License or public domain for public sector information?

Mike Masnick at Techdirt asks Does It Make Sense For Governments To Make Their Content Creative Commons… Or Fully Public Domain?

Ideally all Public Sector Information (PSI; government content and data) would be in the public domain — not restricted by copyright or any related rights. Masnick points to the U.S. federal government’s good policy:

nearly all works produced by the [U.S.] federal government automatically go into the public domain, and don’t receive any form of copyright

Unfortunately it is not quite that good: works produced for the U.S. federal government, but not directly by federal government employees or officers are covered by copyright — including works acquired, produced by contractors, and funded by grants. Furthermore, works produced by U.S. federal government employees are only unambiguously free of copyright in the U.S., thus cannot be considered in the public domain worldwide. This is not to say that the U.S. federal government policy is not stellar — relative to policies of other levels of government within the U.S., and those of other governments worldwide, it truly is, to the particular and tremendous benefit of the U.S. people and economy. But we live in a globalized and highly interconnected world now, and even that stellar policy could be improved.

This brings us to another question: how to improve policy around PSI? The status of U.S. federal government works is specified in the U.S. Copyright Act. Crown Copyright is specified in the copyright acts of various commonwealth jurisdictions. Similarly many other jurisdictions’ copyright acts specify the status of and any special limitations and exceptions to copyright for government works. Clearly changing a jurisdiction’s copyright act or otherwise changing its default status for PSI (preferably to public domain) would be most powerful. But they aren’t changes anyone can effect relatively quickly and deterministically (historically opening up a copyright act has led to more restrictive copyright).

In the meantime (presumably many years) there’s a tremendous desire to make government more accessible and unlock the value of content and data that is funded, held, and produced by governments — and existing public sector copyright defaults are recognized as a barrier to achieving these benefits. Especially in the last few years, governments have been implementing their own directives aimed to modernize PSI while some government agencies and politicians look to move more quickly within their remits, and activist citizens push to clear barriers to the potential of “open government” or “government 2.0” with utmost urgency. This is where government use of a standard public license, usually one of the Creative Commons licenses, makes lots of sense. An agency, province, city or other body that holds copyright or funds the creation of copyrighted works can choose to open its or funded content by releasing under one of the Creative Commons licenses, or if they are really progressive, under the CC0 Public Domain Dedication.

Many governments are using CC tools in just these ways, and we expect that many more will in the coming years. That said, if any do manage to change policy defaults for PSI such that more government content and data is automatically in the public domain — we will be cheering all the way. In fact, we already have a tool for marking and tagging works that are in the public domain worldwide. The CC Public Domain Mark is currently applicable to really old works, but it would be lovely if a government were to decide to by law make all of its content unambiguously public domain, worldwide, thus making the CC Public Domain Mark applicable (of course there is no requirement to use the mark; it is just there for people and institutions that wish to use it to signal to humans and machines the public domain status of a work).

A couple caveats. First, whether they ought to or not, many governments like using copyright to control PSI. Sometimes the desire comes from a good place, e.g, to have the information be used in a way so as to not mislead the public, imply endorsement of the government, or imply that other regulations, e.g., privacy, do not apply. CC licenses have mechanisms to address these concerns where relevant (e.g., attribution to original URL, noting adaptation, non-endorsement) and government licensing frameworks (or non-binding guidelines in the case of the public domain) that explain orthogonal rights and responsibilities (e.g., privacy) but do not create incompatible licenses are key to addressing these concerns.

Second, although as noted above, usually use of any CC license would give the public more rights to PSI than they have now. But, licenses with a NonCommercial or NoDerivatives restriction set the bar too low. Clearly to maximize the value of public sector information, business needs to have access, and to maximize the ability of citizens to do interesting things with content, adaptation needs to be permitted. We strongly prefer governments use fully free/open CC tools — the CC0 Public Domain Dedication and CC Attribution (BY) and Attribution-ShareAlike (BY-SA) licenses. The Definition of Free Cultural Works and Open Knowledge Definition spell out why those tools are preferred in general. We look forward to working with the Open Knowledge Foundation and others to flesh out the specific and even more compelling case for fully free/open PSI.

Further reading:

7 thoughts on “License or public domain for public sector information?”

  1. Thanks for this thought provoking article Mike. As you point out, changing a country’s Copyright Act is not the most straight-forward of tasks. Where that’s not possible or practicable, Creative Commons is an obvious choice in many Commonwealth and other jurisdictions, as Australia and New Zealand have so clearly shown. In my view, where dropping everything into the public domain is not a runner, CC-BY should be the default unless a restriction of some sort applies. Imposing a share-alike obligation on citizens is not really necessary, in my view, as doing so can have a stifling effect on certain types of re-use. For example, let’s say a business wishes to add value to licensed copyright government data in some creative way. Having to share the adapted fruits of its labours on like terms to the original data released could cut against its business model. I don’t really see any need for that when the original is still available. I think open government data policies should be focused on allowing people to re-use government data for their own purposes, with as few restrictions as possible.

    (The above views are my own.)

  2. Hi Richard,

    New Zealand is an especially great example. For those reading along, check out and much more via

    I think the rationale you articulated against adding copyleft/ShareAlike to PSI licensing is fairly well accepted among people working on opening PSI, though there are a couple clear uses:

    * When government is adding to an existing copyleft project, as is often the case with software, occasionally with content and databases (of course government contributions could be under more permissive terms, but the whole project would be copyleft).

    * When there is a strong desire for more restrictive license terms (ie NonCommercial, NoDerivatives, or something ad hoc), copyleft/ShareAlike can be a useful fallback position, as it is far less problematic than other restrictions.

    There also is at least in theory a public policy rationale for preferring copyleft — one could posit that net of less direct reuse and greater secondary sharing is positive, especially in the long term. I don’t know of data to strongly support either view, though perhaps as governments adopt various open licenses there will turn out to be a good natural experiment. Whatever the differences turn out to be, they’ll probably be much less stark than say the difference between the U.S. (public domain) and European (restricted) public data policy, as powerfully illustrated by the value obtained from weather data — which is to say, good luck and thanks to all working to move PSI to fully open terms, whatever slight differences in strategy people might have.

  3. I much prefer at least some CC labelling than a default non-labelled “everything Gov’t produces is in the public domain unless we say otherwise” scenario. Jo Public struggles to understand copyright offline let alone online where everything just seems to be “available”. A CC label educates people who know nothing about IP both as to the terms on which that particular material may be used and more widely as to how copyright generally works.

  4. Would love to see some CC work to bring clarity to the software/code realm, in addition to content that is currently covered. I’ve done work for state agencies that would like to make their code available for free non-commercial use, but almost all licensing out there that deals with code focuses more on *redistribution* than *usage*. The conventional wisdom is that ‘CC is not for software’. Great – but can that be changed?

  5. Rick,

    Defaults when they can be set are extremely powerful, but labeling is certainly valuable even when the default is zero restrictions. That’s the use case for the CC Public Domain Mark.


    No. Free/Libre/Open Source software licenses do permit redistribution, but also use to the extent that is needed, eg from “This License explicitly affirms your unlimited permission to run the unmodified Program” and search for “run” throughout the license. Non-commercial use licenses have been thoroughly rejected for software. Any state agency that wants to make their code available should do so under an established GPL-compatible FLOSS license.

    Some further reading:
    * (some reasons why non-commercial only permissions are (even) less useful for software than content)

  6. What governments produce with public money (tax dollars) should be in the public domain so perhaps there should be a CC-PD designation to make this abundantly clear. Of even greater concern to me is the absence of any provision in CC language to prevent commercial and other entities from taking CC-BY-SA content and placing in a container that restricts the letter and spirit of the CC license. Specifically, paywalls, proprietary formats, obfuscated code and things like that. The SA part should be defined broadly to include everyone, not just the original author.

  7. Frank,

    As mentioned above, there is a CC Public Domain Mark; the problem not the “should be” being abundantly clear, but in turning that “should be” into “are”.

    Regarding BY-SA and non-copyright restrictions, the license does not permit “effective technological measures on the Work that restrict the ability of a recipient of the Work from You to exercise the rights granted to that recipient under the terms of the License” (ie DRM) and it does “include the right to make such modifications as are technically necessary to exercise the rights in other media and formats.”

    I think that’s a pretty happy medium. Not permitting proprietary formats or paid distribution would make the license less than fully free/libre/open and harder to interpret. Can you point out places where BY-SA content is locked up in ways you find offensive and not otherwise available? Is this why you use BY-NC-SA (as I see you do on your site — note that this license doesn’t prohibit proprietary formats either)?

Comments are closed.