flowchart TB
start("We want<br>to choose<br>a license.") --"We adapted a work by<br>others shared under a<br>free/open license."--> use_existing_license["<em>Use its license</em>"]
start --"We created the work<br>entirely by ourselves."--> regulation("Other regulation<br>(by community or funder)<br>concerning license?")
regulation --"Exists"--> follow_existing_norms["<em>Follow that</em>"]
regulation --"Does not<br>exist"--> type("Work type?")
type --"Software"--> apache["Apache 2.0"]
type --"Writing, image, audio, video"--> cc0["CC0 1.0"]
type --"Data(base)"--> existing_license_data("Adapting individual<br>data entries by others?")
existing_license_data --"No, we created them<br>entirely by ourselves."--> cc0_data["CC0 1.0 <em>for database<br>and its content</em>"]
existing_license_data --"Yes, they were<br>shared under a<br>free/open license."--> use_existing_license_data["<em>Use their license<br>for content and</em><br>CC0 1.0 <em>for<br>the database</em>"]
click apache href "https://choosealicense.com/licenses/apache-2.0/"
click cc0 href "https://creativecommons.org/publicdomain/zero/1.0/"
click cc0_data href "https://creativecommons.org/publicdomain/zero/1.0/"
Choose a License
TL;DR
Every expression of an idea, such as a literary or an artistic works, is automatically covered by copyright. Put simply, this means that nobody else than the copyright holder is allowed to copy, modify, or share it. Copyright licenses can grant others some of the necessary rights, but mostly don’t cover publicity, privacy, moral, patent, or trademark rights. Certain exceptions and limitations to copyright (such as fair use or the right to quote) may obviate the need for a license.
Researchers are only allowed to share material they have the necessary rights to. This includes copyrights, which are granted by free/open copyright licenses. These licenses often require to attribute the authors, indicate whether changes were made, and provide the text of the license, among other things. In addition, researchers should share any work by their own under such a license.
Choosing a license involves:
- Recording the copyright notices and licenses for content by others (for example, in a file called
LICENSE.txt) and following all their additional requirements. - In the simplest case, choosing two licenses (one suitable for code, one for everything else) which to apply simultaneously to all own content (for example, again by recording the license in the file
LICENSE.txt). Recipients are then free to choose under which of the two licenses to use the material.
We recommend to choose CC0 1.0 and Apache 2.0 for all own works. Take your time to read the summaries behind the links and/or the actual license text to understand its legal effect.
You can adapt the following wording to your use case:
Except where noted otherwise, all files in this project are made available under CC0 1.0 or (at your option) under the terms of the Apache Software License 2.0.
A Primer on Licenses
Whenever you create a literary or artistic work (such as a text, image, video, or software), the copyright law in most countries limits other people from copying, modifying, and sharing it without your express permission. This even applies if you make it available to others (e.g., on your website): First and foremost they are not allowed to copy, modify, or share it. This legal default of “all rights reserved” was created to benefit publishers, not authors (Fogel, 2006), and runs counter to many cultural and scientific processes. Copyright licenses enable authors to free up their works for reuse by others.
If the work was created as part of your job, it might be your employer who holds the copyright to the work, depending on the country and contract.
A license is a legal document that regulates what others are allowed to do with a copyright-protected work – the licenses we discuss are non-exclusive licenses and do not limit the author or copyright holder in their rights. And while you could write your own license, there are already many pre-formulated licenses available to choose from and to apply to your work.1
For the purpose of this tutorial, by license we mean copyright license.
Warning 1: Giving Publishers Exclusive Rights
Some journals offer to publish your article under a Creative Commons license, but still demand an exclusive publishing and distribution license or a copyright assignment from you. This would give them more rights than the readers of the article have through the respective Creative Commons license (Rumsey & Labastida, 2022) and exceeds by far what is necessary to make publication possible (Suber, 2022). Consequently, authors should oppose this practice and grant publishers the same rights that every other reader of the article has. To facilitate self-archiving, one can also modify the contract with publishers via a rights retention statement (UK Reproducibility Network & Eglen, 2023) / author’s addendum (SPARC, 2006). If you have published a closed-access paper before, you can consult ShareYourPaper for legal options to still make it available free of charge to readers.
If your chosen publisher insists on an exclusive license, you may at least retain the copyright for your figures – follow the guide “Retaining copyright for figures in academic publications to allow easy citation and reuse” by Elson (2016) to learn how to do that.
The licenses we discuss here mostly regard copyright. Therefore, recipients may lack other rights such as publicity, privacy, moral, patent,2 or trademark rights. For example, sharing photos that depict people is not only a matter of copyright, but also of privacy rights. Conversely, the licenses do not apply if recipients are allowed to use the works for other reasons such as fair use, the right to quote, or because they made a different arrangement with the author.
Which License to Choose for a Work?
Many boilerplate licenses are available to apply to your work. Which license is appropriate depends on several factors, including existing licenses in place and the type of work, but also your personal considerations. We strongly recommend to apply a free/open license to your work, which means that the work “can be freely studied, applied, copied and/or modified, by anyone, for any purpose” (Möller et al., 2015). Importantly, this also means that others do not need to ask or notify the author and that they can use it for commercial purposes. By the choice of license, authors can, however, demand that they are credited, that the original license is indicated, that modifications are indicated, that derivative works are only shared under the same license, and that no further restrictions are imposed on the work. Software licenses may additionally require to make the source code available to everybody the software is shared with and often require to display the full text of the license upon usage. Because there are many free/open licenses available, the licenses discussed here only represent a recommended subset.
If you would like to choose a license not listed here, it should be appropriate for the type of work in question and be compatible with the dominant copyleft license in the respective community (see also Note 3). For software that’s almost universally the (A)GPLv3 and for writing, image, audio, and video that’s mostly the CC BY-SA 4.0. For data, no dominant copyleft license has emerged yet, so any of ODbL 1.0, CDLA Sharing 1.0, and CC BY-SA 4.0 are acceptable.
Note 1: A Note on Terminology
The terms free and open, especially with regard to software, come with a history. The Free Software Foundation (FSF) was founded in 1985 to protect “four essential freedoms” (2021) of a program’s user. These are the freedoms to use, study, share, and improve a program. Software whose users legally and practically have these freedoms (because, among other things, they have access to its source code) is considered free. The four freedoms are seen as vital for a society as a whole in the sense that they enable sharing, cooperation and ultimately freedom in general. The FSF maintains a list of software licenses that it considers to be protecting the four freedoms. Sometimes the term libre (Spanish and French for free) is used to make a distinction from gratis software. You can learn more about free software at Write Free Software.
The Open Source Initiative (OSI), which was founded in 1998, follows a more pragmatic approach. It is concerned with developing high-quality software, for which everyone’s ability to obtain, modify and contribute back the source code is considered beneficial. Access to the source code is one out of multiple conditions for software to be considered open source by the OSI (2007), which equally maintains a list of approved licenses.
Access to the source code is a necessary, but not sufficient requirement both for free software and for open source software. Conversely, both can (and frequently are) sold for money, as their respective criteria only apply once one has access to the software. Throughout this tutorial, we write “free/open license” to mean a license that is approved by either the FSF or the OSI. Software which is neither but makes available its source code is sometimes referred to as source-available software.
Apart from software, the term open access has often been used for works that are available at no cost. For example, this is the commonality of bronze, green, hybrid, gold, and diamond/platinum open access articles, which otherwise vary in the rights that are granted to readers. In 2002, the Budapest Open Access Initiative declared that open access additionally includes the right to use articles for any purpose, and in 2003, the Bethesda Statement and the Berlin Declaration added the right to make derivative works.
Two other notable definitions include the Open Definition (Open Knowledge Foundation, 2016), which was first drafted in 2005, and the definition of Free Cultural Works (Möller et al., 2015), for which the open editing phase began in 2006. They are largely viewed as compatible with one another.
Existing License?
First, if you adapt (i.e., modify, remix, build on) a work by others you need to determine if it is provided to you under a free/open license. If yes, we recommend you to make your contribution available under the same license.3 For example, if you adapt code published in another paper, choose the same license for your modifications. The same applies if there are strong community norms to use a particular free/open license4 or if a particular license choice is mandated by your funder. Importantly, as discussed before, you are generally not allowed to adapt a work not published under a free/open license.
You can determine whether a license is free/open by searching for its name in the SPDX License List and looking for at least one
Y in the two columns FSF Free/Libre? and OSI Approved?Work Type?
If you create a new work and no strong community norms suggest a particular license, you need to choose the license yourself. Which license to choose depends on the type of work you create. Software licenses, for example, may consider that the source code is the preferred form for making modifications, while licenses for data can differentiate between the database and any works produced from it. We have created a flowchart that covers the most likely types of works you will create as a researcher: software, writing (i.e., text), images, audio, video, and data (see Figure 1). This flowchart always recommends the most permissive license possible to maximize reuse – though we provide two additional flowcharts below that allow for more choices. Click on the name of a license to learn more about it.
Tip 1: Multi-licensing
Sometimes, the type of a work is not obvious. For example, a Quarto document…
- …contains both R code and writing, and
- …may be distributed in the source format or as rendered document, possibly including images.
One may wonder which license to apply in this case, because Creative Commons licenses are not recommended for source code5 and applying software licenses to PDFs or images can lead to confusion or nuisance.6
One solution is to make such a work simultaneously available under two (or more) licenses, at the choice of the recipient: Either under a specified software license, or under a Creative Commons license7. This is called multi-licensing and makes it easier to reuse both the rendered document as well as the code. For example, one could write:
The Quarto files in this project are made available under CC0 1.0 or (at your option) under the terms of the Apache Software License 2.0.
Note, that for data there is, at least in principle, the possibility that a different license applies to the individual entries than to the collective database. For example, if you were to create a database of artworks by others, those artworks would be licensed individually as chosen by the artists, but the license for the database as a whole could be chosen by you. The latter includes the selection of entries, field names and any sui generis database rights. However, if the content was created by you, we recommend you to choose the same license for both content and database. Factual data (like measurements or metadata) should be licensed under CC0 1.0 (Villa, 2016a, 2016b) – otherwise, consult Note 2 for some caveats.
Note 2: Other Work Types
One should take special care when applying license to the following types of works:
- fonts: Copyleft licenses applied to fonts can be a special case: If a font is put under the license CC BY-SA 4.0, any documents containing texts using that font will probably be derivative works and have to be put under the same license if shared. If the intent is that only derived fonts, if published, have to be put under the same license, the SIL Open Font License 1.1 is an appropriate choice. Note, however, that it doesn’t require attribution for usage. If no copyleft mechanism is intended, CC0 1.0 also works for fonts.
- templates and TeX packages: If a template or, for this purpose, a \(\TeX\) package, is licensed under a copyleft license such as the (A)GPLv3, every work that is a derivative of the template or that uses the \(\TeX\) package has to put under the same license, if shared (Koppor, 2016). And if a document contained source code covered by the (A)GPLv3, the same license would also apply to the document (MadHatter, 2022).
- database content: If the work produced from a database (the “output”) is a derivative of the content in the database, the output is subject to the restrictions laid out in the license. For example, if geospatial data were to be licensed under CC BY 4.0, all maps produced from the data would likely need to fulfill this license’s obligation for unrestricted access (its Anti-DRM provision) if shared (see Poole, 2017). Similarly, following an example from Matt (2009), if one were to choose the copyleft license CC BY-SA 4.0 for this purpose, any map that is a derivative of the data would also need to be licensed under CC BY-SA 4.0 (or a compatible license) if shared. If the intention is to only have derivative databases under the same license, one might want to choose the ODbL 1.0 for the database, as it was specifically designed not to apply to works produced from the data in the database. Otherwise CC0 1.0 is an excellent choice for data.
- works in the public domain: If a work is already in the public domain, it should be marked using the PDM 1.0, rather than applying a waiver such as the CC0 1.0 (or another license).
Copyleft?
The consequence of applying a permissive license to your work is that others may distribute their adaptations from it under stricter terms, possibly without a free/open license. You may prefer that everybody who receives somebody else’s adaptations of your work enjoys the same rights that you granted to them. This can be achieved by copyleft licenses.8 For software, they come in two flavors:
- Weak copyleft licenses only require that modifications to the software itself are licensed under the same or a compatible license if shared. For example, if you create and publish software under a weak copyleft license, others who modify it and put their version on the internet have to apply the same license. However, people who merely use your software in their own work which they make publicly available can choose any license for it.
- Strong copyleft licenses, on the other side, insist that any larger works that use the copyleft-licensed software must also be licensed under the same or a compatible license if shared. For example, if your software were to be put under a strong copyleft license, everybody publishing software that uses your software would need to put it under the same license. Because of only few rulings in courts, the extent of this requirement is disputed (Wikipedia contributors, 2024).
Note, however, that the copyleft licenses we discuss here do not mandate sharing. Copyleft (and attribution) clauses are only triggered if the work is shared (Creative Commons, 2015). This means that if you only use a work internally, you do not need to share your derivative works. It is also worth reiterating that these licenses do not restrict the original author(s): They are still permitted to distribute their work under a different license and without sharing the source code.
Tip 2: Projects Involving R Code
In most cases, the output of software, like images or tables, does not depend on the software’s license.9 Therefore, if you use an R package under a copyleft license to create a figure, you are likely the copyright owner. However, if the output is based on data, it can be considered a derivative work of the data and the license of the data also applies. For example, maps may be considered as a derivative of the geographic data they are based on.
It is disputed whether software that uses an R package under AGPLv3 or GPLv310 can only be published under a GPL-compatible license – or even has to be published under the same license. Posit, the company behind RStudio, does not believe that to be the case (see also Wickham & Bryan, 2023).
You can learn which license an installed package uses via packageDescription("<PACKAGE_NAME>", fields = "License"). And to identify which licenses are being used by the R packages your project depends on, you can use the following code:
deps <- renv::dependencies()$Package |>
unique() |>
pak::pkg_deps(dependencies = NA) |>
getElement("package")
unique(installed.packages(fields="License")[deps, "License"])We have prepared two advanced license flowcharts, one for software, writing, images, audio, and video in Figure 2 and one for data in Figure 3 where you can make additional choices. Note, however, that especially the advanced flowchart for licensing data is quite complex and we recommend you to seek legal counsel if you want to be sure.
Attribution, State Changes, and Anti-DRM?
The advanced license flowcharts also allow you to make additional decisions:
- Attribution means whether recipients of your work are required to provide attribution to you. For software licenses, the attribution is part of a copyright notice. Note that even if attribution is not a requirement of the license, good scientific practice demands that appropriate citations are made.
- State changes means that recipients need to indicate if changes were made.
- Anti-DRM means that when others share your work they are not allowed to apply technological measures restricting anything that the license permits. Implementing such restrictive measures is also referred to as digital rights management (DRM).
From the Creative Commons licenses, only CC0 1.0 does not require providing attribution. All software licenses in Figure 2 require providing attribution, although only Apache 2.0 and AGPLv3 require others to indicate changes. For data, the ODbL 1.0 allows for technological measures that restrict the work only if a version of the database is provided in parallel without such measures.
Note 3: Other Restrictions
As indicated before, a free/open license must allow creating derivative works and must allow exercising the rights granted by it for any purpose, including commercial use. There are a few good resources on the reasoning behind that (e.g., see Klimpel, 2013; Möller, 2023; Open Data Institute, 2015; Stallman, 2022b; Vézina, 2020; Wikimedia Commons contributors, 2009), but we would like to highlight one reason in particular.
With the rights provided by free/open licenses comes the possibility to build on and combine multiple works by different authors, which is essential for any cultural and scientific activity. However, this is only possible if the various licenses involved are compatible with each other. For example, one is allowed to remix two figures if the first is licensed under CC BY 4.0 and the second under CC BY-SA 4.0 because the licenses were written to be compatible with each other (compare Creative Commons, 2024b). Also note that CC BY-SA 4.0 is one-way compatible with the GPLv3, which in turn is compatible with the AGPLv3.
However, applying a restriction such as only permitting non-commercial use or not allowing derivative works creates silos of works which are mutually incompatible with each other. Put differently, one cannot share a remix of two works where one is licensed under CC BY-SA 4.0 and the other only allows non-commercial use. In order to avoid silos, one should only choose licenses which are compatible with the dominant copyleft license in the respective community (Lämmerhirt, 2017; Wheeler, 2014). If you would like to learn more about the different types of compatibility, we recommend you to read the article “A Quick Guide to Software Licensing for the Scientist-Programmer” by Morin et al. (2012). The following diagram provides an overview of the compatibility of various licenses:
Which Licenses to Choose for a Project?
So far, we only discussed how to choose a license for works of one type. But what if you want to share a project with all kinds of files? For example, the project from this tutorial (among other things) includes a data file, a manuscript file with intermingled code and writing, and an R file. And what if you also want to share files by others, as is the case with apaquarto which you may have installed in this project?
The answer is that you need to indicate the license on a per-file or per-folder basis (rather than choosing one for the whole project). Sometimes, even different parts of a file might be subject to different licenses. The easiest approach is to make note of every foreign work included in your project and record its license. Then, multi-license all the remaining files, which are yours, under a code and a non-code license in parallel as explained in Tip 2.
Applying the License
Having selected the licenses of your choice – again, you might need multiple ones depending on the types of works your project contains –, we encourage you to read through the full license text (or at least a legal summary) to understand their effect. Then, you can record the license of existing content and apply the licenses of your own contributions. Mostly, this just means indicating which file or folder is covered by which license, usually in the project’s README (among other places), whose creation will be discussed later. Notably, no sort of registration is required.
Caution 1: License Versions Are Important
You may have noticed that we mostly refer to licenses using a name and a version number. This is because the organizations that created the licenses sometimes publish updated versions to accommodate for developments in copyright law and the communities that use the licenses. For example, the Creative Commons licenses (that start with CC) were first published in 2002. Since then, the possibility to relicense under later (v2.0, 2004) and compatible licenses (v3.0, 2007) has been added, a 30-day window to correct license violations has been established to combat copyleft trolls, and sui generis database rights are covered explicitly (v4.0, 2013). There are many more subtle differences between license versions, therefore it is important to indicate which license version exactly one is referring to, as the license of a work does not “update” automatically. CC0, officially introduced in 2009, is still at version 1.0.
For the AGPLv3 it is even recommended to state whether a work is licensed under exactly the indicated version of the license or, alternatively, also under newer versions of the license (Stallman, 2022a).
For example, the apaquarto extension that you included in your project is a work by others.11 You need to indicate its license so that others know what they are allowed to do – and, of course, you need to comply with any terms yourself, such as retaining the copyright notice.12 In contrast, if it were not for this tutorial, the manuscript would contain your own work and you would need to indicate under which license you provide it to others.
Tip 3: Follow the Help Provided by the License Authors
For all the licenses recommended in this tutorial, the organizations that created these licenses provide more information on how to apply them to your work:
Creative Commons even provides a range of considerations for licensors and licensees (Creative Commons, 2013) and an interactive chooser which you can use to create text snippet that you can copy and paste to the desired location.
While it is common to state the chosen license(s) in the README, usually one of the following actions is taken in addition:
In the simplest case, one just creates a file called LICENSE.txt where the full text of the license is copied verbatim. This is a practice propagated by GitHub, which provides instructions for and comparisons of many licenses via ChooseALicense.com. However, if the project is not completely covered by one single license, this practice may become unwieldy. For example, if a project contains different types of works by different authors, the LICENSE.txt needs to detail which file is covered by which license(s), along with any copyright notices.
Individual programming languages also have their own way of stating which license a package is distributed under. For R packages, this is usually set by the field License in the file DESCRIPTION (Wickham & Bryan, 2023).
Finally, Note 4 explains how to use the REUSE specification to make the choice of license machine-readable. This is the approach we recommend taking.
Note 4: Using REUSE to Record Licenses
Every major free/open license has a unique SPDX identifier which allows communicating the license choice unequivocally. We will be using that to indicate the license for every file in your project, along with the year of publication and the copyright holder. To do this, we add a comment to the beginning of every file and include the two tags SPDX-FileCopyrightText and SPDX-License-Identifier. How this works depends on the file type, as the syntax for a comment varies.
For example, if you previously created the file create_data_dictionary.R, you can now add the following comment to the beginning of the file, replacing <YEAR> and <NAME> with the current year and your name – of course, you can also choose a different license:
create_data_dictionary.R
# SPDX-FileCopyrightText: <YEAR> <NAME>
#
# SPDX-License-Identifier: CC0-1.0You need to use # to start the comment because this is the symbol that starts comment lines in R scripts. Alternatively, you can use the reuse tool to add these information for you. After installing it with…
Terminal
pipx install reuse…you can add the copyright information using the following command – the current year will be added automatically:
Terminal
reuse annotate --copyright="<NAME>" --license="CC0-1.0" create_data_dictionary.RIn many cases, the reuse tool will figure out the appropriate comment style for you. If this is not the case, as currently with Quarto files, you can tell it directly which comment style to use (html in this case):
Terminal
reuse annotate --copyright="Josephine Zerna <josephine.zerna@tu-dresden.de>" --copyright="Christoph Scheffel <christoph_scheffel@tu-dresden.de>" --copyright="Florian Kohrt" --license="CC-BY-4.0" --style=html Manuscript.qmdThis adds the following header to Manuscript.qmd:
Manuscript.qmd
<!--
SPDX-FileCopyrightText: 2024 Christoph Scheffel <christoph_scheffel@tu-dresden.de>
SPDX-FileCopyrightText: 2024 Florian Kohrt
SPDX-FileCopyrightText: 2024 Josephine Zerna <josephine.zerna@tu-dresden.de>
SPDX-License-Identifier: CC-BY-4.0
-->Note that Manuscript.qmd was provided to you under CC BY 4.0, which is what you indicate with the previous comment. If you edited the file, you may also add yourself.13
Sometimes, there are file types which do not allow for adding the license information inside them, such as PDF and CSV files. For these, a corresponding .license file can be created. Try the following command which indicates that the data were published under CC0 1.0:
Terminal
reuse annotate --copyright="Kristen Gorman" --license="CC0-1.0" data.csvYou will notice that this creates another file called data.csv.license containing the relevant information:
data.csv.license
SPDX-FileCopyrightText: 2024 Kristen Gorman
SPDX-License-Identifier: CC0-1.0If you want to indicate the license for all files in a particular folder, you can create a file called REUSE.toml and add an [[annotations]] table for them:
REUSE.toml
version = 1
# apaquarto extension from https://github.com/wjschne/apaquarto
[[annotations]]
path = "_extensions/wjschne/apaquarto/*"
SPDX-FileCopyrightText = "2024 William Joel Schneider <w.joel.schneider@gmail.com>"
SPDX-License-Identifier = "CC0-1.0"Finally, there may be some minor files which are build artifacts. You can either add them to your .gitignore file or use the CC0 1.0 license/waiver with a copyright tag such as SPDX-FileCopyrightText: NONE to assert that there is no copyright holder. For more information, also discussing other corner cases, you can read their Frequently Asked Questions.
Once you are done, you can download the texts of all indicated licenses using…
Terminal
reuse download --all…and verify that you did not miss a file by running…
Terminal
reuse lintRegardless of how exactly the licenses are added to the project, this is a good opportunity to verify one last time that all third party content is provided to you under a free/open license and that you comply with it. Please add a license to your project now, either creating a file LICENSE.txt or following the REUSE standard.
Wrap-up
If you would like to learn more about copyright and licenses you might find the following resources interesting:
- “Open Content – A Practical Guide to Using Creative Commons Licences” by Kreutzer & Deutsche UNESCO-Kommission (2014)
- “Creative Commons Certificate for Educators, Academic Librarians, and Open Culture” by Creative Commons (2024a)
- “Freie Software – Zwischen Privat- und Gemeineigentum” by Grassmuck (2004)
Additional Figures
flowchart TB
start("We want to choose a<br>license for software,<br>writing, image, audio,<br>or video.") --"We adapted a work by<br>others shared under a<br>free/open license."--> use_existing_license["<em>Use its license</em>"]
start --"We created the work<br>entirely by ourselves."--> regulation("Other regulation<br>(by community or funder)<br>concerning license?")
regulation --"Exists"--> follow_existing_norms["<em>Follow that</em>"]
regulation --"Does not<br>exist"--> type("Work type?")
type --"Software"--> code_sa("Attribution?<br>State changes?<br>Copyleft?")
type --"Writing, image, audio, video"--> nocode_cc("Attribution?<br>State changes?<br>Anti-DRM?<br>Copyleft?")
code_sa --"Attribution &<br>State changes"--> apache["Apache 2.0"]
code_sa --"Attribution &<br>Weak copyleft"--> mpl["MPL 2.0"]
code_sa --"Attribution &<br>State changes &<br>Strong copyleft"--> agpl["AGPLv3"]
nocode_cc --"Neither"--> cc0["CC0 1.0"]
nocode_cc --"Attribution &<br>State changes &<br>Anti-DRM"--> cc_by["CC BY 4.0"]
nocode_cc --"Attribution &<br>State changes &<br>Anti-DRM &<br>Copyleft"--> cc_by_sa["CC BY-SA 4.0"]
click apache href "https://choosealicense.com/licenses/apache-2.0/"
click mpl href "https://choosealicense.com/licenses/mpl-2.0/"
click agpl href "https://choosealicense.com/licenses/agpl-3.0/"
click cc0 href "https://creativecommons.org/publicdomain/zero/1.0/"
click cc_by href "https://creativecommons.org/licenses/by/4.0/"
click cc_by_sa href "https://creativecommons.org/licenses/by-sa/4.0/"
Note. DRM = digital rights management
flowchart TB
start("We want to choose a<br>license for data.") --"We adapted a database by<br>others shared under a<br>free/open license."--> use_existing_license_db["<em>Use its license(s)<br>for content and database</em>"]
start --"We created a database<br>entirely by ourselves."--> regulation("Other regulation<br>(by community or funder)<br>concerning license?")
regulation --"Exists"--> follow_existing_norms["<em>Follow that</em>"]
regulation --"Does not<br>exist"--> existing_license_content("Adapting content<br> by others?")
subgraph content["<strong>License for content</strong>"]
existing_license_content --"No, we created the content<br>entirely by ourselves."--> facts("Entries are facts<br>(like measurements<br>or metadata)?")
existing_license_content --"Yes, it was<br>shared under a<br>free/open license."--> use_existing_license_content["<em>Use that license</em>"]
facts --"Yes"--> cc0_content_metadata["CC0 1.0"]
facts --"No"--> choose_license["<em>Consult flowchart for<br>software, writing,<br>image, audio, and video</em>"]
end
subgraph database["<strong>License for database</strong>"]
choose_license --> switch_license["<em>Depending on<br>content license</em>"]
use_existing_license_content --> switch_license
cc0_content_metadata --> cc0_db["CC0 1.0"]
switch_license --"CC0 or<br>non-CC license"--> sa("Attribution?<br>Anti-DRM?<br>Copyleft?")
switch_license --"CC BY or<br>CC BY-SA"--> same["<em>Same license for DB</em>"]
sa --"Neither"--> cc0_db
sa --"Attribution &<br>Anti-DRM &<br>Copyleft"--> odbl["ODbL 1.0"]
%% the following link is only added to have terminal nodes on the same level
sa ~~~ same
end
click cc0_content_metadata href "https://creativecommons.org/publicdomain/zero/1.0/"
click cc0_db href "https://creativecommons.org/publicdomain/zero/1.0/"
click odbl href "https://opendatacommons.org/licenses/odbl/summary/"
References
apsillers. (2019, March 11). Answer to "does "the GPL doesn’t cover the output of a program" also apply if the output is source code?". Open source stack exchange. https://opensource.stackexchange.com/a/8060/34940
Creative Commons. (2013, November 25). Considerations for licensors and licensees. https://creativecommons.org/share-your-work/licensing-considerations/version4/
Creative Commons. (2015, August 8). ShareAlike compatibility analysis: GPL. https://wiki.creativecommons.org/index.php?title=ShareAlike_compatibility_analysis:_GPL&oldid=114476
Creative Commons. (2024a). Creative commons certificate for educators, academic librarians, and open culture. https://certificates.creativecommons.org/cccertedu/
Creative Commons. (2024b, June 3). Answer to ’can I combine material under different creative commons licenses in my work?’ in frequently asked questions. https://creativecommons.org/faq/#can-i-combine-material-under-different-creative-commons-licenses-in-my-work
Elson, M. (2016, November 11). Retaining copyright for figures in academic publications to allow easy citation and reuse. Medium. https://medium.com/@malte.elson/retaining-copyright-for-figures-in-academic-publications-to-allow-easy-citation-and-reuse-77c6e2b511fe
Fogel, K. (2006, May 27). The surprising history of copyright and the promise of a post-copyright world. Question copyright. https://questioncopyright.org/promise/
Free Software Foundation. (2021, February 2). What is free software? https://www.gnu.org/philosophy/free-sw.html
Grassmuck, V. (2004). Freie Software: zwischen Privat- und Gemeineigentum (2., korrigierte Aufl). Bundeszentrale für politische Bildung. https://freie-software.bpb.de/
Klimpel, P. (2013). Free knowledge based on creative commons licenses. Wikimedia. https://meta.wikimedia.org/wiki/Free_knowledge_based_on_Creative_Commons_licenses
Koppor. (2016, July 23). Answer to ’GPL-licensed LaTeX template - implications for resulting work?’. Open source stack exchange. https://opensource.stackexchange.com/questions/2735/gpl-licensed-latex-template-implications-for-resulting-work/4206#4206
Kreutzer, T., & Deutsche UNESCO-Kommission (Eds.). (2014). Open content: A practical guide to using creative commons licences. German Comm. for UNESCO. https://meta.wikimedia.org/wiki/Open_Content_-_A_Practical_Guide_to_Using_Creative_Commons_Licences
Lämmerhirt, D. (2017). Avoiding data use silos: How governments can simplify the open licensing landscape. Open Knowledge International. https://web.archive.org/web/20231002141440/https://research.okfn.org/avoiding-data-use-silos/
MadHatter. (2022, March 18). Answer to ’is the PDF output of LaTeX code licensed under GPLv3 also covered?’. Open source stack exchange. https://opensource.stackexchange.com/questions/12738/is-the-pdf-output-of-latex-code-licensed-under-gplv3-also-covered/12739#12739
Matt. (2009, October 23). Why CC BY-SA is unsuitable. OpenStreetMap foundation. https://osmfoundation.org/w/index.php?title=Licence_and_Legal_FAQ/Why_CC_BY-SA_is_Unsuitable&oldid=3927
Möller, E. (2023, December 15). The case for free use: Reasons not to use a creative commons -NC license. https://freedomdefined.org/index.php?title=Licenses/NC&oldid=27327
Möller, E., Hill, B. M., Garlick, M., Stark, E., Knapp, J. A., Kadlec, P., Stallman, R., Lessig, L., & Beesley, A. (2015, February 17). Definition of free cultural works. https://freedomdefined.org/Definition/1.1
Morin, A., Urban, J., & Sliz, P. (2012). A quick guide to software licensing for the scientist-programmer. PLoS Computational Biology, 8(7), e1002598. https://doi.org/10.1371/journal.pcbi.1002598
Morrison, R. (2024, August 27). Which open data license? Open energy modelling initiative forum. https://forum.openmod.org/t/which-open-data-license/2611
Open Data Institute. (2015, September 22). What are the impacts of non-open licences? https://theodi.org/insights/guides/what-are-the-impacts-of-non-open-licences/
Open Knowledge Foundation. (2016). The open definition - open definition - defining open in open data, open content and open knowledge. https://opendefinition.org/
Open Source Initiative. (2007, March 22). The open source definition. https://opensource.org/osd
Poole, S. (2017, March 17). Use of CC BY 4.0 licensed data in OpenStreetMap. OpenStreetMap blog. https://blog.openstreetmap.org/2017/03/17/use-of-cc-by-data/
Rumsey, S., & Labastida, I. (2022, July 25). Exclusive licence to publish – now here’s a thing. Plan S. https://www.coalition-s.org/blog/exclusive-licence-to-publish-now-heres-a-thing/
SPARC. (2006). Author rights: Using the SPARC author addendum. https://sparcopen.org/our-work/author-rights/brochure-html/
Stallman, R. (2022a, January 22). For clarity’s sake, please don’t say “licensed under GNU GPL 2”! https://www.gnu.org/licenses/identify-licenses-clearly.html
Stallman, R. (2022b, September 6). Why programs must not limit the freedom to run them. https://www.gnu.org/philosophy/programs-must-not-limit-freedom-to-run.html
Suber, P. (2022). Publishing without exclusive rights. The Journal of Electronic Publishing, 25(1). https://doi.org/10.3998/jep.1869
UK Reproducibility Network, & Eglen, S. (2023). Rights retention strategy: A primer from UKRN. Open Science Framework. https://doi.org/10.31219/osf.io/2ajsg
Vézina, B. (2020, April 21). Why sharing academic publications under “no derivatives” licenses is misguided. Creative commons. https://creativecommons.org/2020/04/21/academic-publications-under-no-derivatives-licenses-is-misguided/
Villa, L. (2016a, September 21). Copyleft, attribution, and data: Other considerations. Luis villa: Open tech and policy. https://lu.is/2016/09/copyleft-attribution-and-data-other-considerations/
Villa, L. (2016b, September 26). Public licenses and data: So what to do instead? Luis villa: Open tech and policy. https://lu.is/2016/09/public-licenses-and-data-so-what-to-do-instead/
Wheeler, D. A. (2014, February 16). Make your open source software GPL-compatible. Or else. https://dwheeler.com/essays/gpl-compatible.html
Wickham, H., & Bryan, J. (2023). Licensing. In R packages (Second edition). O’Reilly Media. https://r-pkgs.org/license.html
Wikimedia Commons contributors. (2009, June 30). Licensing/justifications. Wikimedia commons. https://commons.wikimedia.org/w/index.php?title=Commons:Licensing/Justifications&oldid=917608913
Wikipedia contributors. (2024). GNU general public license – linking and derived works. In Wikipedia. Wikipedia, The Free Encyclopedia. https://en.wikipedia.org/w/index.php?title=GNU_General_Public_License&oldid=1244139410#Linking_and_derived_works
Footnotes
In fact, you should not create your own license if the purpose is to share your work freely and openly with others. By using a boilerplate license others know what to expect, also because they have been tested in courts.↩︎
In fact, the software licenses we recommend on this page have been specifically selected to provide an express grant of patent rights.↩︎
Copyleft licenses even require you to choose the same or a compatible license.↩︎
Of course, this is only a heuristic and there might be good reasons to deviate from community norms.↩︎
because (among other reasons) they explicitly disclaim any conveyance of patent rights↩︎
because they often require to display the full text of the license↩︎
that is, a license for writing, image, audio, and video↩︎
So called because copyleft licenses make content permanently free, thus turning copyright around: © → 🄯. In the realm of Creative Commons licenses, copyleft is also called share alike.↩︎
The exception is when the software is part of the output (see apsillers, 2019).↩︎
a variant of the AGPLv3 that does not cover software running as a service↩︎
Unless you are the author of the
apaquartoextension.↩︎In this particular case,
apaquartois licensed under CC0 1.0, which does not require you to retain a copyright notice. In fact, it’s an extremely permissive license.↩︎For licenses that require that modifications are indicated, this is an easy way to comply with them, although you do not need to provide your real name.↩︎