apologies, I hit a thingy too early...
it has links to modern solutions too, especially an archival specific scheme. With PURL came purl resolver systems, ensuring redirections at the server level with built in notions of content negotiation. In some domains people have created their own PURL servers, for example in the biomedical domain:
https://github.com/OBOFoundry/purl.obolibrary.org On top of this there are approaches to publish resolution for namespaces
https://identifiers.org/ This is all a big issue of community governance.
These are outside the archival world yet the problem exists in every domain; every authority designing schemes for this is always going through the hazing of finding a syntax for identifiers, and then making them resolvable as URIs. One extreme solution is to simply create globally unique ids (modulo namespace), a good example is:
https://permid.org/1-5035507360 in the finance world. This is extreme -- and arguably good -- because it treats URIs as things for machines, On the other hand of the spectrum, the UK government issued a while back some guidance on building identifiers:
https://www.gov.uk/government/publications/designing-uri-sets-for-the-uk-public-sector It's the other hand of the spectrum because a lot of semantics and case analysis is built into the URIs.
In the archival world, in the UK, there was a really neat project at the National Archives a few years ago. Maybe someone on this list knows the status, there's a small series of blogs that remain most of the discussions are about IDs and you may find them useful:
https://catalogueprojects.medium.com/I had never come across PID grafs -- the idea of a graph of identifiers does seem in essence to be the idea of Linked (Open) Data and in relation to this, one of the basic FAIR principles
https://www.go-fair.org/fair-principles/f1-meta-data-assigned-globally-unique-persistent-identifiers/Again, not totally sure this is what you are asking about or if it helps, apologies if I misread your question.
Pierre