17 December 2007

Open Access Data - A Question of Protocol

Something calling itself a “Protocol for Implementing Open Access Data” sounds about as exciting as a list of ingredients for paint. But this memo from the Science Commons is one of the most important documents in this field to date. Its scope is explained in the opening paragraph:

This memo provides information for the Internet community interested in distributing data or databases under an “open access” structure. There are several definitions of “open” and “open access” on the Internet, including the Open Knowledge Definition and the Budapest Declaration on Open Access; the protocol laid out herein is intended to conform to the Open Knowledge Definition and extend the ideas of the Budapest Declaration to data and databases.

Again, that may not sound very exciting, but trying to come up with definitions of “open data” or “open access data” have proved extraordinarily hard, and in the course of the memo we learn why:
3. Principles of open access data
Legal tools for an open access data sharing protocol must be developed with three key principles in mind:
3.1 The protocol must promote legal predictability and certainty.
3.2 The protocol must be easy to use and understand.
3.3 The protocol must impose the lowest possible transaction costs on users.

These principles are motivated by Science Commons’ experience in distributing a database licensing Frequently Asked Questions (FAQ) file. Scientists are uncomfortable applying the FAQ because they find it hard to apply the distinction between what is copyrightable and what is not copyrightable, among other elements. A lack of simplicity restricts usage and as such restricts the open access flow of data. Thus any usage system must both be legally accurate while simultaneously very simple for scientists, reducing or eliminating the need to make the distinction between copyrightable and non-copyrightable elements.

The terms also need to satisfy the norms and expectations of the disciplines providing the database. This makes a single license approach difficult – archaeology data norms for citation will differ from those in physics, and yet again from those in biology, and yet again from those in the cultural or educational spaces. But those norms must be attached in a form that imposes the lowest possible costs on users (now and in the future).

The solution is at once obvious and radical:

4. Implementing the Science Commons Database Protocol for open access data
4.1 Converge on the public domain by waiving all rights based on intellectual property

The conflict between simplicity and legal certainty can be best resolved by a twofold measure: 1) a reconstruction of the public domain and 2) the use of scientific norms to express the wishes of the data provider.

Reconstructing the public domain can be achieved through the use of a legal tool (waiving the relevant rights on data and asserting that the provider makes no claims on the data).

Requesting behavior, such as citation, through norms and terms of use rather than as a legal requirement based on copyright or contracts, allows for different scientific disciplines to develop different norms for citation. This allows for legal certainty without constraining one community to the norms of another.

Thus, to facilitate data integration and open access data sharing, any implementation of this protocol MUST waive all rights necessary for data extraction and re-use (including copyright, sui generis database rights, claims of unfair competition, implied contracts, and other legal rights), and MUST NOT apply any obligations on the user of the data or database such as “copyleft” or “share alike”, or even the legal requirement to provide attribution. Any implementation SHOULD define a non-legally binding set of citation norms in clear, lay-readable language.

The solution is obvious because the public domain is the zero state of copyright (in fact, the new Creative Commons public domain licence is called simply CCZero.) It is radical because previous attempts have tried to build on the evident success of the GNU GPL by taking a kind of copyleft approach: using copyright to limit copyright. But the new protocol explicitly negates the use of both GPL's copyleft and the Creative Commons Sharealike licences because, minimal as they are, they are still too restrictive – even though they are both predicated on maximising sharing.

One knock-on consequence of this is that attribution requirements are out. This is not just a matter of belief or principle, but of practicality:

In a world of database integration and federation, attribution can easily cascade into a burden for scientists if a category error is made. Would a scientist need to attribute 40,000 data depositors in the event of a query across 40,000 data sets? How does this relate to the evolved norms of citation within a discipline, and does the attribution requirement indeed conflict with accepted norms in some disciplines? Indeed, failing to give attribution to all 40,000 sources could be the basis for a copyright infringement suit at worst, and at best, imposes a significant transaction cost on the scientist using the data.

It is this pragmatism, rooted in how science actually works, that makes the current protocol particularly important: it might actually be useful. It's also significant that it plugs in to previously existing work in related fields. For example, as the accompanying blog post explains:

We are also pleased to announce that the Open Knowledge Foundation has certified the Protocol as conforming to the Open Knowledge Definition. We think it’s important to avoid legal fragmentation at the early stages, and that one way to avoid that fragmentation is to work with the existing thought leaders like the OKF.

Moreover, the protocol has already been applied in drawing up another important text, the Open Data Commons Public Domain Dedication & Licence:

The Open Data Commons Public Domain Dedication & Licence is a document intended to allow you to freely share, modify, and use this work for any purpose and without any restrictions. This licence is intended for use on databases or their contents (”data”), either together or individually.

Many databases are covered by copyright. Some jurisdictions, mainly in Europe, have specific special rights that cover databases called the “sui generis” database right. Both of these sets of rights, as well as other legal rights used to protect databases and data, can create uncertainty or practical difficulty for those wishing to share databases and their underlying data but retain a limited amount of rights under a “some rights reserved” approach to licensing. As a result, this waiver and licence tries to the fullest extent possible to eliminate or fully license any rights that cover this database and data.

Again, however dry and legalistic this stuff may seem it's not: we're talking about the rigorous foundations of new kinds of sharing - and we all know how important and powerful that can be.

Update: John Wilbanks has pointed me to his post about the winnowing process that led to this protocol - fascinating stuff.

No comments: