The following principles and design assumptions underlie the technological recommendations for the next generation of repositories:
Distribution of control – Distributed control, or governance, of scholarly resources (preprints, post-prints, research data, supporting software, etc.) and scholarly infrastructures is an important principle which underpins this work. Without this, a small number of actors can gain too much control and can establish a quasi-monopolistic position. Distributed networks are more sustainable and at less risk to buy-out or failure.
Inclusiveness and diversity – Different institutions and regions have unique and particular needs and contexts (e.g diverse language, policies and priorities). A distributed network of repositories will aim to reflect and be responsive to the different needs and contexts of different regions, disciplines and countries.
Public good – The technologies, architectures and protocols adopted in the context of the global network for repositories will be available to everyone, using global standards when that are available.
Intelligent openness and accessibility – Scholarly resources, will be made openly available and in accessible formats, whenever possible, in order increase their value and maximize their re-use for the benefit for scholarship and society.
Sustainability – Institutions and research organizations will be major participants in the global network, contributing to the long term sustainability of resources.
Interoperability – Repositories will adopt common behaviours, functionalities and standards ensuring interoperability across institutions and enabling them to engage in a common way with external service providers
Focus on the resources themselves, not just associated metadata – For historical reasons, technical solutions have focused on metadata that describes scholarly resources instead of on the resources themselves. By considering both the scholarly resource and its metadata as web resources identified by distinct URIs, they can be treated on equal footing and can be appropriately interlinked.
Pragmatism – Given the choice, we tend to favour the simpler approach. Where possible, we choose technologies, solutions and paradigms which are already widely deployed. In practical terms, this means that we favour using standard Web technologies wherever possible.
Evolution, not revolution – We prefer to evolve solutions, adjusting existing software and systems where possible, to better exploit the ubiquitous Web environment within which they are situated.
Convention over configuration – Our preference is to adopt widely recognised conventions and standards, and encouraging everyone to use these where possible, rather than accommodating richer, more complex and varied approaches. As a corollary to this, we believe new standards should be introduced only when concrete and pragmatic needs arise, with the intention of keeping constraints to a minimum so that those implementing our systems can readily understand the constraints under which they must operate.
Engage with users where they are – Instead of always asking users to leave their environment and engage with one of our systems, integrate tools into the environments and systems where they are already engaged.