February 6, 2019 – This is a slightly updated version of the COAR initial feedback published on December 13, 2018.

The Confederation of Open Access Repositories (COAR) reiterates our strong support for the goal of Plan S to achieve “immediate Open Access to all scholarly publications from research” and we appreciate the opportunity to provide feedback on the guidance on the implementation of the Plan.

COAR is an international association with members from over 140 organizations on 5 continents. COAR promotes global interoperability of repositories, helps to build capacity across regions, and offers an international perspective related to repositories, open access and open science with the aim of creating a global ecosystem for research dissemination.

General comments

COAR agrees with many of implementation guidelines outlined in Plan S, but has some concerns that the technical requirements are too high and will result in only large, well-funded publishers and repositories to become compliant. The unintended consequence is that Plan S would further entrench the position and control of a small number of players in the scholarly communication system, hampering innovation and our ability to control costs.

10. Deposition of Scholarly Content in Open Access Repositories

We recognize and agree with the aim of transforming the publishing industry, however to truly improve and transform the system there needs to be a multipronged approach, with a number of actions undertaken concurrently.

Repositories are important mechanisms for advancing innovation in research communications, as detailed in the COAR Next Generation Repositories report, and should be considered on equal footing in terms of their role in providing open access, while also recognized for their contribution towards a more sustainable system that can support management and access to a variety of research outputs.

To avoid further enclosure by a small number of large publishers and the negative consequences that would come with this (cost inflation, misaligned incentives, and publication barriers), we recommend that Plan S actively foster and explicitly acknowledge the importance of a diversity of publication models and non-APC based service, such as the vision outlined in the COAR next generation repository work. This model offers a sustainable alternative to traditional publishing by building peer review layers on top of the global repository network, distributing the services and costs across institutions.

Specific comments related to implementation guidance

10.2 Requirements for Plan S compliant Open Access repositories

COAR and others in the repository community have concerns related to several of the requirements for repositories, a number of which we argue are not necessary and will create artificial barriers to the participation of universities and other research organizations in the scholarly communication system.

While some of these recommendations may be ‘nice to have’, they are not prerequisites for robust and interoperable repository services. Instead they could result in driving repository functionality in the wrong direction, create too high of a bar for less resourced institutions, and further centralize research infrastructures and services because they cannot be adopted, leading to a replication of the existing inequalities in the scholarly communication system.

We urge authors of Plan S to remove or reword some of the requirements, and move others into a “Recommended additional criteria” section, and establish a transition period for repositories to become compliant with some of the requirements.

Below you will find COAR’s detailed feedback related to the “Requirements for Plan S compliant Open Access repositories”. These recommendations are based on widespread expertise of the COAR community and input from the COAR Next Generation Repositories Editorial Group, representing some of the foremost experts in repository technologies. The technical recommendations are based on a 1.5 year in-depth examination of the future behaviours and functionalities required for repositories undertaken by COAR and the Next Generation Repositories Working Group. We would be happy to explain these comments further and would be pleased to work with cOAlition S to finalize the recommendations. This will ensure that the implementation of Plan S for repositories is achievable and supports our common goal of accelerating open access.

Automated manuscript ingest facility

Apart from the SWORD protocol and a few regional router services, very few repositories currently have an automated ingest process for manuscripts. Indeed, most repositories employ a human-mediated deposit approach which involves librarians and/or repository managers recruiting and depositing content into the repository on behalf of researchers. This approach ensures there is some quality control of metadata and that the appropriate version of the article is being deposited. We do not believe that an automated manuscript ingest functionality is needed for repositories to comply with the intent of Plan S, which is to provide immediate access to the Author’s Accepted Manuscript. Furthermore, there are numerous different ways which “automated ingest” could be interpreted and implemented by different repositories. Unless there is a common, standardized mechanism required by all repositories, this requirement will not be effective in aiding with automated population of repositories. With further explanation and specificity, this functionality could be included in a “recommended additional criteria” section.

Full text stored in XML in JATS standard (or equivalent)

The adoption of XML is extremely resource intensive. While we agree that full text articles need to be available for TDM, this does not equate to full text needing to be pre-processed through XML-JATS. This requirement supposes that text mining is part of the repository system, but the preferred approach to TDM for repository content is that external services aggregate and convert resources into text-minable format. It is sufficient to require that the content in repositories are available/open for automated discovery of full text (e.g. pdf or word). To that end, Plan S should recommend the Signposting protocol as the standard convention for making full text easily discoverable and retrievable by external services and processes.

Quality assured metadata in standard interoperable format, including information on the DOI of the original publication, on the version deposited (AAM/VoR), on the open access status and the license of the deposited version.

We agree with this requirement and have nothing to add.

Open API to allow others (including machines) to access the content

We agree that repositories should have an open API, however with hundreds of different open APIs, it is important to specify which APIs should be adopted in order to ensure machine interoperability and that service providers can develop cross-repository services. Plan S should recommend and eventually require, after a transition period, the adoption of Resource Sync as an API. ResourceSync is a modern successor to OAI-PMH. It is a specification based on Sitemaps that can be used by repository managers to provide information that allows third-party systems to remain in sync with the resources in their repository as they evolve, i.e. are created, updated, deleted. ResourceSync has been found to improve aggregation services, is scalable and is being adopted by service providers.

QA process to integrate full text with core abstract and indexing services

Many repositories are already connected and indexed through regional, international or national harvesting services (e.g. OpenAIRE, BASE, CORE, LA Referencia, etc.) We recommend that the requirement that repositories integrate into abstract and indexing services be replaced by a requirement that repositories be harvested by a national, regional or international aggregator.

Continuous availability

We agree with this requirement and have nothing to add.


The aim of this requirement is to ensure that the repository service is contactable and will respond to issues or queries from users. However, this seems unnecessarily onerous for the repositories. Many complex websites and services do not have helpdesks. We recommend that instead of a helpdesk, Plan S require a mechanism for users to get help or support for using the repository, which could be as simple as an email address or a simple online form. This is in line with existing practice for mainstream web services.

We would be pleased to provide further explanation, and happy to work with cOAlition S members to ensure Plan S can be successfully implemented. For more information, please contact Kathleen Shearer, Executive Director: kathleen.shearer@coar-repositories.org

Download a PDF version of the response.