論文

This study aimed (a) to understand the actual state of reference relations in RDA guidelines and instructions, expressed as RDF data, and (b) to grasp its characteristics at a global level by applying network analysis to those reference relations. The reference relations were examined and extracted from four datasets–RDA guidelines, instructions on entities, and those on Manifestation and Work elements. Furthermore, they were aggregated by source and destinations. The characteristics of the four graphs formed by the reference relations were analyzed using measures such as centrality.

Purpose: The purpose of this study is to confirm the feasibility of RDF data conversion by presenting necessary considerations and options and then identifying appropriate options, with the intention of converting NCR2018 descriptive rules into RDF data. Method: In order to be suitable for effective utilization as linked data, (1) we considered possible options for expressing individual rules as an RDF class and expressing various relationships between the classes; (2) based on the choices adopted, the conversion to RDF data was performed for the rules in the three chapters; and (3) mutual references between the RDF-expressed NCR rules and metadata in line with the rules were examined. Results: (1) In addition to the RDF class being determined based on the rule numbers of NCR2018, it was appropriate to adopt a division unit corresponding to each instruction applied independently under a given rule. Each rule and division unit was given a URI as an RDF class, and various relationships between them were properly expressed including hierarchical relationships, reference relationships, those between a rule and its alternatives, etc. (2) RDF data from the three chapters of NCR2018 were prepared by performing mechanical conversion to RDF data with manual correction. It was confirmed that RDF data can be converted without major barriers using the adopted policy and choices. (3) It was shown that mutual references between the NCR rules and metadata created under the NCR rules can be implemented, leading to a possible utilization of RDF-expressed NCR rules.

Among metadata-related standards, data content standards like metadata guidelines and instructions for creating metadata still remain in legacy forms. This study investigates a way to transform data content standards to linked data (LD) through conversion from other formats, while referring to the proposed layered framework. Under the basic policies for making LD on which this study is based, several principal matters were examined: (a) defining units to be assigned with Universal Resource Identifiers (URIs), (b) defining relationships among the instructions with URIs and (c) expressing instructed content in instructions properly with certain Resource Description Framework (RDF) properties. With the proper choice(s) for each matter, some actual standards were converted to LD: Resource Description and Access (RDA) and Dublin Core User Guide. The results showed that the adopted way of transforming data content standards to LD is valid and proper, and the resultant LD would be expected to be utilised in various manners.

We examined how data provenance and additional information of element values including nomens, and administrative information on the metadata should be modeled and represented in the Resource Description Framework (RDF) for linked data of library catalogs. First, we classified such information types into categories and organized the combination with recording-units, i.e., a description statement or description set. Next, we listed the appropriate RDF representation patterns for each recording-unit. Then, we reviewed the methods to examine such information in Resource Description and Access (RDA) in RDF, BIBFRAME, and Wikidata, and pointed out the issues involved in them.

Purpose: For reconstructing the JAPAN/MARC (JPMARC) bibliographic records created by the National Diet Library in accordance with the Nippon Cataloging Rules, 2018 ed. (NCR2018) into metadata that is composed of NCR2018 entities and their elements, the purpose of this study is to develop the necessary schema-level mapping and tools that make use of it, and then to perform data conversion. This will contribute to a considerable increase of the NCR2018 data examples available as learning content. Methods: We developed our own mapping to convert and reconstruct the JPMARC data while basically considering all the data elements in the JPMARC format for mapping. Based on this policy, appropriate methods were introduced to solve the individual mapping issues. In addition, we implemented a tool to perform data conversion based on the mapping. Results: The characteristics of the mapping and conversion tool are as follows: 1) separate the mapping from the subsequent adjustment process, 2) introduce qualifiers to NCR elements in the mapping and utilize them in the subsequent adjustment process, 3) adopt the JPMARC data element unit as a combination of "field tag, indicator, and subfield code" and in some cases "material designation" is also combined for more accurate mapping, 4) mapping is divided into two levels with priority, and 5) the provenance information of each value is recorded. Furthermore, the mapping was divided into patterns of cardinality, and a detailed examination of the remaining issues was conducted. The implemented tool performed data conversion of 87,000 records, thereby resulting in a vast amount of metadata in line with the NCR2018 structure.

This paper proposes a framework to understand and review the present and possible RDA Vocabularies in RDF after the 3 R Project, and also examines RDA Vocabularies under this framework to confirm the validity and usefulness of the framework. Further, it proposes an alternate for the vocabularies at each layer of the framework. (1) The proposed framework consists of four layers: (a) layer 1 constructs RDA Vocabularies in line with the RDA conceptual model; (b) layer 2 constructs vocabularies in RDF representation that accommodate the recording methods prescribed in RDA for individual elements; (c) layer 3 constructs them by reconstructing the relationship elements; and (d) layer 4 accommodates additional ways to enforce the structural representation of RDA metadata. RDA Vocabularies are examined at each layer to demonstrate the framework’s validity and usefulness. (2) Alternate vocabularies in RDF are proposed at layers 2 and 3 of the framework. Those proposed vocabularies could be useful, but the examination of their usefulness is limited. Extensions at layer 4 would also be useful for the schema for RDA metadata creation. Therefore, the proposed framework is valuable for understanding and reviewing RDA Vocabularies, while some points need further consideration.

© Shoichi Taniguchi. This article examines BIBFRAME 2.0 from the viewpoint of a metadata schema for diverse descriptive metadata about bibliographic resources, that is, whether the present BIBFRAME is suitable for the exchange and sharing of such metadata. First, an overview of the BIBFRAME model built with core classes and properties is provided. Secondly, under the multi-class structure of BIBFRAME, a way of defining the domain of the properties in BIBFRAME is examined. Thirdly, possible other ways of defining the properties’ domain are discussed and compared with that of BIBFRAME. Finally, a more suitable way of defining the domain is explored.