Metadata Features in the Domain of Frameworks

Metadata Features in the Domain of Frameworks

Beyond using metadata for page organization (HTML/XHTML), display (CSS/XSL), or description (DC, etc) there is an additional need to use metadata to describe the metadata schemas used in creating an electronic resource.

Shelley Powers (2003) in Practical RDF suggests an analogy with relational databases which states that the database software program which allows the same table, field, record, value structure to be used for any configuration of unique labels, values, and relationships (ie. Persons: Name: 01= “Heather”) is basically how Resource Description Framework (RDF) provides the underlying structure which allows the same element, attribute and value structure (ie. ) to be described and interrelated (¶5.2.2.). RDF is like MS Access, not Clients.mdb.

The issue of frameworks or containers for metadata schemes is also closely tied to a longstanding desire to create a semantic or ontological web (Lagoze, Lynch & Daniel, 1996, p.2). The semantic web is working towards a situation where data which is now only processed or read, can also be aggregated and interpreted automatically by software applications. Powers (2003) states that RDF is an, “overlay of additional constraints that allow for easier interchange, collection, and mergence of data from multiple models” (p.29). This statement shows the fundamental interest in the domain of Frameworks. When the constraints of the RDF data model are taken a step farther with semantic developments such as Web Ontological Language (OWL), they allow software applications to both compile and read data and make inferences about the relationships between data.

Hausenblas, Slany & Ayers (n.d.) state that most data on the web is non-compliant to the RDF model, but show that this is not a problem because data can be converted from HTML, syndicated feeds, web APIs, microformats, and some RDBMSs through the use of Gleaning Resource Descriptions from Dialects of Languages (GRDDL) (¶3.1). DuCharme (2007) afirms the improvement of GRDDL and Semantic Web (SW) Agents for aggregating data and metadata, over current ‘screen scraping applications’ (¶Intro). It is these movements towards automated data aggregation which are encouraging further development of RDFa and OWL.

Representative Schemes that are used in the Domain of Frameworks

•Dublin Core (DC) & Warwick Framework •Dublin Metadata Workshop put aside resource description in order to achieve results, Warwick Workshop provided containers and packages for aggregating metadata sets (Lagoze, Lynch & Daniel, 1996, p.9).

•Meta Content Framework (MCL) •Research project at Apple Comp. from 1995-7, R.V. Guha later helped to create RDF at Netscape (Meta Content Framework, n.d.).

•DARPA •Kahn/Wilensky Framework for a distributed information infrastructure was used by the Warwick Framework (Lagoze, Lynch & Daniel, 1996, p.23)

•Agent Markup Language (DAML) + Ontology Inference Layer (OIL) built on XML & RDF allows software programs to interpret relations between data (The DARPA Agent Markup Language Homepage, 2006).

•Resourse Description Framework (RDF), RDF:XML & RDFa •RDF is a data model which creates graphs of relationships using triples (subject-predicate-object), this allows data to aggregated from various sources

•RDFa is a simplified version of RDF/XML which can be embed in XHTML

•Web Ontological Language (OWL) •allows more constraints on RDF relationships, thereby allowing software agents to make more inferences about data (Powers, 2003, p.229)

•Semantic Markup, Ontology & RDF (SMORE) editor (Powers, 2003, 248).

Strengths & Weaknesses

The following table represents some significant strengths and weakness of the major schemes in the domain of Frameworks:

Scheme Strength Weakness

DC •

• •

•

Warwick •

• •

•

DAML •

• •

•

MCL •

• •

•

RDF •a simple & familiar premis of subject-predicate-object for explaining relationships

•allows relative description of complex & infinite resources •cannot