Invited Panel: Metadata as the Foundation for Dataspaces

As global data economies transition toward federated ecosystems, establishing cross-domain trust and seamless semantic alignment remains a pivotal policy objective for emerging digital nations. This presentation shares the strategic governance design and preliminary policy insights from the "Taiwan Data Space Creation Guidance Framework," an initiative spearheaded by the Ministry of Digital Affairs (moda) to guide Taiwan through its preparatory and pre-implementation phases. Reflecting international benchmarks, such as the European Data Spaces Support Centre (DSSC) blueprints and Japan’s Information-technology Promotion Agency (IPA) guidebooks, Taiwan’s forthcoming guidance draft formulates a phased roadmap across business, legal, operational, and technical dimensions, customized for an ecosystem at the inception stage. Prior to deploying physical infrastructure, the framework prioritizes the "soft infrastructure" principles needed to operationalize trust. It focuses on the strategic design of Identity and Attestation Management. Crucially, this session highlights how metadata governance serves as the architectural backbone during this foundational blueprinting phase. By integrating standardized specifications, such as DCAT application profiles, Taiwan envisions a model to guarantee future resource discoverability and data quality without proprietary vendor lock-in. Furthermore, the framework translates abstract legal frameworks into machine-readable readiness models, utilizing Open Digital Rights Language (ODRL) formalisms to design policy enforcement mechanisms long before data plane transactions occur. By charting Taiwan’s journey in defining clear rules, stakeholder roles, and interoperability standards during its inception phase, this presentation underscores the significance of institutional preparation. It outlines a visionary roadmap for a resilient data mesh where public sectors, national access points, and private enterprises can confidently co-create value in a secure, sovereign environment.

As an opening remark for the panel on "Metadata-Governed Data Spaces: Trust, Usage Control, and AI-Ready Data Sharing", I will present W3C's current and upcoming work related to this topic

Open Data Spaces (ODA) is an open and scalable architecture for distributed data management (DDM), while respecting the diversity of organizations and countries. Rooted in the concept of data mesh and classical data spaces (first proposed in the U.S. in 2005), ODS is evolving architecture as the Double Product Quanta Model (DPQM) to enable the DDM across the organization and a trustworthy post-training data pipeline for Agentic AI.