Thanks @bsterner. To be fair, the use of the term “communities of practice” here is probably rather sloppy - something for which I take responsibility. I see this mainly as a recognition that we need to find a model that avoids the possible pitfalls you identify in your post on Topic 4.4 Governance Arrangements.
I find it easiest to approach this from the aspect of GBIF’s role. I realise that this reflects my own biases and may make me blind to some aspects, but I do think it’s helpful.
GBIF has a role both as an aggregation point for information on biodiversity and as a facilitator for the community of biodiversity stakeholders represented by the GBIF Participants (although of course GBIF aims to be a global public good and serve other stakeholders as well).
The role as an aggregation point addresses the fact that, in a complex field such as biodiversity research, there are some use cases that cannot be properly satisfied or that are seriously compromised if there is no efficient way to search for and relevant information regardless of where it is held. GBIF’s role in trying to facilitate discovery of all spatiotemporal evidence for the occurrence of species is an example of this. A comprehensive aggregation point can vastly simplify data access for many purposes. In the same way, a comprehensive look-up service for discovering “all” natural history collections plays a role and can answer questions that a multitude of partial collections cannot address. Some of these questions are of course important for GBIF for its other data management activities, but the ability to slice this information in many different ways and the ability to offer resolution services around a comprehensive resource will support many platforms and researchers.
However, GBIF does not itself need to support or take ownership for all of the domain-specific or national requirements that are important to IH, ALA, GGBN, CETAF and other communities. All of these have needs that probably include the common elements needed by GBIF but also have additional needs and may have learned lessons of their own that do not apply to the other communities.
GBIF’s second role as a facilitator for the whole GBIF community means that it has a strong interest in helping each of these communities achieve their goals. This means that, as GBIF works out how to achieve the data integration it requires around collection information, it will aim as far as possible to do so in ways that also benefit these other communities.
In some cases, the balance of interests is best met by GBIF assuming responsibilities that are not strictly part of its own mandate or mission - the adoption of the whole of GRSciColl (rather than just GRBio) is an example of this. In the same way, this consultation has partly been to understand in which ways the GBIF/GRSciColl core should be extended to take responsibility for what to GBIF are peripheral needs. If this means that the resulting “central catalogue” simplifies life for one or more communities, then it may make sense to aim for a unified solution/implementation.
In other cases, for GBIF to try to solve the problems of other communities would be inappropriate, undercutting the ability of these communities to remain sustainable and to serve their own constituents. This means that the overall approach to building a global catalogue will almost certainly at some level be a mosaic that federates across different communities and services, each with their own focus and strengths.
You are correct that IH, CETAF, ALA, etc. are themselves organisations rather than strictly “communities of practice” but each of these to a variable extent is the expression of a sense of community bringing together a subset of collection stakeholders. Each of these has its own vision, goals and needs. I hope that we can use a model of overlapping communities to strengthen all of these, while bringing their core information together as a comprehensive global catalogue (whether genuinely unitary or more accurately federated) to support the services that require such a centralised endpoint.