1.1. One of the best licenses you can use for your repository is a CC BY 4.0 license, specifying that “unless otherwise noted, this repository is under a CC BY 4.0 license”.

We recommend using a CC BY 4.0 license as a repository license for the following reasons:

• Creative Commons licenses are internationally recognised, well-established, and both human-readable and machine-readable;
• CC BY 4.0 licenses meet the definition of “open access” as defined in the Budapest, Bethesda, and Berlin declarations on open access;
• CC BY 4.0 is one of the most compatible licenses for interoperability purposes.

1.1.1. If you follow point 1.1 it means that the license applies to all “works” or other subject matter in the repository.

This includes:

• The repository as a copyright protected database (in case it qualifies);
• The repository as a suis generis database right protected database (in case it qualifies);
• The elements composing the database which can be:
  • Not protected, such as a database of temperature measurements. In this case, as these data are not protected in themselves you don’t need a license. CC licenses are written in a way that you only have to accept them if you need permission to use something.
  • Protected (e.g. a database of journal articles)

For University repositories, it is likely that several of these elements co-exist, but it could also well be that the repository is not a protected database. In either case CC licenses are a good choice because (avoiding technicalities) they only regulate the use if that use requires a permission. Under this point of view it could be said that CC licenses are self-contained to when permission is necessary.

1.1.2. However, this could become problematic, when, as in the case of University repositories, the owner of the repository (the University) and the owner of the journal article (the author unless they transferred the copyright) are different people.

Therefore, by using the recommended “unless otherwise noted” wording, you clarify that the elements that belong to third parties (e.g. journal articles) are distributed under their own license terms (which as you will see later, ideally is a CC0 or a CC BY).

It is important to license the repository as a database under an open access compliant license. This is because when a user uses aggregated data (such as in data analytics, text and data mining, etc.) in order to crawl, scrape or analyse the database, authorisation (e.g. a license or an exception if it exists) is often necessary. But if you have applied a CC BY to your repository this is already taken care of!

1.2. The CC BY 4.0 license should be incorporated into the terms of service of the repository.

1.2.1. The CC BY 4.0 license exists as a separate legal document from the terms of service.

As such, it must be incorporated by reference into the contractual, and broader terms of service which govern all uses of the repository. Creative Commons provide guidance on how to incorporate the CC BY 4.0 license into the repository terms of use.

2.1. You should provide metadata in order to enhance discoverability of your resources.

2.1.1. Providing machine-readable bibliographic metadata is a requirement for projects which are funded under Horizon 2020.

2.2. Metadata often are not protected as such because they are factual information, thus not original or not substantial.

However, in certain cases, complex and elaborate metadata could perhaps be protected. For the avoidance of doubt, apply a CC0 to your metadata. In this way, in those cases when a right exists, you are waiving it and allowing other people to reuse the metadata information.

CC BY should be avoided unless you know exactly what it entails.

2.2.1. In those few cases when metadata can be considered original works, thus protected by copyright, they will enjoy both economic and moral rights.

Moral rights are recognised in most countries (but with important exceptions, such as the US), and may be unwaivable.

This should not represent an issue in the case of CC0, as the waiver clarifies that it only waives the rights as long as this is permitted under applicable law. So if you enjoy unwaivable moral rights, CC0 will not affect your moral rights.

2.3. For metadata to be used meaningfully, it must be standardised to optimise machine-reading (a requirement of H2020 projects). A commonly used format in libraries and cultural heritage institutions is Dublin Core.

3.1. In point 1 you have applied a license to your repository, and to its content “unless otherwise noted”. Now let’s take a look at the “unless otherwise noted” part.

As a repository manager, you (or the University) usually don’t own the copyright in the articles uploaded (unless you have written them).

Therefore, the repository has to implement a license selection procedure that allows the uploader (author or rightsholder) to choose the proper license. As further detailed in point 6, this process should offer a number of choices to the author, but it should be the author who ultimately decides what license to use.

Nevertheless, in order to help researchers to make the right choice, you can offer and implement some guidance that will help researchers to make the right choice and to adhere to Open Science principles.

In points 4 and 5 you will see what choices are recommended for a) data and databases; and b) articles.

3.2. We recommend a CC BY 4.0 license in respect of the content of the repository. This is detailed further in Point 5. This may not be appropriate for data or datasets as detailed in Point 4.

3.3. Creative Commons licenses are not appropriate for software.

Instead, we would recommend that a GNU GPL v3.0, BSD/Apache style license is applied. These are some of the most well-established public licenses for free software, and are based on the “copyleft” concept for open source software. They are also highly interoperable with other licenses.

3.4. Whilst Creative Commons licenses may apply to both digital and non-digital content, this guide currently provides advice only in respect of fully digital repositories.

Libraries and cultural heritage institutions may need to audit their non-digitised resources, and check for complex or multilayered content (e.g. multiple authors, orphan works etc.)

3.4.1. If you need to digitise materials for use on an online repository, note that there may be special considerations for traditional knowledge works.

4.1. Data and datasets and databases should be offered without restrictions on use, meaning under a CC0.

Remember that a database can be protected by copyright (database structure) and/or SGDR (substantial investment in obtaining verifying and presenting data), without prejudice to any copyright or other rights in the underlying material.

Applying a CC0 to a database means that if any rights exist they are waived, if they don’t exist CC0 does not create any obligation. If waiver is not possible then CC0 operates as a waiver or as a license to the same effect within the limits of applicable law.

4.1.1. The advantages of making data available without restrictions include:

• Greater availability and accessibility of publicly funded scientific research outputs;
• Possibility for rigorous peer-review processes;
• Greater reproducibility and transparency of scientific works;
• Greater impact of scientific research.

4.2. Where the uploader is concerned with regards attribution they can ‘kindly request’ to be attributed, rather than using the more legally restrictive CC-BY license.

This is not legally binding although follows standard scholarly practice in crediting researchers for their work.

4.2.1. When ‘kindly requesting’ attribution of a work, the uploader should be advised to offer a citation which can be easily copy and pasted by subsequent users.

4.2.2. Tools such as the Creative Commons ‘Open Attribute’ tool are available to assist with ensuring adequate attribution.

4.3. Where there are concerns with regards privacy issues or data protection, these should be dealt with under the relevant legislation or ethics policies.

4.3.1. Both the BMC consultation and the PLOS data policy address these issues.

Suggest which license should be chosen to meet OS requirements, but let the uploader choose.

5.1. Give the uploader the possibility to choose the license.

You can indicate which licenses are better for OA/OS, but don’t choose for them.

5.2. Repositories may play an important role in educating uploaders with regards open licensing.

5.2.1. The importance of making their work open access, should be explained to uploaders prior to upload.

The benefits of open access work include: Researchers and their institutions benefit from having a wider audience Open access allows use of text and data mining tools, without legal barriers. Funders receive a greater return on their investment when results of research can be utilised by more people and at an earlier date.

5.2.2. The Sparc Europe resource offers a useful summary of the benefits of making work open access.

6.1. Uploaders should be offered all possible guidance and explanation with regards the various licences open to them, and the degree to which these are compatible with open access principles.

6.1.1. This can be done by incorporating some form of ‘Licence Selector’ tool into the upload process. The tools featured here offer examples of how this can be achieved.

6.2. CC-BY 4.0 may be considered as a default standard licence, except in the case of data and datasets.

However any default licence provided should always be accompanied by a selection of alternative licences and comprehensive explanations about the function of each.

The CC-BY 4.0 licence is often considered the ‘gold standard’ open access licence, since it is the least restrictive and allows people to use the licensed content as they choose, provided attribution is provided, and is fully OA compliant. As a note of caution, however, it should always be the uploader who makes the final licence selection.

6.3. Where uploaders select a licence which is less compatible with open access/science requirements, this should be made clear to them.

This is particularly relevant where uploaders choose Creative Commons licences with NC (non-commercial) or ND (no derivatives) conditions. These licences have been described by Creative Commons as failing to promote ‘free culture’.

6.4. In the case of software, application of a GNU GPL or BSD/ Apache style licence is recommended.

These licences are:

• The most well-established public licences for free software, and
• The most interoperable licence both in terms of general use and for TDM purposes.

6.5. In the case of public sector information, application of an Open Government Licence is mandated by the UK Government Licensing Framework (UKGLF) for all public sector information.

6.6. Ultimately, however, the final decision with regards which licence is applied should rest with the uploader.

6.7. The resources featured here offer comprehensive discussion regarding the benefits of open access principles, and provide an example of how these might be expressed to uploaders.

6.8. Account must be taken of any external limitations on the uploader’s choice of licence.

This may be as a result of funding body stipulations or publishers’ requirements.

6.8.1. The resources featured here, including the European Commission H2020 guidance, provide an example of possible funding body stipulations, with regards making work open access and how this should be done.