Full Papers

As research becomes increasingly data-driven and interdisciplinary, metadata interoperability has become essential for efficient data discovery and reuse. However, many domain-specific metadata schemas lack clear specifications and standardized structures, making cross-domain integration difficult. To address this issue, this study proposes a systematic three-phase workflow for cross-domain metadata mapping. The workflow consists of Phase (i) preliminary assessment of metadata schemas, Phase (ii) mapping relationship analysis, and Phase (iii) XSLT-based metadata transformation. To evaluate the effectiveness of the proposed workflow, a case study is conducted using a set of 125 metadata schemas. The results show that only 10 out of 125 schemas satisfy the requirements for mapping to the six mandatory DataCite properties, while the remaining schemas are limited by incomplete documentation or structural heterogeneity. In particular, challenges in representing properties such as ``Identifier'' and ``ResourceType'' highlight persistent semantic and structural mismatches across metadata standards. XSLT transformation files were developed and released, enabling practical implementation of the proposed mapping approach. This study contributes a systematic workflow for metadata mapping and provides empirical evidence on the limitations of current metadata standardization practices, supporting future efforts toward improved cross-domain metadata interoperability.

The rapid expansion of digital repositories and scholarly resources has increased the demand for scalable and intelligent metadata management systems. Traditional metadata creation methods, which rely on manual cataloguing by information professionals, struggle to keep pace with the growing volume and heterogeneity of digital content. This study develops and evaluates an AI-guided framework, Metadata-Building-AI-Guidance V.1.0 that combines document ingestion, chunking, OpenAI text-embedding-3-small embeddings, and the gpt-4o-mini large language model into a single Streamlit application supporting both automated Dublin Core-aligned metadata extraction and embedding-based semantic retrieval. We position the system as a design-science artefact in which retrieval is not a side feature but a feedback loop: the same vector index that powers similarity search is also used to surface the contextual chunks from which structured metadata are extracted and to support retrieval-augmented question answering over uploaded collections. The system was evaluated on a purposively sampled corpus of 30 open-access academic documents and 20 expert queries, using (i) field-level F1 against librarian-curated ground truth with Cohen's κ for inter-annotator agreement and (ii) Precision@5 and Mean Reciprocal Rank with binary relevance judgements from two librarians. Field-level F1 ranged from 0.67 to 0.73 for dc:title, dc:creator, dc:subject, and dc:description, with overall κ = 0.83 (almost perfect agreement); the lowest F1 was 0.57 for dc:type, traced to a fixed generic prompt output rather than to a model-capability limitation. Semantic retrieval reached Precision@5 = 0.61 and MRR = 0.69 with κ = 0.66 (substantial agreement) across the 20 queries. We discuss the limits of LLM-only evaluation—including the absence of a head-to-head comparison with established non-LLM extractors such as GROBID—and identify controlled baseline comparison together with a refined dc:type prompt as the immediate next steps. Prompts, JSON schema, library versions, sampling log, and evaluation queries are released to support replication. The contribution is a reproducible reference implementation that aligns AI-assisted metadata extraction with Dublin Core Terms and the FAIR principles for digital libraries, archives, and cultural-heritage repositories.

As online comic reading continues to grow in Taiwan, online platforms have become important sites for both comic consumption and discovery. This exploratory study examines genre-labeling practices on three major online comic platforms in Taiwan. We collected genre terms from the platforms’ Traditional Chinese and English interfaces and generated a list of 35 unique English genre terms. These terms were then mapped onto a facet framework drawn from previous literature. Our preliminary findings show that platform genre labels function as multidimensional access points rather than simple genre categories, representing aspects such as setting, mood, plot or narrative, and production context. Cross-platform differences in labeling granularity and cross-language differences in semantic scope were also observed. The findings may inform the development of more consistent and user-friendly genre labels for digital comic environments.

MARC-and MODS-based bibliographic data, due to their record-based structure, have limitations in supporting entity identification and semantic relationships in linked data environments. Existing approaches to converting these formats into BIBFRAME primarily rely on field-level mappings, which fail to capture bibliographic meaning and relationships adequately. To address this issue, this study proposes a semantic-based transformation framework that reinterprets bibliographic data at the level of semantic units. The framework includes a normalization model and a transformation module consisting of semantic analysis, entity extraction, relationship generation, and RDF transformation. This approach enables the reconstruction of bibliographic data into an entity–relationship structure, facilitating their conversion into BIBFRAME while preserving semantic integrity and enhancing interoperability and reusability.

MARC- and MODS-based bibliographic data, due to their record-based structure, have limitations in supporting entity identification and semantic relationships in linked data environments. Existing approaches to converting these formats into BIBFRAME primarily rely on field-level mappings, which fail to capture bibliographic meaning and relationships adequately. To address this issue, this study proposes a semantic-based transformation framework that reinterprets bibliographic data at the level of semantic units. The framework includes a normalization model and a transformation module consisting of semantic analysis, entity extraction, relationship generation, and RDF transformation. This approach enables the reconstruction of bibliographic data into an entity–relationship structure, facilitating their conversion into BIBFRAME while preserving semantic integrity and enhancing interoperability and reusability.

National web-scale discovery services (WSDS) depend on the sustained interoperability of institutional repositories to deliver reliable access to scholarly content. Existing evaluations predominantly assess interoperability through metadata completeness at registration, overlooking the operational dimensions that determine long-term integration. This paper introduces the Multidimensional Interoperability Readiness (MIR) framework, which integrates three analytically distinct dimensions: metadata completeness, harvesting sustainability, and metadata capacity. The framework is validated empirically through analysis of the Indonesia One Search (IOS) registry and OAI-PMH harvesting dataset comprising 29 registration fields across four functional categories. Findings reveal a structural decoupling between metadata completeness and harvesting sustainability: most repositories register adequate descriptive metadata but fail to sustain active harvesting over time. Journal repositories demonstrate substantially higher interoperability readiness than dataset and ETD repositories. The MIR framework offers a principled basis for evaluating national discovery infrastructure, with concrete governance implications for repository onboarding, monitoring, and differentiated intervention strategies. This study contributes to the DCMI 2026 theme of Data Integrity and Reliability, arguing that trustworthy discovery infrastructure requires verified, sustained metadata flow—not merely administrative registration.

Metadata should be understood as an interpretive practice and not just as a technical framework for describing digital objects. It holds power and facilitates community engagement. Within the digital humanities arena, metadata plays a pivotal role in narrating and recovering stories that have remained obscured or misrepresented in historical records. This raises a fundamental question: how can we more humanly describe Indigenous communities whose identities and relationships to kinship and culture have been misrepresented in colonial records? This paper examines the role of controlled vocabularies in shaping the representation of Indigenous histories in Stolen Relations’ digital humanities project, positioning metadata as a form of archival intervention. It demonstrates how iterative feedback informs the refinement, implementation, or creation of controlled vocabularies and positions metadata as a space where descriptive practices are examined and reshaped.

AIs are increasingly used to support metadata processing and investigation, which depends on whether AI-encoded meaning aligns with human meaning. However, AI encodes word meaning through distributional and contextual representations, and it remains unclear whether such representations preserve the meaning value of human system. We answer this question through Zipf’s meaning law, which links word frequency to number of word meanings. We compare multiple AI-induced meaning estimates with human-measured meaning. To quantify alignment, we propose Meaning-Zipf Deviation (MZD), which covers continuous meaning distributions and measures their divergence with reliability adjustment. Extensive experiments show that human words consistently follow Zipf’s meaning law. AI-encoded meanings also exhibit Zipfian regularities, inheriting part of the statistical structure of human language. However, AI meaning distributions remain flatter than human distributions, with lower scaling exponents and non-negligible MZD values. Larger models do not reduce this gap. AI tends to bind words to context-conditioned senses rather than preserve their broader polysemous potential.

KOS registry entries often cover multidimensional topics that resist representation by a single classmark, making multi-notation assignment a common but semantically limited approach. This study explores how selected K-KOS entries can be reclassified under the developing version of the Integrative Levels Classification (ILC) and synthesized into structured classmarks using the ILC Classmark Builder, with multi-notation assignment serving as the baseline representation. Based on three representative cases, candidate classmarks were manually constructed and examined through baseline assignment, free-facet combination, and facet synthesis, with support from the Builder for retrieval, combination, and syntactic validation, and followed by expert review of the resulting classmarks. The findings show that notation synthesis makes semantic relations more explicit and enhances the structural expressiveness of KOS representation, while also revealing that successful synthesis depends on human judgment in syntactic disambiguation, conceptual interpretation, and evaluation among alternative formulations. The study demonstrates both the feasibility and the practical challenges of this transition, and confirms the value of tool support — while underscoring that human judgment remains indispensable.

Large language models (LLMs) applied to subject cataloguing using Library of Congress Subject Headings (LCSH) tend to generate headings and strings that are syntactically plausible but policy-invalid, bypassing the controlled vocabularies and governing rules that subject collocation depends on. This paper describes a two-track experiment to address the lack of grounding in standards and policy. The first track is lc-vocabularies-mcp, a Model Context Protocol (MCP) server connecting an LLM to live Library of Congress (LC) linked data APIs, enabling real-time authority validation for LCSH and related controlled vocabularies. The second track is a conversion of the Subject Headings Manual (SHM) from PDF to structured DITA (Darwin Information Typing Architecture), designed as a machine-actionable retrieval-augmented generation (RAG) corpus. Together, the two tracks support a five-part subject cataloguing workflow in which non-parametric knowledge is supplied to Claude at each part where parametric knowledge alone is insufficient. The paper reports on the architecture of lc-vocabularies-mcp, the DITA conversion methodology, and an evaluation design that tests whether policy-grounded retrieval improves AI-assisted subject cataloguing.

This paper calls for a rethinking of metadata literacy in the age of AI. Previous discussions often defined metadata literacy as knowledge of descriptive structures or skills in creating and using metadata records. But this understanding is no longer enough – though still necessary – when AI systems create, transform, rank, summarise and recommend information at scale. In such environments, metadata is not simply about post hoc description of resources; it structures provenance, visibility, accountability, cultural representation, and the conditions under which machine outputs can be interpreted and trusted. Using metadata, metadata instruction, AI literacy, and information literacy scholarship, this conceptual paper presents a reconstructed model of metadata literacy for LIS education. The model is built on five dimensions: understanding of infrastructure, contextual description and representation, provenance and disclosure, evaluation of algorithmically mediated outputs and intervention in terms of ethics and governance. The paper also gives examples of learning tasks and assessment evidence to illustrate how the model can be applied. It argues for viewing metadata literacy not as a narrow technical specialisation but as a fundamental educational response to AI-mediated knowledge environments.

Library of Congress Classification (LCC) and Dewey Decimal Classification (DDC) are two of the most widely used knowledge organization systems in libraries, yet empirical understanding of how they align and diverge in cataloging practice at scale remains limited. This paper examines co-assignment patterns between LCC and DDC classes using 4,042,962 dual-classified bibliographic records drawn from the Library of Congress's book catalog. Through descriptive quantitative analysis and bipartite network analysis, we identify areas of strong and weak structural correspondence between the two systems. Results reveal that well-defined humanities disciplines — including law, fine arts, religion, literature, and history — exhibit high one-to-one alignment, while broader and more applied domains such as social sciences, technology, and computer and information science show markedly dispersed cross-system mappings. A structural asymmetry is also evident: LCC classes tend to map more sharply to single DDC counterparts than vice versa. Network analysis identifies seven disciplinary communities and highlights Social Sciences and Technology as key interdisciplinary hubs, while second-level classes reveal contrasting topologies — a hub-centric star structure for LCC:G and a fragmented, multi-polar constellation for DDC:6XX. These findings carry practical implications for library reclassification projects, cataloging workflows, and the reuse of bibliographic metadata in emerging technological environments.