Reasons to use a CC BY-SA licence instead of CC BY?


Many governments have adopted the Creative Commons Attribution (CC BY) licence as a default. On these same Governments open data portals I come across data licensed using the Creative Commons Attribution-ShareAlike (CC BY-SA) licence.

The main difference between the two licences is CC BY-SA requires that:

If you remix, transform, or build upon the material, you must distribute your contributions under the same license as the original.

Assuming governments are publishing their own data, and not re-publishing other data that has a CC BY-SA licence, is there any good reason to use the more restrictive CC BY-SA licence instead of CC BY?


Any creative work is subject to the “automatic protection” of the Berne Convention, that is a kind of implied license. In response to this lack of freedom, we must to express the protections that are waiving. The CC0 is the response for “waiving as many rights as legally possible”, and Governments can use CC0… Any other license is a “less freedom than CC0” license. See types of copyright license.

On my opinion, the inclusion of the share-alike (SA) clause in a license is not exactly a matter of “Berne/CC0 balance”, as CC-BY, but a market strategy, that isolates the “open works” from the non-open market. The SA clause enhances the “open heritage”.

PS: there are a trade off. Software houses, for example, when using open licenses, prefer MIT or Apache licenses than GPL (has a SA clause), because can’t to isolated their products.


Good analysis, a couple quibbles.

In the first sentence I think you’re misusing the term implied license, “an unwritten license which permits a party (the licensee) to do something that would normally require the express permission of another party (the licensor).”

“Automatic protection” does not fit this definition.

I don’t know what you mean by the phrase “Berne/CC0 balance”.

You’re quite right that SA licenses are a market strategy (except perhaps when mandated by law, if/when that is the case), but then so are any licenses; each just reflects a different market strategy.


So if a government has said they are “open by default” and the data they are publishing is only theirs (i.e. No embedded data from others that is licenced BY-SA), then I can see no reason not to publish using CC-BY.


Thanks @mlinksva. Sorry my English (is not an usual language for me), I will try to explain… But at the risk of too much explanations (less is more :slight_smile:

The “balance” is a compensation: the Berne Convention imposes some “default rules”, with no choice, except with (at the expense of) some transaction cost. The standardization of the “opposite of Berne” licenses, in the last years, as the CC0, and other popular options like GPL, CC-BY, etc. are now the low-cost choices. This “balance” is an important context to describe the main and most popular license clauses (BY, NC, ND, and others in Berne or in opposition to Berne clauses).

About formal definition of “implied license” (or implicit license), I think there are no international agreement, so, it is an informal concept… But is sufficient as label of “opposite of explicit license”. A link to CC-BY in a webpage, an attachment in a Gutenberg’s book, or a symbol “© Crown copyright” (OGL-UK) in a PDF law of, all these are examples of “explicit licenses”… There are a license when it is not explicit? Yes it is possible (!). All 7.9 million of documents listed here by LexML have an implied license, and it is materialized here.

If this notion of “implied license” makes sense, we can say that the “automatic protection” is an implied license (!)… And will be a good idea to express (to materialize in a license format), as we trying here.

@Stephen, to say “open by default” is possible, since the government “say” something permitted by Berne, as Article 2.4, and on a document that have force of law, like the evidences cited here.
PS: of course, usually, “explicit licenses” are better than “implied licenses”.

About “to publish using CC-BY” in the government, is a good practice and have some good references as UK (that use OGL-UK=CC-BY), but for public sector the usual recommendation is CC0 or a CC0 variation. Example: the “BY clause” in a map causes more cost tham CC0, for practical work in the openstreetmap community, and map features reuse and distribution.


Implied/implicit is not what I have a problem with, license is. Automatic protection is not a license.


@mlinksva You showed a good point, let’s check how to get back some conceptual agreement…

The Berne’s automatic protection is like a license in its “contractual effects” (in this sense CC0 is opposed to Berne), if we not consider a “license-like constraint”, what kind of constraint automatic protection is it?


Licenses don’t constrain. They give permissions that come with conditions which you may either accept or not take the license. Automatic restriction is just not like a license.


@mlinksva, Your point make sense, and I agree. Perhaps I need to split my position in contexts…

For a rigorous debate and a “checklist of the consensual truth”, we need a legal/juristical reference. For final user needs, that is my aim, need a cause/effect analysis. So I will try to review from this two main contexts,

  • in a rigorous use of language (legal/juristical jargon and conceptual frameworks) we have two distinct things,

    • Berne: their clauses are components of statutory law of each country.
      NOTE: user not need to read or understand all this law-documents, but at each country, each person will suffer the effects of the clauses. How to “translate” (in licenses or complementar FAQ/help documents) this effects for commom people, who do not have time to spend and no special legal knowledge?

    • public (reusable) licenses, license templates and individual contracts: they give permissions, and when a user accepts a license it is like to sign a contract, so is in the contractual right framework.
      NOTE: … even when the statutory law applies (is implicit in the contract) the usual contrac-text not need to cite or copy/paste all law-text, but is important to user know… In a country where the statutory law is under the ignorantia juris non excusat principle, unfortunately, the usual reality is the ignorance (so that is the big problem!).

  • in a usual/colloquial language (like the “commons deed text” in a “Three layers architecture of contents”), we need to express cause/effect to final users… How to say that “a kind of license exists by default”? (and express the clauses that are in effect)
    PS: I suppose we are in consensus that “The cause is Berne, the effect is like a set of license clauses”.

PS: perhaps a solution to the “big problem” is to express “Berne license-like clauses” for usual people, in a kind of implied-license deed.


@Stephen to return to your original question: “is there a good reason to use the more restrictive Share-Alike license?”

The main answer is that share-alike could be a way to create a larger open commons by

a) ensuring downstream users "share back"
b) giving encouragement original providers to share by reassuring them that there won’t be too much “free-riding” on their efforts (e.g. people using it “for free” and not sharing back)

It could also be used in dual-licensing model by a savvy government. This would involve them sharing their data openly under share-alike and also making it available (for a fee) without the share-alike restriction for users who want that (this is the “MySQL business model”). As far as I know, no government has tried this, but it is an option and also a potential way to address gov concerns about losing revenue (I think I alluded to this in the various papers I wrote on financing and supplying public sector information).

I am not saying btw that gov should do this - just explaining a potential logic.

There is more on this in this post of mine The post as a whole is about scientific data and a specific debate there but parts are relevant here and I’ve excerpted some)

Share-alike provisions are more controversial. It has been argued that share-alike conditions are problematic because of the potential for incompatibility between two share-alike licenses (or community norms). At the same time share-alike may provide an important incentive for individuals and communities to make their data openly available since it provides some assurance that this data will remain open. Thus, any evaluation comes down to the balance between:

  1. The costs, if any, of allowing share-alike in terms of e.g. complexity and compatibility.
  2. The benefits, if any, that share-alike provides by encouraging the creation of open data in the first place and in ensuring subsequent ‘sharing back’ by those who build upon that data.

In my view the benefits are substantial while the costs are not. Incompatibility can largely be avoided by only ‘approving’ share-alike licenses that are compatible. At the same time, share-alike enshrines a principle that is important to many communities in the code and content spheres and same seems true of data (consider e.g. Open Street Map). …


While it’s practically true that share-alike can work sort of like “share back”, it’s better to make it totally clear that it’s a downstream clause. There’s no requirement to share back upstream at all. The motivation can be best described as “pay it forward”. It’s saying that you’re providing openness and freedom to people under the requirement that they pass on the same openness and freedoms to other to keep things open always.

There’s too often a confusion about copyleft / share-alike being an upstream license, and that framing is upstream-centric, i.e. focuses on the concerns of developers or authors or researchers first and less on the concerns of citizens otherwise.


I would counsel strongly against the sort of dual licensing arrangements suggested as an option by Rufus for governments.

The evidence from the POPSIS study and elsewhere suggests that the use of data once made Open goes up by much more than one would expect from the reduction in price alone. (Indeed, although I have not gone back to check, I seem to recall that Rufus’ original papers had a multiplier reflecting this.) I have always considered that at least some of this effect is due to the removal of non-financial conditions on how the data can be used - which, if one is charging, are essential to protect one’s revenue stream from other customers. It’s the absence of conditions, and the need to negotiate them, that maximises the opportunity for innovation and for new challenger entrants to established markets.


Dual licensing (odbl or fee) has been discussed for a while in France,
we’ve been advocating it at Regards Citoyens as a solution to start opening
public payant datasets that some administrations have always sold before
and refuse to open fearing to lose their income
It was largely discussed by MPs in the National Assembly last month during
the debates on a big digital bill for instance with this amendmen:t

Debates can be read in french here, the idea being supported by MPs from
the right to the left wing:

In addition to such cases, I personnally believe Share-Alike can often be a
better choice to enable small reusers to compete with the largely
established companies by allowing them to develop apps in a common
ecosystem whereas Google or such refuse to share back their preestablished
data assets

Benjamin Ooghe-Tabanou