Add section on Controller Document data model.

[msporny]

This PR is an attempt to address issue #31 by defining the properties that are required and optional in controller documents. This text is ported from the DID Core specification (it was deleted, by mistake, when the original document was ported over).

---

Preview | Diff

[iherman]

The new text includes this:

<p class="note" title="Property names used in map of different types">
The property names `id`, `type`, and `controller` can be present in map of
different types with possible differences in constraints.
 </p>

Unless I completely misunderstand, I believe this note contains essential information that should be part of the core text and elaborated upon. This note means that different other specification can use the controller document "structure" in their own, security related specification. I.e., a Verifiable Credential, a Verifiable Presentation, or a DID document are all examples of controller documents (and this is how this specification relates to the VCDM or DID).

If this is all true, it should be made much more clear and explicit somewhere in the document, and shown also with examples.

From an ontology/vocabulary point of view, the clean way would be to define a class for ControllerDocument on which some properties are defined, and then a ~VerifiableCredential, a VerifiablePresentation~ ProofGraph or a DIDDocument is defined as a subclass. (It may be too late to do that.)

(Edited 12 hours later, realizing that a VC or a VP is probably not a Controller Document. My mistake.)

[iherman]

Also related to https://github.com/w3c/controller-document/pull/32#issuecomment-2169118081: the terminology on controller document is, sort of, circular:

controller An entity that has the capability to make changes to a controller document. controller document A set of data that specifies one or more relationships between a controller and a set of data, such as a set of public cryptographic keys.

Nowhere is the term "controller document" properly defined; the definition above fits any kind of resource under the Sun ("such as" means that any property can be present...). The Data model section, which includes the aforementioned note, is also void of real information.

We should come up with a more specific definition in the terminology section to make the term more "restrictive", i.e., to really separate Controller Documents from other resources.

(I must admit that I do not understand how this term "Controller" fits here. But it may be too late to come up with a better name. All the more important to provide a better descriptive text.)

[iherman]

The issue was discussed in a meeting on 2024-06-19

• no resolutions were taken

View the transcript

#### 3.1. Add section on Controller Document data model. (pr controller-document#32) _See github pull request [controller-document#32](https://github.com/w3c/controller-document/pull/32)._ **Manu Sporny:** This replaces a section that was accidentally removed a while ago. We never actually said what we would find in the controller document. Lifted text from DIDCORE with modifications. … Talks about what are the mandatory and optional properties beyond just verification methods and relationships. … Some feedback on text from Ivan, some good points but the further we get from what DIDCORE says the harder it is to maintain alignment. … TallTed's changes are also good. … Need guidance from the group. Are we going to address Ivan's more fundamental concerns about the PR? **Brent Zundel:** Not seeing anyone on the queue. It seems like Ivan's comments might be better addressed as a separate issue rather than as part of this PR. **Ted Thibodeau Jr.:** Table changes are also editorial. **David Chadwick:** Question to Manu. If the controller property is present and says "Fred and Mary can do this", what does it say about the subject? … Does the subject still have all the properties as if the controller is absent? **Manu Sporny:** If there is no controller field, then the subject is the controller. If there is a controller, I believe that most implementations allow the subject to have the same power. **David Chadwick:** Will raise an issue. **Manu Sporny:** TallTed, trying to bunch the same description in the fields, will try to fix the table. **KevinDean:** Supports the need for clarification of controller. Parents may be controllers of a child's VC without wanting the child to have the same capability.

[iherman]

The issue was discussed in a meeting on 2024-06-26

• no resolutions were taken

View the transcript

#### 3.1. Add section on Controller Document data model. (pr controller-document#32) _See github pull request [controller-document#32](https://github.com/w3c/controller-document/pull/32)._ **Manu Sporny:** we have a number of approvals already on these additions to the data model. … the text is mainly from DID core so I prefer not to change it in the Controller document, but rather raise issues on DID core. … specifically to ivan's comments. **Ivan Herman:** I could not properly explain what a controller document is from the existing text. … need a higher level section to explain what the controller document is actually about.

[msporny]

Normative, multiple reviews, changes requested and made, no objections, merging.