In order to discuss the development of the taxonomy, one first has to consider the ERM. The taxonomy is inexorably linked with the ERM, and also must take into account the implementation models based on the ERM3 – the MODAF meta-model and the web services specification for MARS. The implementation format for MODAF data exchange is to be XMI, the OMG’s model interchange standard. The XMI content will be constrained by a UML profile which is founded on the ERM. XMI is an XML specification. The use of XMI for MODAF tool interchange does not necessarily dictate that XML be used for the taxonomy, but there are several reasons to use XML:
• XML permits simple referencing between files using URIs, and implementations are capable of navigating such references. In the case of the W3C OWL XML specification, this enables the taxonomy to be distributed and also allows individual teams to extend the “standard” taxonomy for their own use.
• There is an abundance of reliable XML tools on the market, many are even free
• There is a strong IT skill base in XML – virtually every programmer has worked with it at some time
• XML Schema is an effective data modelling language for XML data, and has most of the capabilities of an entity-relationship model
• XSLt, the XML transformation language, allows XML data to be presented in a number of different formats, which need not be XML. Hence, an XML taxonomy can be presented in a tabular format for inclusion in the MODAF documentation, or as a
navigable HTML tree
Replaced/Superseded by document(s)
|File||MIME type||Size (KB)||Language||Download|
|Implementation Strategy for the MODAF Taxonomy.pdf||application/pdf||101.17 KB||English||DOWNLOAD!|
This paper proposes a way forward for developing the MODAF1 taxonomy that makes use of semantic web standards under development by the World Wide Web Consortium (W3C). The proposed solution is to use OWL (the Web Ontology Language) which was originally intended for defining structures and relationships that lend meaning to information on the
web. In its simplest form, however, OWL is rather good for defining classification hierarchies– taxonomies.
The OWL approach offers the following advantages:
- It is XML, as specified by the MOD for all future data exchange projects
- It allows taxonomies to be developed and hosted at different sites, with links such that the set of taxonomies can behave as one integrated taxonomy
- There are a wide variety of free authoring and development tools
- It is a standard from the same stable as HTML, XML and Web Services
- It can be easily converted into formatted text for publication
- It is already being trialled in other parts of the MOD for taxonomy and ontology development
This paper does not address implementation architectures for the taxonomy, but does suggest possible technologies that should be considered – including web services to provide applications with online access to taxonomy.