[Cbor] Re: Actual features used: EDN Usage in RFCs

[Rohan Mahy <[e-mail]>]

Correction:
The line about hex integers was supposed to say "found in at least 1 RFC".

For features where there seems to be no objection, I spent less time
counting the actual RFCs.
-rohan

On Sat, 11 Jul 2026, 08:59 Rohan Mahy, <[e-mail]> wrote:

> Hi All,
>
> There are 53 RFCs that reference RFC8610
> <https://datatracker.ietf.org/doc/rfc8610/referencedby/> (35 normative
> and 18 informative).
>
> I manually perused *all of them* for EDN looking for usage of specific
> features. Here are my findings:
>
> *RFC 8949 diagnostic notation features*
> indefinite length arrays, maps — one instance of indefinite length array
> found in RFC 9891.
> indefinite length strings — no instances found
> b32, h32 — no instances found
>
> *RFC 8610 EDN features*
> whitespace in h’’ and b64’’ (§G.1) — widely used
> text in byte string notation (§G.2) — found in RFC 9052 (kid in protected
> header), perhaps in others.
> embedded CBOR (§G.3) — used in several COSE documents: RFC 9052, RFC 9338;
> plus RFC 9594, RFC 9891
> concatenated strings (§G.4) — no instances found
> hexadecimal integers (§G.5) — found in 1 RFC
> octal and binary integers (§G.5) — no instances found
> hexfloat notation (§G.5) — no instances found
> /comments/ outside of quoted strings (§G.6) — universally used
> /comments/ inside h’’ (§G.6) — one use (/x/) in Figure 4 of RFC 9528
> /comments/ inside b64’’ (§G.6) — doesn’t work
> # comments (alluded to in §G.6) — RFC 9886
>
> *Features in draft-ietf-cbor-edn-literals*
> ellipsis (…) — used in RFC 9770, RFC 9880, RFC 9942, RFC 9943†
> // comments — no instances found
> extra trailing comma — one example found in RFC 9526 (possibly
> unintentional)
>
> †Note: RFC 9443 contains a travesty of EDN, with unquoted text strings and
> nill instead of null
>
> *RFC 8742* introduced CBOR/ EDN sequences
>
> Thanks,
> -rohan
>
>