[ View/Print a PDF Version of this document.]

A data dictionary is a set of terms, with their definitions, used in a computerized system. Some data dictionaries are structured, with terms related through hierarchies and other relationships: structured data dictionaries are derived from ontologies. An ontology combines a data dictionary with a logical data model, providing a consistent and logical world view. It differs from the traditional taxonomic approach to knowledge representation in that it does not follow a rigid/parent child hierarchical structure (terms may inherit meaning from more than one parent) and a more complex relationship is maintained.

An interoperable data dictionary contains terms from different computerized systems or metadata schemes, and shows the relationships they have with one another in a formal way. The purpose of an interoperable data dictionary is to support the use together of terms from different systems.

The DOI System uses an interoperable structured data dictionary. The IDF's Data Dictionary (IDD) is a structured ontology, in which meaning, once it has been defined, can be passed on from one term to another by logical rules of association such as inheritance and opposition. In recognition of the great diversity and complexity associated with multimedia content, it is designed to represent as many different specialisations of meaning as its users require, and to show their relationships in a structured way in order to support the mapping and transformation of terms between different schemas and systems.

The IDD is currently in use as an internal tool. However it is built on exactly the same principles as the public ISO/IEC 21000-6 Dictionary, the MPEG-21 RDD (http://www.iso21000-6.net), for which the IDF is the ISO Registration Authority (IDD and RDD are instantiations of the same methodology). The ISO/IEC 21000-6 dictionary is designed as a component of MPEG-21 to support the MPEG REL (Rights Expression Language), but is based on a generic methodology. IDF also intends to promote the adoption of ISO/IEC 21000-6 as the registry for metadata semantics for standard content identifiers, and to this end IDF is investigating how the IDD and RDD can converge.

The DOI^® Data Model embraces both this data dictionary and a framework for applying it. The data dictionary component is designed to ensure maximum interoperability with existing metadata element sets; the framework allows the terms to be grouped in meaningful ways (DOI^® Application Profiles) so that certain types of DOI^® names all behave predictably in an application through association with specified Services. This provides a means of integrating the features of Handle resolution with a structured data approach. DOI names need not make use of this data model, but it is envisaged that many will: any DOI name intended to allow interoperability (i.e. which has the possibility of use in services outside of the direct control of the issuing Registration Agency) is subject to DOI System Metadata policy, which is based on the registration of terms in the IDD.

IDD exists to solve an obvious but difficult problem: how does one computer system know what the terms from another computer system mean? (If A says "owner" and B says "owner", are they referring to the same thing? If A says "released" and B says "disseminated", do they mean different things?) The data dictionary provides a way of describing relationships between terms, and confirming agreement about this, so that A or B (or anyone else) can make use of one another's metadata with confidence and in a highly automated way.

• It may be assumed that A knows what he means, and B knows what she means; but they may be assuming totally different concepts from each other. This is true of any term: concepts (e.g. "depression" as understood by the mental health, economics, and meteorological communities), roles (e.g. "publisher" as understood by music, newspaper and book industries), and physical formats (e.g. "folio" as understood by the bookkeeping, legal, and printing communities).
• The only way of unambiguously deciding if one term means the same as another, irrespective of what it is called, is by sharing a single frame of reference: a structured ontology (an explicit formal specification of how to represent the entities that are assumed to exist in some area of interest and the relationships that hold among them) with an underlying model which allows the generation of consistent new relationships, and a method of recording the agreement between the parties whose terms are included in it.
• Mapping terms from one scheme to another is not always straightforward. Terms may be expressed in different parts of speech and tenses, and meanings are often "contextual" (e.g. the same term "Identifier" in one place may mean "Product Identifier" and in another place mean "Party Identifier", within the same scheme). The IDD is designed to support these levels of complexity and contextuality.
• Whilst there are many ontology approaches, few address the semantic interoperability requirement; IDD uses the most well-developed method for this.

• Whilst "crosswalks" can be constructed to compare terms in any two metadata schemes, the total number of such crosswalks grows much faster as the number of schemes grows linearly (N schemes require (N/2)(N-1) mappings). The existence of one dictionary reduces this to N mappings, one for each scheme.
• Bilateral agreement between dictionary and scheme ensure that the existence of agreed mapped terms enables extensibility -- mapping to another scheme -- without reference to the originators of each scheme. Such mappings will increasingly be computable and thus automated.

The DOI System provides a ready-to-use system of several components: a specified numbering syntax, a resolution service (based on the Handle System), a data model system (including the indecs Data Dictionary), and policies and procedures for the implementation of DOI names through a federation of Registration Agencies.

• One component of the DOI System is the IDD. Its implementation in the DOI System has been supplemented by expanded technical infrastructure and features specific to DOI System applications.
• The IDD is the repository for all data elements and allowed values used in DOI System metadata declarations; it is the heart of the DOI^® Data Model process. The IDD enables the definition and ontology of all metadata elements to be available to all RAs, and provides the necessary mappings to support metadata integration and transformations required for data interchange between RAs who require it.
• The functions of the IDD are to support:
• interchange of metadata between Registration Agencies using standard messaging (RMDs);
• automated use of Kernel metadata declarations;
• interoperability between Application Profiles by common semantics for DOI System Services and ResourceTypes.

DOI System development aimed to use existing or developing standards, rather than develop unique tools. The IDD has a long history which began around the same time as the DOI System, and is used in several major activities.

• The IDD is built using methodology from the indecs (interoperability of data in e-commerce) framework, an influential multimedia metadata project from 1998-2000 backed by groups from the content, author, creator, library, publisher and rights communities, which pioneered a model of event-based metadata as a solution for integrating rights. Indecs in turn drew on earlier work from the library community (FRBR) and music community (CIS).
• Subsequent versions of the methodology have been used as the basis for the DOI System, for the MPEG-21 Rights Data Dictionary (RDD), and heavily influence the current development of messaging systems for the publishing industry (ONIX) and music industry (MI3P).
• The methodology for the development of such data dictionaries was initially codified during the development of the MPEG 21 Rights Data Dictionary by the CONTECS:DD consortium (which included the IDF).
• The International DOI Foundation (IDF) and EDItEUR (the International Group for electronic commerce in the book and serial sectors) harmonise ONIX and DOI System metadata through the use of this common data dictionary (and welcome collaboration with others adopting a similar approach).
• The methodology has been validated against the W3C ontology language OWL-DL.
• The methodology for constructing interoperable Data Dictionaries which underlies IDD is in use commercially as Ontologyx. Ontologyx is a comprehensive data meta-model, an analytical tool and an ontology; a computing platform to articulate the tools is under development. In some ways Ontologyx has the same relationship to the DOI System as the Handle system: it is a component, but also used in other ways outside the DOI System.

The semantic analysis underlying the IDD is independent of any implementation model.

• It was fundamental to indecs (despite "e-commerce" in its name) that it had no inherent commercial model, and it remains so for all the work that has followed it. It is just as critical to be able to say "this is not subject to copyright" as to say the opposite; one of the problems any "non-commercial" person or organization has is to be able to state that something is freely available and under what circumstances. A broad ontology supporting rights expressions must be able to support any kind of expression of any kind of right, agreement or licence or any terms or none. Most organizations have the need for both freedom and protection of intellectual property in different contexts. The IDD is not solely a tool for intellectual property as "commercial property" but is neutral as to the intellectual property regime being used.

• An RA must be capable of producing a Kernel Metadata Declaration for each DOI name, using a small set of standardised terms from the IDD.
• Metadata exchanged between RAs supporting DOI System services should be exchanged using an agreed message format, a DOI^® Resource Metadata Declaration ("RMD").
• Proprietary terms (data elements and values) used by RAs in Kernel and Resource Metadata Declarations should be registered in the IDD.
• RAs are otherwise free to use any metadata schemes for gathering, storing or disseminating metadata.

Every term entered into the IDD carries information on its status as to origin and mapping agreement.

• If a mapping from scheme A to IDD is reciprocally agreed with the governance authority of scheme A, then the dictionary can embody an assured mapping which will enable users of the dictionary to interpolate mappings from their own schemes, through IDD, to scheme A and know that this will be considered authoritative by scheme A. Such mappings will be dynamically updated as new versions of schemes are made available.
• Any RA contributing terms to the IDD can specify who is allowed to see or specify their own terms.
• Any public terms are accessible to all IDF agencies; e.g. ONIX, DOI System terms from the kernel and RMD, and the MPEG21 RDD.

• It is no more a requirement to know the details than it is for the design of a web page to require one to read all the underlying internet protocol RFCs.
• A fundamental role of the IDF with the IDD is to provide assurance to users that the work has been peer-reviewed, tested in practical implementations, and is based on sound principles. The IDD structure has been tested in several implementations, public and private, including the MPEG 21 RDD evaluation, the DOI System, and mapping to the W3C ontology language OWL-DL.
• Some key features of the IDD methodology are:
• Extensible and granular: the ontology extends its core "Context Model" to whatever level of detail and granularity is required.
• Multiple, different, specialized views are available: these include the Resource Model, based on ten core data elements, and the Rights Model, based on a set of specialized Contexts.
• Local terms: an RA can add all its own local data elements and names into the ontology, and use only those terms it needs. It can include different terms from different internal systems and map them together.
• External terms: it incorporates external and standard schemes such ISO territory, currency and language codes, and sector specific external schemes, allowing them to be treated seamlessly alongside local terms.

IDD and the MPEG-21 Rights Data Dictionary (ISO/IEC 21000-6)

• IDD and RDD are instantiations of the same methodology. The ISO/IEC 21000-6 dictionary is designed as a component of MPEG-21 to support the MPEG REL (Rights Expression Language), but is based on a generic methodology. IDF also intends to promote the adoption of ISO/IEC 21000-6 as the registry for metadata semantics for standard content identifiers, and to this end IDF is investigating how the IDD and RDD can converge.
• As currently specified, all terms in the RDD are mapped into the IDD; that is, RDD is one of the authorities specifying terms within IDD. RDD is therefore a sub set of IDD. It is conceivable that some future RDD terms might be added to the RDD which are not within IDD; the two Data Dictionaries would then overlap and share some common terms.
• The RDD Registration Authority will establish an automated web based look-up system. It is anticipated that access will be granular, with levels of privilege being established to ensure that those with authority to access the Dictionary are able to view what is appropriate while private TermSets, if they exist, are kept confidential.

[ View/Print a PDF Version of this document.]