Auditability
Standards-compliant audit logs
Can oracle solutions be trusted if they cannot, or will not, demonstrate how their source data was collected and validated? We don't think they should be.
As a response to this question, Orcfax generates standards-compliant archival packages that are made available permissionlessly so that anybody has the ability to audit the flow of Orcfax data collection, validation, and publication.
The Orcfax Explorer
To demonstrate the audit-enabling function and re-usability of these archival packages, while further simplifying the auditing process for users, Orcfax launched its Orcfax Explorer; this tool provides a user friendly dashboard which enables users to quickly and easily navigate all data published by Orcfax, along with complete documentation for all processes leading to publication.
Each Orcfax Fact Statement that is published on the Cardano blockchain has a corresponding audit trail that documents how the primary source data was collected and validated before it was published as a Cardano transaction datum. These audit logs are marked up using both industry-standards and open-data standards that enable machine-readablility, like IETF Bagit, which improves digital repository interoperability for digital archival packages.
While navigating through the Explorer, users can click on a Fact Statement card to see the detailed summary for that Fact Statement. The Archive Explorer viewer loads the archival package from the Arweave network (via Arkly.io) and provides the user with an intuitive view of all the audit log files.
Click on the link below to view a walkthrough of the Orcfax Explorer.
This product brings additional value to the Orcfax solution by providing tooling for users which enables them to freely search for, and review, all Fact Statements published by Orcfax. The Explorer does this through purposeful user-centered design decisions which have prioritized accessibility and ease of use for a dynamic and diverse user group.
Orcfax will continue to build upon this service in order to deliver meaningful user tools which allow them to "trust but verify".
Future value
The archival packages created by Orcfax are described using Schema.org and JSON-LD compliant taxonomies, which have been purposefully selected in order to enrich Orcfax data. Additionally, by leveraging these taxonomies, Orcfax makes its outputs Linked Data and AI-training compatible, which dramatically enhances downstream value.
JSON-LD is the most widely used serialization of the Resource Description Framework (RDF); this framework ensures that data is machine-readable. And schema.org is the single most impactful markup language for making any type of metadata machine-readable. In Orcfax's case, that means metadata related to Claims made about events happening in the real world.
The use of these taxonomies, and the resulting highly structured data, will perfectly position Orcfax feeds to serve as the source of a growing data lake of real world Fact Statements that will have many secondary reference uses beyond DeFi oracle publication. The highly machine and human readable data lake of validated Fact Statements can be used, for example, to train enterprise or personal AI models with reliable real-world data.
Some of the most current research on AI makes it clear that the in-deterministic nature of Large Language Models (LLM) cannot be relied upon for factual citations; these technologies need to be supported and trained by highly structured Fact Statements about the real world that humans inhabit. We anticipate that the Orcfax protocol, anchored by the decentralized resiliency of the Cardano blockchain network, will have the ability to provide this foundational layer.
We anticipate a suite of enterprise service layers on top of this data lake of decentralized, validated Fact Statements as a secondary revenue stream for the Orcfax protocol once we reach a critical mass of feed types, data sources, and validator nodes.