[rpp] Fwd: [art] JSON Schema work

[Pawel Kowalik <[e-mail]>]

This might be relevant to RPP work.


-------- Forwarded Message --------
Subject: 	[art] JSON Schema work
Date: 	Fri, 27 Feb 2026 10:23:04 -0800
From: 	Lisa Dusseault <[e-mail]>
To: 	[e-mail]




Hi ART,

Along with much help from several experts in the JSON Schema community, 
I've prepared a draft that I think would be a good basis for a consensus 
process and open standardization of JSON Schema.

https://datatracker.ietf.org/doc/html/draft-dusseault-json-schema 
<https://datatracker.ietf.org/doc/html/draft-dusseault-json-schema>

I wanted to start from the same functionality the community has been 
using for the last five years (if you've seen the "2020-12" tag in 
schema docs, that's the most recent, common, stable spec).  Moving 
forward from there, anything we add or delete will be visible in 
changelogs and diffs.  But before that, this offers a large-scale 
re-organization of the document (made for some terrible diffs I tell ya) 
, with I feel some progress towards improved readability.
* Terminology section is only terminology, not specification text
* Terms "instance" and "input" have been narrowed and used consistently
* Important sections like the definition of a keyword are never deeper 
than H3 - easier to find in TOC
* Core functionality is described first, then implementation 
requirements especially around loading and following references, then 
extensibility functionality

I welcome feedback on high-level direction: what is included in the spec 
and what is not,  intended status, and WG/charter plans. The highest 
order bit within the spec is probably the knot of related decisions 
around what value goes in "$schema", when that changes, and what that 
means for implementations supporting a given spec version and 
encountering keywords not in that spec version.  The JSON Schema 
community (there's a slack and forums linked off json-schema.org 
<http://json-schema.org>) has a bunch of experience and different plans 
for this so determining consensus will be important.

Lisa