Linking information requirements to their regulatory clauses

[odd-Innocent]

Hello,

I work on den, a remote MCP server that answers questions about Korean building codes and construction standards and attaches the source clause to each answer.

I would like to ask one question. When an information requirement comes from a regulatory clause, is it useful for the requirement itself to carry a link to that clause?

If you have experience with linking requirements to their regulatory sources, I would be grateful for your view. Thank you for your time.

den: https://den.archi

[andyward]

I had to use AI to make sense of this idea 😄

The crux of the question seems to be:

when an information requirement depends on regulatory clauses, does exposing the cited path make the requirement easier to verify or change?

I don't think it makes model verification easier, because in IDS verification is explicitly defined by the Requirement facets. An IDS validator does not need to know why a requirement exists in order to test whether a model complies with it.

It may, however, make it easier to verify, justify, maintain, or change the requirement itself, because it provides traceability back to the source that drove the requirement. In other words, it helps answer WHY the client has this requirement.

Whether that traceability should point directly to a regulatory clause, or instead to a client's information requirements that themselves have regulatory drivers, is probably use-case dependent.

In IDS today, the closest mechanism would be the URI on a requirement. However, for a canonical requirement such as:

Doors must have Pset_DoorCommon.FireRating populated

I'm not sure how useful it would be for the URI to point directly to, say, a clause in the UK Fire Safety Act 2021, rather than to the FireRating property definition, a client requirement, or some other authoritative source.

This feels less like a validation problem and more like a provenance or requirements-traceability problem.

One idea for a future IDS version might be support for multiple linked sources, potentially with metadata describing the relationship. This could allow separate references to:

• the regulatory clause;
• the client's information requirements;
• relevant standards or classifications (bSDD, IFC property definitions, etc.);
• other project- or organisation-specific sources.

That would preserve the machine-testable requirement while also making its rationale and lineage more transparent.

[odd-Innocent]

Thank you for taking the time to reply, and I apologise that my original post was hard to follow. I have shortened it.

You have stated the question better than I did. I agree this is about the provenance and traceability of the requirement, not about model validation. In den, each answer keeps the regulatory clause and the technical standard as separate cited sources, so a requirement with several linked sources, each with its relationship stated, would fit that well.

I will follow how this develops in IDS. Thank you again.