As if the current political climate isn't going to result in the sabotage of scientific infrastructure if some state actor decides that it could provide some economic or military advantage. (hello three body problem)
DOIs should have been hashes, that would have been cheaper, more resilient, and more covenient. But sadly librarians tend to re-build paper workflows digitally instead of building digitally native infrastructure.
Blockchain would be fine as a timestamp service to replace publishers, although a consensus based system hosted by the worlds libraries would also be fine for that purpose and require a lot less machinery.
Users would generate and centrally register or receive a generated W3C DID keypair with which to sign their ScholarlyArticles and peer review CreativeWorks.
W3D DID Decentralized Identifiers solve for what DOI and ORCID solve for without requiring a central registry.
W3C PROV is for describing provenance. PROV RDF can be signed with a DID sk.
PDFs can be signed with their own digital signature scheme, but there's no good way to publish Linked Data in a PDF (prepared as a LaTeX manuscript for example).
Bibliographic and experimental control metadata is only so useful in assuring provenance and authenticity of article and data and peer reviews which legitimize.
From https://www.hackerneue.com/item?id=28382186 :
>> JOSS (Journal of Open Source Software) has managed to get articles indexed by Google Scholar [rescience_gscholar]. They publish their costs [joss_costs]: $275 Crossref membership, DOIs: $1/paper:
I’m no blockchain evangelist in its current state of “value” but this seems like a great test case for resolving the academic or legitimate origin of published material.