be, copyright does not protect them…. Articles whose intrinsic purpose is to portray a useful article or convey information—such as toys, scientific or fine-art models, newspapers, architectural plans, and maps—are not useful articles and may qualify for copyright protection. But protection would not extend to any mechanical or utilitarian parts they may have.”
So, is any connector between two things functional? If that’s the case, then interlocking jigsaw puzzle pieces seem functional: they make the picture hold together, much as the spine of a book holds the pages together. Other interesting questions: Computer-generated artworks (if artworks they are) where the piece shapes are determined by mathematical algorithms, and then perhaps hand-tweaked. Copyrightable? Here’s a real case where the output is independently valuable, compared to the program itself; the program’s obvious copyrightability doesn’t matter very much if a copier can use the output freely because there’s no human authorship in the output.
|Second puzzle made out of blue pieces of first puzzle|
Separately: If the pleasure added by the jigsaw puzzle to the image is haptic, then isn’t it more like a chair with a picture printed on it (or like a skateboard with an image printed on it, for that matter, another non-necessary item that is nonetheless a “useful article”) than like a sculpture? Things US copyright generally doesn’t cover: Perfumes; soft woven scarves with specially pleasing textures; recipes. Cf. Christopher Buccafusco, Making Sense of Intellectual Property Law, 97 Cornell Law Review 501 (2012); he has a followup article here which I will shortly read.
|An unassembled Liberty puzzle|
|Liberty puzzle, assembled|
Stave puzzles—the image intertwined with the cut. Is it a sculpture? If the cut isn't part of the work, then someone else might be able to copy the cut even though they can't copy the image.
Color line cutting: integrated with the picture in another way, in order to trick the eye or delight the puzzler when the piece falls into place. Do these techniques mean that puzzle cuts might be derivative works of the image, with creativity that intertwines with the expression of the image?
Comments indicated that some people would just try to register the cut as a 2D image, though I think that wouldn’t give you rights to control the cut if a puzzle is a useful article. The puzzle does what it does because the design is cut into different pieces. Back to the question: Does giving intellectual/haptic pleasure do something more than conveying information? I wouldn’t say petting my cat conveys information, but it certainly provides tactile pleasure.
Tom Kjellberg, Counsel, Cowan Liebowitz & Latman
Separability seems to set a higher bar for copyrightability than something that’s treated as purely aesthetic: pure abstract art in the ordinary 2D and 3D senses is usually protectable; would have higher bar as useful articles. Jewelry easy; button harder.
Newspaper is useful, but not a useful article because its intrinsic function is to convey information. 3 separability cases sought cert.; the cheerleading uniform case is scheduled for conference this month. Inhale v. Starbuzz: 9th Circuit gave Skidmore deference to Office’s holding that distinctiveness of shape doesn’t affect separability; any part of a container that merely accomplishes containment isn’t copyrightable/separable. [Which seems to me to have implications for my jigsaws …]
Legislative history: a 2D graphic work is still capable of being identified as such when it is printed on or applied to utilitarian articles such as textile fabrics, wallpaper, containers, and the like. [This is why I think the cheerleading case should have been decided on originality grounds, not separability grounds; the basic stripes just aren’t creative enough to pass through the copyright turnstile, but some other stripe arrangement might well be.]
Mannington Mills: heavily photoshopped photo of wood flooring used in laminated wood flooring. Designers chose particular planks, colored them, aged them, arranged them into aesthetically pleasing patterns; manipulated hi-res photos to create the pattern. Registered as 2D artwork. Court of appeals found it both physically and conceptually separable. The design doesn’t perform the function of hiding wear; the layer on which it’s printed performs the function of hiding wear.
Varsity Brands v. Star Athletica: involved in this case from the beginning, including having registered some of these designs. Created by in-house designers. How to characterize what it is we’re claiming? Don’t want protection for the drawing of the uniform; want protection for the uniform. Settled on “design on garment.” District court said this was a utilitarian object: nothing more than a cheerleading uniform.
But what is a useful article? The function of the uniform design is to convey the information that the person wearing it is a cheerleader, which should take it out of the category of useful article. (The function of the uniform is to cover the body, so in the first instance a shirt/skirt is a useful article.)
Q: Male cheerleaders just wear a T-shirt—so you can convey “I’m a cheerleader” in other ways. Should the design on a T-shirt have a separate copyright?
A: no, the © is in the 2D graphic work that appears on the surface. Could be a maple laminated floor pattern!
There is no circuit split, just different analytical tests on separability; just as there isn’t a split on substantial similarity just because the 9th and 2d Circuits express the test differently. [I note that the SCt suggested in B&B that there might be a similar problem with the trademark infringement multifactor test.] The 2d Circuit alone has taken multiple approaches; it’s subjective/you know it when you see it, and maybe should be revisited once more important issues in © have been resolved.
Carroll: isn’t there a split with the Harrod’s uniform case, or do the epaulets in the casino uniform serve a nonseparable function?
A: Star Athletica argues that there is a split with this 5th Circuit decision finding that the designs of casino uniforms lacked separability, and also pointed to an unpublished 2d Circuit holding on prom dresses, but in both cases they were talking about the 3D design of a garment, not a 2D design on a garment. In the Compendium’s next revision, some attention should be paid to that.
Q: compare your approach to the jigsaw puzzle. The puzzle has to have a thickness, which the design on fabric doesn’t.
Q: TM can protect the uniforms without ©. Smack Apparel etc. Recent Santa Claus w/snow globe in its belly: added enough artistic elements to make 3D thing; cut off the rights of the original artist, according to the court, b/c the original artist only had rights in a 2D drawing.
Carroll: Doesn’t think the 3D/2D distinction works; it’s really about the utilitarian function of the uniforms. The court describes cut of a prom dress, but also lace—the utilitarian function is to mark the wearer as going to a special occasion, not to cover the body; same with the design features in the Harrod’s case and the cheerleading uniforms—social utility. If that’s what makes it useful, 2D and 3D isn’t important.
A: definitionally, it’s some blend of portraying its own appearance and conveying information. So that’s true, but that doesn’t make it a useful article. [What if a sports car conveys the information “I am insecure in my masculinity,” as well as providing transportation? We know that the sports car is a useful article but so is a cheerleading uniform, because it provides transportation/covers the body respectively. What we’re asking is what’s separable from that useful article.] 2d Cir. distinguishes between dress designs as a term of art and fabric designs, which doesn’t just include a design printed on a bolt of fabric but also little squirrel appliques on the sweaters. So 2D v. 3D does some work.