Open bblazicevic opened 3 months ago
onthebreeze
commented
2 months ago Chat with Marcus about how a similar problem was handled in the case of regulatory compliance with climate related financial disclosures in Australia. It may bring some useful lessons about how to iterate the conformity credential to suit regulatory compliance use cases.
seewodg
commented
2 months ago Thanks @onthebreeze for raising this important issue, and thanks to @bblazicevic for raising it.
Here is my rather long response (sorry)...
In a team discussion, while seeking to close outstanding issues, Steve brought this issue into focus and team discussion ensured. I offered the observation that one of my customers, a reasonably large multinational accounting firm, had a requirement to link reporting metrics for climate change (under Australian legislation) to parts relevant parts of, parts of relevant legislation. This customer is seeking to assist their clients with both tax offset claims and with climate change legislated reporting requirements, as mandated under Australian Legislation (Acts of Parliament) and Legislative instruments.
Mostly, each of this particular accounting enterprise clients, will have (or must develop) more or less distinct climate reporting metrics, whereas those metrics must conform to the spectified in setting methods that links them. Methods that are somewhat encoded in regulation (inc: references to standards bodies). Same for tax offset claims in that the descriptions in claims need to apply to legislated conformance criteria. According to my customer, it is "prudent" for their clients to link specific metrics to specific parts of legislation, same for descriptions for tax offset claims.
In assisting my customer I harness the situation whereby a consistent publication approach is applied when publishing legislation on the web in Australia (as specified within legislation). This provides structure for mining the metadata that itself specifies parts of legislation, including mining links to the parts.
That enables automated mining function outputs legislation part addresses as RDF. The resulting RDF can be serialised as linked data in the usual formats. Then the mined metadata from legislation is expressed as a DCAT Dataset, scoped to an Act of Parliament.
Each entry in the dataset will include a hyperlink to the part of the legislation.
The RDF (DCAT) dataset(s) generated as RDF (including hyperlinks and available related metadata) will then be indexed in a DCAT catalog. The dataset contents are transitively referenceable from the catalog.
Cataloged datasets can be referenced from individually tailored Profiles (of the catalog) providing links to cataloged datasets in the context of a particular accounting firm's client. Thanks to transitivity (declared and/or inferred using inference rules), individual part of legislation can be reference in a Profile.
I point all this out because in assisting my customer with capability in this space it seems to present some promise regarding the issue of linkage assessment scheme to regulation.
DCAT catalogs can index DCAT catalogs, which itself is in keeping with the UNTP for a distributed information sharing ecosystem.
Not to overload a concept (I hope), it is worth considering that catalogs themselves will be "Profiled", as is the cast today. Perhaps the UNTP doesn't require a Profile for a DCAT Catalog but instead can simply adopt an existing one.
Governance regarding update of catalogs is in the UNTP, it is perhaps a matter of trust anchor ownership of catalogs. Governance over profiles is likely a vertical industry concern, or even individual entities. If there existed a UNTP DCAT Catalog (and perhaps there should be), the governance could be distributed among trust anchor participants or via other governance body.
The use of DCAT is a well established semantic architectural pattern for referencing regulation that works today, within the constraints of machine readability at least regarding regulation, and in particular regarding Australian legislation (which I know about), and linking to structure in regulation via semi persistent URL's is available (inc: referencing versions and changes). Globally that assertion itself lives on a spectrum.
Hopefully this feedback is on topic for this issue.
In any case, one of my goals when working in any transaction ecosystem includes seeking to "keep the message payload as lightweight as possible while delivering maximum available context".
Is my goal immediately above captured in the UNTP goals Steve, I must go back and reread them?
Does this respon help highlight the issue Bree or throw more mud on the pond?
Discussion point - needs further clarification from UNTP regarding linkage of assessment scheme to regulation. Will assume for the purpose of demo that this artifact relates to the governance of the credential (application layer).
https://uncefact.github.io/spec-untp/docs/specification/ConformityCredential/#conformityassessmentscheme