We're going to have structured UUIDs for books, in addition to a field for ISBNs, but that depth of analysis is pointless, when you also have the source in the same database because you can have AI analyze the text and add a label event telling _you_ what it is.