moq-streaming-format: declare WAI as a MoQ streaming format #85

Merged

dcharlot merged 1 commit from wai/moq-streaming-format into main 2026-09-10 12:21:10 -04:00

Conversation 0 Commits 1 Files changed 1 +139

dcharlot commented 2026-09-07 04:20:21 -04:00

Owner

Copy link

WAI already references MoQ for confidentiality (MLS + SFrame) — a binding that treats MoQ as a pipe. MoQ is not a pipe. It is a two-layer design, and the upper layer has an open slot exactly where a container of capability-dispatched objects belongs.

• MOQT moves objects in groups on tracks, and does not know what an object means.
• MSF (draft-ietf-moq-msf) says what objects mean — fragments a stream into objects, declares what a publisher emits in a catalog, sets prioritisation.

The alignment is structural, not aspirational: MSF's unit is an object in a group; WAI envelopes are already objects; jwp-receipts already commits per object and signs per group. Neither layer has to change to agree.

The slot is open, and cheaply

MSF's packaging field has a fixed value set and no central registry — which reads closed. draft-ietf-moq-cmsf shows it isn't: a sibling I-D declaring packaging value cmaf, stating it extends the allowed packaging values defined in MSF, inheriting MSF wholesale, adding two track-level catalog fields, and registering nothing — "this document has no IANA actions." LOCMAF is a second instance. A sibling draft, no registry action, no permission sought.

What CMSF did not take is the per-object slot. MSF §12 defines properties at Track and Object scope with an IANA type registry holding essentially one entry (MSF_COMPRESSION); CMSF defines no per-object property. §5 takes it with WAI_RECEIPT — content hash, joules, acquisition class, group-receipt hash — the shape jwp-receipts already has.

The load-bearing rule

§3: the object boundary MUST be the envelope boundary. That is what makes MOQT's object identity and the receipt's per-object commitment the same unit. If they diverge, every guarantee that composes them is void.

§4 makes the catalog do work: wai_capabilities MUST be exhaustive, a sink MUST NOT subscribe assuming an undeclared capability won't appear, and wai_determinism is the weakest tier on the track — declaring the strongest would advertise a guarantee most objects don't meet.

What it does not do

No MOQT wire encoding, no replacement for CMSF or LOC, no making MoQ WAI's transport, and no pinning of a draft revision. MOQT is pre-Last-Call; where it and this text disagree, MOQT wins and this text is what changes.

Ingest queue item 13b (#80 §4.4). Sourced from direct reads of draft-ietf-moq-msf and draft-ietf-moq-cmsf-00.

🤖 Generated with Claude Code

WAI already references MoQ for confidentiality (MLS + SFrame) — a binding that treats MoQ as a **pipe**. MoQ is not a pipe. It is a two-layer design, and the upper layer has an open slot exactly where a container of capability-dispatched objects belongs. - **MOQT** moves objects in groups on tracks, and does not know what an object means. - **MSF** (`draft-ietf-moq-msf`) says what objects mean — fragments a stream into objects, declares what a publisher emits in a **catalog**, sets prioritisation. The alignment is structural, not aspirational: MSF's unit is an object in a group; WAI envelopes are already objects; `jwp-receipts` already commits per object and signs per group. **Neither layer has to change to agree.** ### The slot is open, and cheaply MSF's `packaging` field has a fixed value set and no central registry — which reads closed. `draft-ietf-moq-cmsf` shows it isn't: a sibling I-D declaring packaging value `cmaf`, stating it *extends the allowed packaging values defined in MSF*, inheriting MSF wholesale, adding two track-level catalog fields, and registering nothing — **"this document has no IANA actions."** LOCMAF is a second instance. A sibling draft, no registry action, no permission sought. **What CMSF did not take is the per-object slot.** MSF §12 defines properties at Track *and* Object scope with an IANA type registry holding essentially one entry (`MSF_COMPRESSION`); CMSF defines no per-object property. §5 takes it with `WAI_RECEIPT` — content hash, joules, acquisition class, group-receipt hash — the shape `jwp-receipts` already has. ### The load-bearing rule §3: **the object boundary MUST be the envelope boundary.** That is what makes MOQT's object identity and the receipt's per-object commitment the same unit. If they diverge, every guarantee that composes them is void. §4 makes the catalog do work: `wai_capabilities` MUST be exhaustive, a sink MUST NOT subscribe assuming an undeclared capability won't appear, and `wai_determinism` is the **weakest** tier on the track — declaring the strongest would advertise a guarantee most objects don't meet. ### What it does not do No MOQT wire encoding, no replacement for CMSF or LOC, no making MoQ WAI's transport, and **no pinning of a draft revision**. MOQT is pre-Last-Call; where it and this text disagree, MOQT wins and this text is what changes. Ingest queue item 13b (#80 §4.4). Sourced from direct reads of `draft-ietf-moq-msf` and `draft-ietf-moq-cmsf-00`. 🤖 Generated with [Claude Code](https://claude.com/claude-code)

dcharlot added 1 commit 2026-09-07 04:20:21 -04:00

moq-streaming-format: declare WAI as a MoQ streaming format ... a243192a04

WAI already references MoQ for confidentiality (MLS + SFrame), and that binding
treats MoQ as a pipe. MoQ is not a pipe — it is a two-layer design, and the upper
layer has an open slot exactly where a container of capability-dispatched objects
belongs.

MOQT moves objects in groups on tracks and does not know what an object means.
MSF (draft-ietf-moq-msf) says what objects mean: it fragments a stream into
objects, declares what a publisher emits in a catalog, and sets prioritisation.
The alignment with WAI is structural rather than aspirational — MSF's unit is an
object in a group, WAI envelopes are already objects, and jwp-receipts already
commits per object and signs per group. Neither layer has to change to agree.

THE SLOT IS OPEN, AND CHEAPLY. MSF's `packaging` field has a fixed value set and
no central registry, which reads closed. draft-ietf-moq-cmsf shows it is not: a
sibling I-D that declares packaging value `cmaf`, says it extends the allowed
values defined in MSF, inherits MSF wholesale, adds two track-level catalog
fields, and registers nothing — "this document has no IANA actions". LOCMAF is a
second instance. A sibling draft, no registry action, no permission sought.

What CMSF did NOT take is the per-object slot. MSF §12 defines properties at
Track AND Object scope with an IANA type registry holding essentially one entry
(MSF_COMPRESSION); CMSF defines no per-object property. §5 takes that slot with
WAI_RECEIPT — content hash, joules, acquisition class, group-receipt hash — which
is the shape jwp-receipts already has. `acquisition` MUST accompany
`joules_micro`, because a joule figure without its acquisition class is what
energy-measurement §2 forbids, and putting it on a wire does not make it
acceptable.

The load-bearing rule is §3: the object boundary MUST be the envelope boundary.
That is what makes MOQT's object identity and the receipt's per-object commitment
the same unit. If they diverge, every guarantee that composes them is void.

§4 makes the catalog do work rather than decorate: `wai_capabilities` MUST be
exhaustive, a sink MUST NOT subscribe assuming an undeclared capability will not
appear, and `wai_determinism` is the WEAKEST tier on the track — declaring the
strongest would let a track advertise a guarantee most of its objects do not
meet.

§6 states what this does not do, so it is not re-argued: no MOQT wire encoding,
no replacement for CMSF or LOC, no making MoQ WAI's transport, and no pinning of
a draft revision. MOQT is pre-Last-Call; where it and this text disagree, MOQT
wins and this text is what changes.

Ingest queue item 13b from the landscape survey (#80 §4.4). Sourced from direct
reads of draft-ietf-moq-msf and draft-ietf-moq-cmsf-00.

Co-Authored-By: Claude Opus 5 <[email protected]>

dcharlot force-pushed wai/moq-streaming-format from a243192a04 to 0c17e8c846 2026-09-10 10:25:51 -04:00 Compare

dcharlot merged commit 87c34fddb5 into main 2026-09-10 12:21:10 -04:00

dcharlot referenced this pull request from a commit 2026-09-10 12:21:10 -04:00

Merge pull request 'moq-streaming-format: declare WAI as a MoQ streaming format' (#85) from wai/moq-streaming-format into main

Sign in to join this conversation.

Reviewers

No reviewers

Labels

Clear labels

No items

No labels

Milestone

Clear milestone

No items

No milestone

Projects

Clear projects

No items

No project

Assignees

Clear assignees

No assignees

1 participant

Notifications

Subscribe

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference

Transaction-Science/open-standards!85

No description provided.