On 13 Jun 2021, at 12:16, Lassi Kortela <xxxxxx@lassi.io> wrote:

>> "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.

A solution that just occurred to me for human-readability: use URIs instead, especially tag URIs. <https://www.taguri.org/>

For SRFIs, a single tag URI namespace could be established, say <tag:srfi.schemers.org,2000:>. (The choice of 2000 is simply because it’s a nice-looking round number. The tag URI specification requires merely that the person or organization minting the tag (in this case, we’ll claim for the sake of argument that it’s the SRFI Editors) controlled the given domain name or email address on the date given in the tag URI. SRFIs have existed since 1998, so we can safely use 2000 for all SRFIs. Or we might just choose the year/date the SRFI was submitted. Whatever.)

It would then be easy to say (following Marc’s example in the email after the one this one is in-reply-to):

(define-record-type fxmapping
  (sealed #t) (opaque #t)
  (nongenerative |tag:srfi.schemers.org,2000:146:fxmapping|)
  (fields ...)
  ...)

This is human readable, can be used to uniquely identify record types for serialization/deserialization (say, in John Cowan's Twinjo proposal), and unique. The authority for the ID is also clear, which is not the case for UUIDs.

Daphne