ISO 19135 Improvement of figure 1 (Organizational relationships)

[heidivanparys]

Figure 1 should be improved to help the reader of the standard understand the different roles and responsibilities, keeping in mind that this model is not a data model.

Specific proposal for changes:

• make sure that nouns are used as role names (qualifier ✔️, qualifiedBy ❌)
• remove the role names unless they really add value
• add a verb and a reading direction to the relation (more than one word can be used, the idea is to explain the relation to the reader)†
• consider whether more relations than the ones present now should be added (more than one relation between the two same roles/concepts may be present)
• consider whether the multiplicities add value or not
• consider using lower case for the terms, so they match the terms and definitions and the text. Again, this figure is not a data model, so I would argue that it is ok to deviate from the UML specification

A small example of what I mean:

† For more information about the importance of verbs on relations for helping the reader understand the meaning (and for making sure that the writer has thought everything through!), I refer to:

• Verb Concepts in Concept Models
• The Theory Underlying Concept Maps and How to Construct and Use Them

[heidivanparys]

Some input to this discussion: I recently needed a diagram illustrating some of the core concepts described in ISO 19135-1, and my colleagues and I came up with the following diagram:

[PeterParslow]

Do you see this as a concept model, conceptual model, or something else again?

Is it a 'different diagram off the same (consolidated) model' - with attributes suppressed & different association labels? Although I think only the register & register system "boxes" equate to classes in the current model?

Although actually is that how we (TC 211 & OGC conceptual model discussions) would handle the 'different levels of abstraction'?

Would it perhaps be more like a UML use case diagram, with the other things being Actors?

That would make it a fitting replacement for Figure 1, which I'm not sure has even been part of the UML model before.

[heidivanparys]

In the terminology of the agency where I work, that would be a concept model.

It is not a data model, the (conceptual?) data model for ISO 19135-1 is this one.

It does not contain roles, because I don't think they would add much value on this particular diagram. But you could add them for completeness of the model, and hide them on the diagram. But if the role names aren't used (because it isn't a data model, you won't generate any XML schemas from it) and don't give other value, you may perhaps leave them out completely.

You could also make a model that conveys some of the same information differently, and that does use role names:

Here it is illustrated that a register manager, register owner, etc. are organizations¹. But on this UML diagram it is not possible to illustrate the relations between the roles, e.g., that a register owner authorizes a submitting organization etc. I think there is no right or wrong here, it't just the universe of discourse seen from a slightly different perspective. The definitions of these concepts would still be the same.

¹ While drawing this diagram, I realized that ISO 19135-1 actually does not define a control body as an organization, but as a group of experts… I left it in for now, as this was just an example anyway. But perhaps reusing a ISO “organization” concept that matches what it meant in ISO 19135-1 would be an idea… 🙂

[heidivanparys]

Would it perhaps be more like a UML use case diagram, with the other things being Actors?

It should be possible to draw the diagram using UML actors according to the UML spec and https://www.uml-diagrams.org/use-case-actor.html, but EA does not seem to allow this…

A quick test with PlantUML shows that there it is possible to have relations between actors. But I'm not sure it is possible to mix actors and regular classes...