[Yot] Mitigating a squatted code point (CBOR Tag 42)

[Carsten Bormann <[e-mail]>]

Recently, for yang-cbor, Tag number 42 was allocated.
https://www.iana.org/assignments/cbor-tags/cbor-tags.xhtml

Unexpectedly, this got the attention of the IPLD/IPFS people:
https://github.com/core-wg/yang-cbor/issues/13

It looks like they had been squatting on Tag number 42 for a while and were caught by surprise that this now was allocated.

Normally, we don’t encourage squatting, so the usual answer would be “your own fault”.
However, in this case, a mitigation seems easy:

* Move YANG Schema Item iDentifier (sid) from 42 to 47.
* Get a reasonable specification out if the IPLD/IPFS people that we can use for registering their tag 42 (see the above github issue for 80 % of this, which just needs to be stitched together properly).

IPLD/IPFS would have a hard time fixing their tag number, as that is embedded in Merkle trees already out there.  This makes it more likely we would see dual use of Tag 42 both for YANG SID numbers and for IPLD CID byte strings.  I think we should avoid this if it is cheap for us to do this.  The YANG-CBOR people tell us that right now the damage from going from 42 to 47 would be very limited.

I’d like to discuss this at the CBOR Virtual Interim in 40 minutes — both the specific case, and what we could be doing to make further occurrences of this less likely.

Grüße, Carsten