# Limits and trade-offs This page is the honest answer to the sharpest objection this standard receives: if every Universe must adopt a common model to federate, is that not centralization wearing a different hat? The short answer is that there is a real, mandatory, shared layer. We do not hide it. This page names exactly what it is, what it is not, and what you pay for it. If you read only one page before deciding whether to trust the standard, read this one. --- ## What the common layer actually requires To federate, a Universe MUST agree to a fixed set of structural primitives and conduct duties. This is the thin, mandatory form. It is defined by the [Meta-Universe Constitution (MUC)](/spec/#01-constitution/Meta-Universe-Constitution.md), and it is the only thing that is not optional. The MUC is the authoritative enumeration: 21 articles across six chapters. It is not a schema of the world; it is a small set of invariants about how a unit of meaning is shaped, moved, and accounted for. The load-bearing ones are these. **How a unit of meaning is shaped and moved:** - **Identity.** Every unit of meaning carries a globally unique, persistent identifier (MUC Article 2). Labels and representations may change; the identity does not. - **Object and projection are distinct.** Knowledge crosses a boundary as a projection, a view under a specific context and purpose, never a copy that redefines the source (MUC Article 5). What another Universe receives is a Projection, not a transfer of the source object. - **Provenance.** Every significant fact declares where it came from; unknown provenance is marked as unknown, not hidden (MUC Article 7). - **Traceability.** Origin, ownership, evolution, and dependencies of a fact stay determinable (MUC Article 8). - **Shared schemas, not a shared dictionary.** Interoperability begins with each Universe publishing its own schema; publishing a schema never implies publishing the data behind it (MUC Article 9). - **Explicit contracts.** Knowledge is exchanged only through a declared contract that states purpose and scope. Nothing flows implicitly (MUC Article 11). - **Explicit mappings.** Where two models differ, the difference is stated as an explicit mapping; implicit semantic assumptions are disallowed (MUC Article 17). **What you agree to do when you exchange:** - **Least knowledge.** A participant receives only the minimum required for the agreed purpose; projections are preferred over whole objects (MUC Article 12). - **Declared purpose.** Every request declares its purpose, and purpose participates in the authorization decision (MUC Article 13). - **Trust is never assumed.** It is established through verifiable identity, provenance, evidence, or agreement (MUC Article 14). - **Security is independent of trust.** Establishing trust does not remove the obligation to protect the integrity, authenticity, and confidentiality of what you disclose (MUC Article 15). That is a real, mandatory core, and it is larger than a single slogan: it fixes how meaning is identified, projected, traced, and contracted, and it binds you to conduct duties when you exchange. It is a genuine point of central agreement, shared by every participant, and it is not negotiable per node. Pretending it is only a handful of harmless shapes would understate the price, and understating the price is exactly the failure this page exists to prevent. The full, article-by-article text is the [Meta-Universe Constitution](/spec/#01-constitution/Meta-Universe-Constitution.md); read it if you want the whole bill. A note on vocabulary: the landing page presents the model through a presentation vocabulary (Dimension, Namespace, Object, Projection). Those are the surface names for the same invariants the MUC states normatively. The MUC articles are the authoritative enumeration of the mandatory form; the landing terms map onto them, they do not add a separate price. ## What the common layer does NOT require The form says nothing about what your meaning *is*. Domain meaning stays entirely local: - What "customer" means, what a "part" is, what counts as a "contract" in your business: **yours, defined by you, never imposed.** - Your vocabulary, your entity boundaries, your rules, your truth: they stay home. - No single master schema anyone must map into, and no apex authority that owns the definitions. Mappings between models are explicit and bilateral (MUC Article 17), not submission to one central dictionary of domain terms. You conform your models to a structural form. You never surrender their content. Two Universes can federate while disagreeing completely about what a "product" is, because the standard fixes the container, not the cargo. --- ## The deal, stated plainly Every working interoperability standard makes the same trade, and none of them is considered centralization in the objectionable sense: - **TCP/IP fixes the packet, not the traffic.** Every machine on the internet agrees on the packet format. No machine agrees on what you are allowed to say inside it. - **The shipping container fixes the dimensions, not the cargo.** The whole world standardized the corner castings and the outer box. Nobody standardized what you ship in it. - **HTML fixes the markup, not the text.** Every browser agrees on what a paragraph tag means. No standards body approves your prose. In each case the shared layer is thin, mandatory, and deliberately blind to content. That is what makes it safe to adopt. Vercy is the same bargain applied to meaning: **the form is shared so that the content does not have to be.** ## The honest claim We do not claim "no centralization." That claim would be false, and a skeptic would catch it in one read. The honest claim is narrower and defensible: > **Centralization here is minimal, limited to form, and accountable.** - **Minimal**: a small set of constitutional invariants, not a schema of the world. - **Limited to form**: it constrains how a unit of meaning is shaped and exchanged, never what it means. - **Accountable**: who governs the form is itself governed, in the open, by a published process (see below). If any of those three stops being true, the objection is correct and the standard has failed. They are commitments, not decoration. --- ## The hard question, answered Stated in plain words, the objection is this: > **If participation requires adopting a common constitutional model, and the terms of > participation are not optional, then is content sovereignty real, or has centralization > simply moved one floor up?** It is the right question, and half of it lands. Yes, the terms are not optional: to federate, you conform to the form. Calling that "no centralization" is spin. So sovereignty cannot mean "no shared rules," because under that definition no interoperability standard in history has ever been compatible with sovereignty, and that is plainly not how the word is used. Here is the working definition these pages adopt, and that the standard is being aligned to across its canonical sources: > **Sovereignty is not the absence of shared rules. It is the right to leave with your models, > the right to fork the shared layer, and the right to participate in changing it. Conforming > to the shared form is the price of interoperability, and it is named explicitly.** This alignment is not yet complete on every page. Some canonical documents still carry older absolute framing (for example the Vision "quotable beliefs" and the MUC Final Principle), and reconciling them to the definition above is tracked, open work, not a finished claim. We would rather say that plainly here than pretend the whole corpus already speaks with one voice. Against that definition, the objection resolves into a single genuine risk and its answer. **The genuine risk: who governs the constitution?** If one steward can change the mandatory form at will, then your content sovereignty is conditional on that steward's good behavior. Structural separation of form from content does not fix this. Only governance does. It is closed by three things, and the first two already hold today: 1. **Right of exit.** The artifacts are plain files under an open license (Apache-2.0). Leaving costs you the federation, but it never costs you your models. You walk away with everything you authored, in a format nobody can revoke. 2. **Right of fork.** The standard forks like any open standard. If the steward drifts, the community can take the last good version and continue. This is what makes the steward's behavior checkable: the threat of a fork is real, so the incentive is to keep the form thin and the process open. 3. **Right to participate in change.** Changes to the form go through a public, documented process, not an owner's private decision. Any stakeholder MAY propose a change; every change is versioned, reviewed and auditable. See [Governance, in brief](/governance/) for the condensed answer, and the full [Governance](/spec/#01-constitution/Governance.md) and [Change Process](/spec/#01-constitution/Change-Process.md) documents for the mechanics. The governance documents state the mechanics directly: released versions are immutable, frozen documents may only change through a published Change Request, breaking changes must be labeled and carry migration guidance, and the whole revision history stays auditable (why a change happened, who proposed it, when it was accepted, which version introduced it). Ownership is stewardship of the specification, not ownership of anyone's implementations or published knowledge. So: is sovereignty real? Under the honest definition, yes, because you can exit, fork, and help govern. The one part that a skeptic should keep watching is not exit or fork, which are already true, but whether the change process stays genuinely open in practice. That is the right thing to hold the standard accountable to, and it is exactly where we are pointing you to look. --- ## What you give and what you get **What you give** - You conform your models to the constitutional form: identity, the object/projection separation, provenance, traceability, shared schemas, explicit contracts and mappings, plus the conduct duties (least knowledge, declared purpose, trust never assumed, security independent of trust). - You accept that this form is mandatory and shared, not chosen per node. - You accept a versioned change process for that form, rather than changing it unilaterally. **What you get** - Your domain meaning stays local and unaltered. Nobody maps you into a master schema. - Knowledge becomes portable across boundaries as verifiable Projections, not lossy copies. - Every exchange is explicit, contracted, and traceable to its origin. - The right to leave with everything you authored, in open files, at any time. - The right to fork the standard if its stewardship stops serving you. - A vote, through an open process, in how the shared form evolves. The trade is deliberate and it is the same trade every durable standard makes. The form is small so that the content stays free. We charge that price openly rather than pretending it is free. --- ## Not just words: the form can refuse A form that cannot reject anything is not a standard, it is a glossary. This one refuses. The [registry conformance checks](https://github.com/ver-cy/registry) are executable: an invalid model, a missing export, a broken edge in the graph, a date left in a filename, or an index that has drifted out of sync each produce a non-zero failure that cites the specific normative rule violated. Conformance is something a model can fail, which is the only kind of conformance worth claiming. --- ## Where to go next - **Who governs the form, and what if you disagree:** [Governance, in brief](/governance/), then the full [Governance](/spec/#01-constitution/Governance.md) and [Change Process](/spec/#01-constitution/Change-Process.md) documents. - **The mandatory form itself, article by article:** [Meta-Universe Constitution](/spec/#01-constitution/Meta-Universe-Constitution.md). - **The whole standard in one connected read:** [Overview](/spec/#00-foundation/Overview.md).