The 10th International Forum on Metadata
Registries met in New York City,
Using the library and information science (LIS) world to exemplify interoperability and standards is somewhat misleading. In the world at large, data are commonly maintained in locally designed databases. Indeed, it is not uncommon for an organization to maintain conceptually the same data in multiple databases, using different (and perhaps incompatible) structures and different terminology. Standards designed to achieve interoperability usually are not imposed on the underlying databases, but only on how the databases are described. Contrast this with the sharing of bibliographic data between two institutions, even where only one uses MARC-encoded data: The functionality supported by the two bibliographic databases would probably still be very similar, such that a mapping between the two database structures could probably be readily developed. (A crosswalk between Dublin Core and MARC is a salient example.)
The underlying difference between the bibliographic and non-bibliographic data contexts explains why non-bibliographic metadata are not classed with bibliographic metadata in 025.3 Bibliographic analysis and control. Moreover, the Forum on Metadata Registries considers databases from the “narrowly technical” perspective of 005.74 Data files and databases (see Manual at 025.04, 025.06 vs. 005.74). Interoperability is mentioned in a class-here note at 004.6 Interfacing and data communications, which, although not directly applicable to data interoperability, shows that interoperability should be classed with the technology that enables it. That’s where the metadata registry comes in, classed in 005.742 Data dictionaries and directories (where data interoperability is also classed). Metadata standards for non-bibliographic data are classed in 005.740218 (built with 005.74 Data files and databases plus T1—0218 Standards), while terminology (which in this context refers to all of terminology science) is classed in 401.4 Language and communication, the number for interdisciplinary works on terminology.
Posted by: Josepheb | July 31, 2007 at 12:30 PM