This document discusses metadata mapping as a solution for metadata interoperability between systems. It defines metadata as data that describes other data and explains that metadata has building blocks like schemes and element definitions. The document also outlines some of the heterogeneities that can exist with metadata between systems. Metadata mapping is presented as a way to maintain representations of metadata and formally declare relationships between different metadata schemas to allow for metadata transformation and querying across sources.
2. Contents
• Metadata interoperability goals.
• Definition of Metadata.
• Metadata Building Blocks.
• Heterogeneities with metadata .
• Interoperability Solutions.
• Metadata Mapping.
• Conclusion.
3. Metadata
Interoperability Goals
• Metadata interoperability:
• Systems and applications can work with or use metadata across system
boundaries.
• Requirements:
• Machines need to communicate to exchange metadata.
• Machines must be able to read/process the data received.
• Machines + humans must be able to interpret the metadata correctly.
4. What is Metadata
• Metadata:
• “the sum total of what one can say
about any information object at any
level of aggregation, in a machine
understandable representation.”
• Information Object:
• “anything that can be addressed and
manipulated by a human or a
system as a discrete entity.”
5. Metadata Building Blocks
- Define Schemes, meta-meta-model,
UML, XML, SQL DLL.
- Defines how attribute like ‘title’ will be
semantically presented.
- Element Definitions.
- Content Rules.
- Descriptive Metadata elements
7. Interoperability Solutions
• Agreement on a certain model:
• Accredited institution like W3C or ISO.
• Consensus, Standard, or assurance of uniform implementation.
• Agreement on meta-model:
• Schema is defined by the same language (standard model with different
implementations)
• Reconciliation of structural and semantic
heterogeneities:
• Mapping schema languages to others’ language.
• Instance transformation (changing meta attributes to correspond)
8. Metadata Mapping
Maintaining
representations Start Find relationships and
heterogeneities
Metadata transformation.
Formal Declaration of
Answer queries over
mapping relationships
metadata sources.
9. Conclusion
• Mapping suggested over Standards.
• Standards require licensing, software
tools, personnel costs.
• Mapping has high discovery cost.
10. Bibliography
• Haslhofer, Bernhard and Wolfgang Klas. 2010. A survey of techniques
for achieving metadata interoperability. ACM Comput. Surv. 42, 2,
Article 7 (February 2010), 37 pages.
Notes de l'éditeur
\n
\n
Paper focuses on last 2 things.\n\n
\n
\n
To understand the problem, we must define the differences/heterogeneities between metadata.\nHeterogeneities that interfere with interoperability: \nStructural (model-related): \n element definition conflicts\n naming: models elements representing same element given different name\n Identification: If they have an id, having different one (sometimes no id, only name exists)\n Constraints: datatype for example.\n Domain Representation\n Abstraction level: domain representation conflicts, entities arranged into different generalization hierarchies, or distributed into different model elements\n Multidimensional correspondences: Conflict in the multiple relationships drawn up.\n Meta-level discrepancy: information with in different elements (like naming)\n Domain coverage: one model has data x, the other does not.\nSemantic: (language differences in schema)\n Domain conflicts: different expressiveness of languages\n Terminological: naming: synonyms and homonyms\n Scaling/Unit Conflicts: different measurement units\n Representation: format of date value for example\n\n