You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
NovaFabric is at v0.99.0. The next real milestone is freezing the v1.0 capsule format — and once it is frozen, it is a compatibility promise we intend to keep for years.
So this is the moment to argue about it.
What "v1.0" actually means here
Not "we ran out of bugs". It means the Run Capsule and Evidence Bundle formats stop changing incompatibly. Everything else — server mode, the dashboard, the collector — can keep evolving. The formats cannot, because the entire premise is that a capsule you archive today still verifies in five years.
The gate that is actually blocking it
Three independent design-partner sign-offs. Currently zero.
Not "nearly three". Zero. The one entry in the register is the maintainer's own validation, which does not count — self-validation by the person who designed the format is precisely the failure mode the gate exists to prevent.
No amount of engineering unblocks this. It needs people running real workloads to say "yes, this format survives contact with what I actually do".
What we want from you
1. What would break for you?
If the format froze exactly as it is today, what would you be unable to express? Missing field, wrong shape, an assumption that does not hold for your workload?
2. What is over-built?
Which parts of the capsule have you never looked at? A field nobody reads is a field that should not be frozen into a compatibility promise.
3. What is under-specified?
Where did you have to guess what something meant?
4. What would make you adopt it?
The honest version, including "nothing, it is not for me" — that answer is genuinely useful and we would rather hear it now.
Known open questions
Things we already know are unresolved, so you do not spend time rediscovering them:
Schema versioning across the freeze — producers currently write 0.1.0 while the canonical v1 targets are separately identified (ADR-0223). The producer flip is deliberately gated on the freeze itself.
Session vs parent/child grouping — two different grouping concepts coexist. Are both earning their place?
Evidence facets — experimental. Freeze them, or leave them extensible?
How much of the compliance surface belongs in the core format versus in exporters that read it?
Not a vote
To be straight about the process: this is input, weighted seriously, but decisions get made through the RFC process and the project lead has final say until a TSC exists. What you say here can absolutely change the outcome — it just does it by convincing people, not by counting.
Reply with anything: a paragraph, a bullet, a single field name you think is wrong. Half-formed is fine.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
NovaFabric is at v0.99.0. The next real milestone is freezing the v1.0 capsule format — and once it is frozen, it is a compatibility promise we intend to keep for years.
So this is the moment to argue about it.
What "v1.0" actually means here
Not "we ran out of bugs". It means the Run Capsule and Evidence Bundle formats stop changing incompatibly. Everything else — server mode, the dashboard, the collector — can keep evolving. The formats cannot, because the entire premise is that a capsule you archive today still verifies in five years.
The gate that is actually blocking it
Three independent design-partner sign-offs. Currently zero.
Not "nearly three". Zero. The one entry in the register is the maintainer's own validation, which does not count — self-validation by the person who designed the format is precisely the failure mode the gate exists to prevent.
No amount of engineering unblocks this. It needs people running real workloads to say "yes, this format survives contact with what I actually do".
What we want from you
1. What would break for you?
If the format froze exactly as it is today, what would you be unable to express? Missing field, wrong shape, an assumption that does not hold for your workload?
2. What is over-built?
Which parts of the capsule have you never looked at? A field nobody reads is a field that should not be frozen into a compatibility promise.
3. What is under-specified?
Where did you have to guess what something meant?
4. What would make you adopt it?
The honest version, including "nothing, it is not for me" — that answer is genuinely useful and we would rather hear it now.
Known open questions
Things we already know are unresolved, so you do not spend time rediscovering them:
0.1.0while the canonical v1 targets are separately identified (ADR-0223). The producer flip is deliberately gated on the freeze itself.experimental. Freeze them, or leave them extensible?Not a vote
To be straight about the process: this is input, weighted seriously, but decisions get made through the RFC process and the project lead has final say until a TSC exists. What you say here can absolutely change the outcome — it just does it by convincing people, not by counting.
Reply with anything: a paragraph, a bullet, a single field name you think is wrong. Half-formed is fine.
All reactions