A Hybrid Declarative/Procedural Metadata Mapping Language Based on Python Export

@incollection{citeulike:1362955, abstract = {The Alexandria Digital Library (ADL) project has been working on automating the processes of building ADL collections and gathering the collection statistics on which ADL's discovery system is based. As part of this effort, we have created a language and supporting programmatic framework for expressing mappings from XML metadata schemas to the required ADL metadata views. This language, based on the Python scripting language, is largely declarative in nature, corresponding to the fact that mappings can be largely—though not entirely—specified by crosswalk-type specifications. At the same time, the language allows mappings to be specified procedurally, which we argue is necessary to deal effectively with the realities of poor quality, highly variable, and incomplete metadata. An additional key feature of the language is the ability to derive new mappings from existing mappings, thereby making it easy to adapt generic mappings to the idiosyncrasies of particular metadata providers. We evaluate this language on three metadata standards (ADN, FGDC, and MARC) and three corresponding collections of metadata. We also note limitations, future research directions, and generalizations of this work.}, author = {Jan\'{e}e, Greg and Frew, James}, citeulike-article-id = {1362955}, citeulike-linkout-0 = {http://dx.doi.org/10.1007/11551362_27}, doi = {10.1007/11551362_27}, journal = {Research and Advanced Technology for Digital Libraries}, keywords = {adl, metadata--mapping, vocabulary-mapping}, pages = {302--313}, posted-at = {2007-06-04 15:45:04}, priority = {0}, title = {A Hybrid Declarative/Procedural Metadata Mapping Language Based on Python}, url = {http://dx.doi.org/10.1007/11551362_27}, year = {2005} }

TY - CHAP ID - citeulike:1362955 L3 - citeulike-article-id:1362955 N2 - The Alexandria Digital Library (ADL) project has been working on automating the processes of building ADL collections and gathering the collection statistics on which ADL’s discovery system is based. As part of this effort, we have created a language and supporting programmatic framework for expressing mappings from XML metadata schemas to the required ADL metadata views. This language, based on the Python scripting language, is largely declarative in nature, corresponding to the fact that mappings can be largely—though not entirely—specified by crosswalk-type specifications. At the same time, the language allows mappings to be specified procedurally, which we argue is necessary to deal effectively with the realities of poor quality, highly variable, and incomplete metadata. An additional key feature of the language is the ability to derive new mappings from existing mappings, thereby making it easy to adapt generic mappings to the idiosyncrasies of particular metadata providers. We evaluate this language on three metadata standards (ADN, FGDC, and MARC) and three corresponding collections of metadata. We also note limitations, future research directions, and generalizations of this work. JF - Research and Advanced Technology for Digital Libraries EP - 313 TI - A Hybrid Declarative/Procedural Metadata Mapping Language Based on Python SP - 302 KW - adl KW - metadata--mapping KW - vocabulary-mapping AU - Janée, Greg AU - Frew, James PY - 2005/// UR - http://dx.doi.org/10.1007/11551362_27 ER -