> It has already been argued several times that all these attempts so far
> are logically flawed and don't imply what they should.

Is there a succint statement of the flaw?

> Luckily, there is a solution, which can be implemented right now and
> which is future-proof, namely using uids, through which we can exactly
> express what we want.
>
> So the above sentence cited from SRFI 224 would become:
>
> "Fxmappings are instances of a sealed, opaque, nongenerative record type
> with uid
> fxmapping-2bf340e5-304e-436e-8478-926c7040f3f."

UUIDs are trivially unique, but suboptimal to have in contexts where
they are shown to people. Would this be something along the lines of:

(define-record-type fxmapping-2bf340e5-304e-436e-8478-926c7040f3f ...)

Another alternative is to allocate human-readable IDs and track them in
a table in <https://registry.scheme.org/>. Here too we'd be free to add
identifiers at SRFI publication time, RnRS publication time, or any
other time.