Replies: 1 comment
-
|
I am curious of the "semantic_mappings" element proposed in the Extended Metadata proposal would allow for this: https://docs.google.com/document/d/1DLuteF0WwlVbpD5kmHhcqa6vItqNFL4MoAQrImy6dKo/edit?tab=t.0#heading=h.ormr47seafi For example: |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
OSI solves structural interoperability — any tool can read and write semantic models in a common format. What it does not solve is conceptual interoperability. When one semantic model defines a field as
security_typeand another calls the same conceptinstrument_class, both are structurally valid OSI. But a consumer still has to reverse-engineer that they mean the same thing. At scale, this manifests as a significant integration cost.The problem is not theoretical. In financial services, dozens of data providers and institutions independently model the same core concepts — trades, positions, instruments, claims — with no shared vocabulary to anchor them. Every integration rediscovers the same mapping problems. The structural standard gives us a wire format; it does not give us a shared language.
The need is already recognized outside of OSI. FIBO (Financial Industry Business Ontology), for example, provides thousands of open, industry-vetted concept definitions in OWL/RDF ontology format. Efforts like this confirm the demand for shared vocabularies — but these concepts are not yet consumable within an OSI semantic model.
Concept-level standards are not a silver bullet. Not every concept lends itself to standardization — some carry nuance, some embed intellectual property, and some vary enough across contexts that consensus may not be possible. But standardizing even the common core would meaningfully reduce integration friction, particularly when coupled with parallel developments on the OSI roadmap like ontology support and composability.
Concretely, this could take the form of a governed reference dictionary of concept definitions — organized by industry vertical, starting with financial services — that OSI models can optionally cite. Not a schema extension. Not a mandate. A shared vocabulary that gives model authors a common starting point and gives model consumers a basis for alignment without reverse-engineering field names.
This is a gap in the standard. Data providers, platform vendors, and consumers stitching models together are all dealing with it. this discussion is about how we can address this gap.
Beta Was this translation helpful? Give feedback.
All reactions