The Copyrightability of Databases

Recently, a colleague asked me about the copyrightability of information in a database, specifically, in this instance, a phonebook. My colleague wanted to know if you could claim copyright over a database: this is a really important question, as the internet is, by and large, a gigantic database. This is actually two separate questions: 1. Are the contents of databases copyrightable? 2. Are there copyrightable elements of databases in and of themselves? So let's break it down and treat each separate.

Contents of Databases
We do know copyright does not apply to facts, and it would be an unfortunate circumstance if simply putting a fact in a database gave you copyright over that fact. The case that everyone learns in law school that expresses the copyrightable elements of compilations is Feist V. Rural Telephone Co. According to Wikipedia:
Feist Publications, Inc., v. Rural Telephone Service Co., 499 U.S. 340 (1991),[1] commonly called Feist v. Rural, is an important United States Supreme Court case establishing that information alone without a minimum of original creativity cannot be protected by copyright. In the case appealed, Feist had copied information from Rural'stelephone listings to include in its own, after Rural had refused to license the information. Rural sued for copyright infringement. The Court ruled that information contained in Rural's phone directory was not copyrightable and that therefore no infringement existed.
If Feist had literally distributed photocopies of Rural's telephone book, there would be a claim of copyright information. However, merely reproducing the information in a phone did not constitute copyright information. This assertion was largely supported by the decision in Bellsouth V. Donnelly, which basically stated that the selection, arrangement and coordination of a phonebook were the only copyrightable elements of the phonebook. Contrast this, however, to CCC Info Services v. Maclean Hunter Market Reports, where it was held that the valuations of automobiles in the "Redbook" constituted creative, judgment based content, and not simply merely factual reporting. Similarly, in CDN v. Kapes, the coin prices in a guidebook were found to be copyrightable because they were the result of expert judgment, and not just mathematically computed facts. So, when looking at the contents of databases, the question to ask regarding their copyrightability is, "Is this content merely a list of facts? Or is this content the result of a process of creativity or judgment?" If the answer to the former is yes, it is probably not protectable; if the answer to the latter is yes, the content is probably protected by copyright.

Copyright of Databases Themselves
We know that copyright explicitly covers compilations, i.e., that the creative elements involved in the organization of other works (layout, selection, etc.), are also protectable by copyright. A magazine, for instance, has many artistic and creative choices made in its organization and appearance. As does a compilation of selected works, or a series of short stories, photo essays, coffee table art books, etc.
     Importantly, however, databases are not like other creative compilations. A database, is a functional data structure: most of the 'layout' is done by algorithms, and much of the 'creative' choices made are made out of mathematical necessity. At first glance, then, it looks like we may be lead into the uncomfortable position where copyright can be extended over the form of databases themselves - which would be disastrous for the internet, because it means that you would be able to 'own' a data structure (i.e., listing names cross referenced by phone numbers or email addresses, etc.) simply by being the first to commit it to a tangible form.
     This turns out not to be the case, largely due to the idea/expression dichotomy, also called merger doctrine. This doctrine states that one cannot extend copyright protection over expressions that are 'merged' with their underlying ideas, concepts, natural laws, etc. To put it another way, you cannot use copyright to make a claim of ownership over functional elements or to get ownership of concepts. The courts have been recognizing increasingly that  databases in and of themselves are merged structures: they are mathematical, programmatic structures designed to allow for rapid manipulation of data in many forms, i.e., they are fundamentally merged. Or, as stated by the Copyright Office:
In summary, very few of the post-Feist compilation cases have held entire works to be uncopyrightable. In fact, copyrightability of the entire work is seldom even contested. Disputes tend to focus instead on the scope of protection. Consistent with Feist’s pronouncement that copyright affords compilations only “thin” protection, most of the post-Feist appellate cases have found wholesale takings from copyrightable compilations to be non-infringing. This trend is carrying through to district courts as well.(70)
      Translated into normal human, that means that the Courts have recognized that even copying big chunks of other people's databases, verbatim, rarely rises to the level of copyright infringement, so long as the information being copied itself is not subject to copyright protection. The arguments that get into court over databases mostly concern how much copying is allowed, or whether the underlying information itself is subject to copyright. As a result, as that (very lengthy) paper by the Copyright Office goes on to explain, database protection is mostly accomplished these days through a combination of contracts (ever wonder what all those "Click Here to Agree" terms of service agreements are for?) or with technological protection measures, i.e., DRM. So, if you are interested in protecting your databases, you should investigate your terms of service and your security measures.

In the end, the question about copyrightability of databases comes down to this: is the information in the database itself protectable by copyright? If the information in the database is the result of a sophisticated judgment or a creative process, then the answer is almost certainly yes. However, if the database consists of simple facts, or the output of deterministic formulae with no creative input, the answer is very likely no. Unlike artistic compilations, like magazines, photo collections, excerpted and collected works, etc., databases, though they may appear similar to compilations, do not typically receive the same protection of copyright law. As above, this due to the merger doctrine, because databases are functional constructs. So, unfortunately, there is no bright line rule dictating whether a database is per se copyrightable: it will always be a highly fact specific determination. The basic lesson, in the end, is largely consistent with copyright law in general: creative elements get copyright protection, whether those elements are visually creative layout of someone else's works or a series of numbers that resulted from a creative process.

When in doubt, it is probably best to ask your lawyer.


Popular posts from this blog

Contract Drafting: Software Development Agreements

Apple Might Be Forced To Reveal & Share iPhone Unlocking Code Widely

Why Stack Overflow Doesn’t Care About Ad Blockers – Stack Overflow Blog – A destination for all things related to development at Stack Overflow