A presentation by @Paul Knowles and @Robert Mitwicki of the Human Colossus Foundation.
A presentation on Identity Semantics with concepts from Overlays Capture Architecture (OCA), Blinding Identity Taxonomy and more. This is a key area especially for interoperability & consent architecture. they have been exploring Identity in the Semantics WG in TOIP. Their primary interest is in the healthcare field.
- Human Colossus Foundation (HCF): is a not-for-profit and releases most of their work product under the GPL, with appropriate product under Apache (due to adoption and derivation concerns).
- Paul Knowles presented OCA-
- There is a schema-base (which is a machine-readable definition of the semantics of a data structure in its purest form) which includes a blinding block (which allows the issuer to flag any attributes that could potentially unblind the identity of a governing entity in reference to BIT (Blinding Identity Taxonomy). In other words, sensitive fields are blinded through encryption, hiding during presentation etc.
- Overlays can then be layered on top of the schema-base to provide extra contextual and/or conditional information to that base object. They can can be used by an issuer to transform how information is displayed to a viewer or to guide a verifier or holder in how to apply a custom process to schema data. In other words, they enhance or guide the different aspects of the schema-base mapping for automatic processing and to present user interfaces. These include (but not limited to) mapping, source, subset, label, entry, format etc. This transforms the schema-base into a multi-dimensional array. These objects are linked to the schema-base through a hash of the base object. All objects are stored in (but are not limited to) an OCA repository.
- Meta-data is a part of the data that is managed. The schema and the overlays can add items like source etc.
- JSON-LD as linked data co-exists with OCA
- Robert Mitwicki demonstrated the way this works through a public sandbox argo.colossi.network. The sandbox has links to available schemas. There were some glitches mainly due to poor internet connectivity; the bane of WFH. However RM did make the whole thing work.
- Paul demonstrated a sample schema-base and various overlays in JSON. This brought clarity to the concepts and showed the succinct nature of the solution.
- Paul also said that most of the pre-standards work was happening in TOIP Semantics WG.
- He also showed how existing multi-dimensional constructs could be modified to bootstrap new use-cases in new contexts for different enterprises. The presence of a GICS code also showed how meta-data standards could be used to further enhance interoperability work.
- There was a Q&A
- How does it work with NIEM schemas? Although NIEM is specific to the US, there would be no problem converting the schemas into this architecture.
- Any production versions of this? Work was proceeding on getting this into production
- How do they work with GDPR, especially the data retention for legal purposes? Even if there were disconnects, the relying party could retain data if there overriding constructs.
- What about other fields (other than healthcare) can this also be applied to data other than Identity data? Yes, work is ongoing for Financial sector under Philippe Page who is the chairman of HCF.