If you were building a second OpenRiC implementation, what would you need from us?

[johanpiet2]

Heratio is the only implementation of OpenRiC today. That makes "implementation-neutral" a claim, not a proof. We want a second implementation to exist — ideally someone else's — and we want to know what is missing to make that tractable.

If you run AtoM, Archivematica, a national-archive platform, or anything custom and have thought about serving RiC-CM/RiC-O: what would make it worth your time to emit the OpenRiC API? Questions we are sitting with:

• Is the mapping table (spec/mapping.html) the right shape, or would a worked example per source-schema (EAD / ISAD(G) / ISAAR-CPF) be more useful?
• Is the validator CLI enough to self-certify conformance, or do you want a hosted test harness?
• What is the minimum surface you would consider implementing first — just /records and /agents? Full L1? Subgraph?

Rough sketches, strong opinions, and "this is never going to happen unless you do X" are all welcome.