The goal of DFIG (Data Fabric Interest Group) is to identify a set of Common Components and define their characteristics and services that can be used across boundaries in such a way that they can be combined to solve a variety of data scenarios such as replicating data in federations, developing virtual research environments, supporting the exchange of data and so forth. DFIG has had a number of productive discussions and sessions over the last 18 months, and has developed several draft documents identifying Common Components, as well as identifying the need to develop more detailed ideas on principles, guidelines, and recommended practices. Detailed ideas on principles, guidelines, and recommendations, do not only emerge from RDA Working Groups (WG Recommendations), RDA Interest Groups, RDA discussions, but also from other initiatives and projects. Bringing them together to extract coherent messages is now urgent since these will help reduce the solution space.
However, since DFIG is an Interest Group rather than a Working Group, RDA does not currently have an explicit process on how such recommendations can be included or incorporated into official RDA Outputs. The following process is a suggestion on how the discussion and DFIG work efforts in RDA may lead to widely agreed principles, guidelines and recommended practices, which in turn, would be a new kind of output for RDA. DFIG sees itself in a position to kick off this process, but limiting the scope of work to the issues around data fabrics. The following procedure is recommended:
- Identify a number of Bundles or other work products that are obviously related to key components (first suggestions can be seen on the DFIG guidelines pages https://rd-alliance.org/group/data-fabric-ig/wiki/recommendations.html):
- Define characteristics and requirements for each Bundle and Work Product (some of which has already been started and which needs to be extended)
- Establish DFIG Working Group Charrettes that will take responsibility to manage the process and development for each Bundle and Work Product including involvement of external experts
- Develop draft set of principles, guidelines, and recommendations for each Bundle and Work Product by including also recommendations from other well-known initiatives (3-4 months)
- Submit the draft set of principles, guidelines, and recommendations for each Bundle and Work Product to the larger RDA Community and TAB as an RFC (Request for Comment) asking for community review, input and comment (3-4 months)
- Working Group Charrette** then addresses the issues and comments from the RFC process and develops a final set of principles, guidelines, and recommendations to be presented at the next RDA Plenary (3-4 months).
- Once the final set of principles, guidelines, and recommendations of each Bundle and Work Product has been adopted by RDA, they will be made available to the public.
- Each set of final principles, guidelines, and recommendations will be made available in a common format on the RDA website (and other sites as appropriate). The common format (still to be developed) will include details on authorship, references and contacts for additional information.
It is necessary to involve as many data professionals in such a discussion. Therefore, in Europe the GEDE interaction was started involving experts from 45 of the most active research infrastructure projects (https://rd-alliance.org/groups/gede-group-european-data-experts-rda). It will be one of the challenges to closely synchronize these discussion strands.
** A charrette is an intensive planning or set of work sessions which provides a forum where people can collaborate and share ideas to develop a new vision or approach. It also offers the unique advantage of allowing everyone who participates to be a mutual author.
A recent comparison between RDA DFT and FAIR statements is attached.