Coming to Terms with SPM
Robert A. Morris
Abstract
The Species Profile Model (SPM) is a proposed ontology, defined in terms of the emerging new TDWG ontology, centered architecture based on the Resource Description Framework (RDF) language of the World Wide Web Consortium. Although RDF has a representation, one of several, in XML, it and related languages attempt to capture semantics, not only syntax, of data. By contrast, XML-Schema, the principal language for constraining XML for data validation purposes, only constrains syntax. RDF does so by making it easier to express in standard ways the relationships among concepts meant to be addressed by particular data and allow such data to be compared and integrated without requiring transformation into a particular syntactic form. Many intuitions arising from syntax-only constraint languages like XML-Schema give little or no insight into semantic questions (e.g., when is an attribute value, or even its applicability, inherited from objects to subobjects, when are two attributes mutually exclusive, etc.). Consequently, the learning curve for RDF technology can be steep, all the more so without good tools for producing, editing, and visualizing information expressed in RDF and related technologies. This presentation will describe and demonstrate two such tools, Protege and Altova SemanticWorks.
Although the ultimate point is to generate SPM programmatically, the use of tools like Protege and Altova SemanticWorks can ease the RDF learning burden, and of course serve as a debugging aid, if properly configured. Both tools let users interactively tinker with SPM instances in ways that help one better understand the SPM design and clarify the use of it.
No species profile will live in isolation, so one question these tools help you consider is whether to attempt to design an ontology (described in the RDF-based Web Ontology Language, OWL) for descriptions of your group of interest, or whether to be content simply with RDF consistent with the TDWG architecture. To a certain extent, this speaks to the question of whether your goal is to support machine reasoning about the species described by your SPM, or simply to support mashups as a means of integrating data for human analysis. Time permitting, I will discuss how such RDF editors can be used as visualization and debugging aids for RDF produced by screen scraping tools generated by the MIT SIMILE suite(http://simile.mit.edu/wiki/Main_Page).
Although the ultimate point is to generate SPM programmatically, the use of tools like Protege and Altova SemanticWorks can ease the RDF learning burden, and of course serve as a debugging aid, if properly configured. Both tools let users interactively tinker with SPM instances in ways that help one better understand the SPM design and clarify the use of it.
No species profile will live in isolation, so one question these tools help you consider is whether to attempt to design an ontology (described in the RDF-based Web Ontology Language, OWL) for descriptions of your group of interest, or whether to be content simply with RDF consistent with the TDWG architecture. To a certain extent, this speaks to the question of whether your goal is to support machine reasoning about the species described by your SPM, or simply to support mashups as a means of integrating data for human analysis. Time permitting, I will discuss how such RDF editors can be used as visualization and debugging aids for RDF produced by screen scraping tools generated by the MIT SIMILE suite(http://simile.mit.edu/wiki/Main_Page).