How can an existing repository join DataONE?

[mbjones]

Several groups have expressed interest in what mechanisms are available for joining DataONE as a member repository. While the website has a lot of information, is there a more succinct overview of approaches to joining?

[mbjones]

There are several different paths for joining DataONE, each of which involves making your data available to our federated search subsystems and has a number of benefits for your repository. The easiest of these only provide read-only APIs that are compatible with DataONE, whereas the latter solutions in this list provide more comprehensive APIs for interacting with your repository. These are:

1. publish SOSO-compliant schema.org metadata on your dataset landing pages
   • pros: fast and lightweight; - cons: minimal metadata, not as rich as domain standards for search
2. deploy one of our software stacks that implement the DataONE API (Metacat, GMN)
   • pros: supports domain metadata, both read and write APIs available; cons: more integration/development work
3. implement the DataONE API in your own repository software stack
   • pros: support domain metadata, both read and write APIs available; cons: more integration/development work
4. use the DataONE Hosted Repository service to host your repository for an annual fee on our servers
   • pros: no infrastructure to maintain or develop, detailed metadata, rich reporting services; cons: external service with a fee

In addition, projects that have only a modest amount of data, or who might not have funding to sustain a longer term repository, may instead be best served by contributing data to an existing repository (e.g., KNB, Dryad, EDI) and create a custom DataONE Data Portal to represent your collection of data. This gives you a place to showcase your data and a project web site without the overhead of maintaining a repository.