GBIF Vocabulary Server

From Terminology Sandbox Wiki
Jump to: navigation, search

GBIF Vocabulary Server

The GBIF Vocabulary Server (Harman et al, 2009) is developed using Drupal 6 and the Scratchpads (Smith et al, 2009) of the ViBRANT project and was first released in 2009. The recently established Vocabulary Management Task Group (VoMaG) has initially recommended to build Darwin Core extensions and controlled value vocabularies for the GBIF infrastructure based on the reuse of terms and concepts from RDF vocabularies. The GBIF Vocabulary Server is designed to provide a software tool for collaborative development of Darwin Core extensions and controlled vocabularies.

http://kos.gbif.org/images/illustrations/GBIF_Vocabularies_extensions.png

The terms included in Darwin Core extensions and controlled vocabularies developed at the GBIF Vocabulary Server should preferably be backed-up by a machine readable resource that better supports the "reuse" of such terms by other resources. The extensions and vocabularies under development at the GBIF Vocabulary Server is rather intended with a more limited scope to provide a list of controlled values for data exchange tools such as the GBIF IPT. These controlled values are rather instances of the source terms for a particular use case (here aimed at the Darwin Core Archive format) than the actual source definition of the term. The GBIF Vocabulary Server is probably not an appropriate software tool to manage such source term definitions (eg. no RDF output).

We see the GBIF Vocabularies as a tool to support collaborative development of the extensions and controlled vocabulary resources to be published at the GBIF Resources Registry (http://rs.gbif.org) when these resources are finalized and frozen from further modifications. The extensions and controlled vocabularies would be intended for the consumption by applications such as the GBIF IPT (and perhaps also by the Scratchpads...?). The extensions and controlled vocabularies would therefore be immutable once finalized and published at the GBIF Resources Registry/Repository (http://rs.gbif.org). Extensions and controlled vocabularies would therefore not be updated and modified, only deprecated and replaced by a new replacement resource. Published extensions and controlled vocabularies could include a version number, possibly as a postfix to the file name.

Our current thinking is that all concepts included in such a flat list of controlled values would be "reused" from a vocabulary of terms (or from an ontology). Each term should have or get a URI to persistently identify the term so that the term can be "reused". Our current thinking is that all terms should as a guideline recommendation be declared (with URI and definition) in a basic and flat vocabulary of terms and "reused" from this RDF vocabulary by resources such as more elaborate ontologies (eg. OWL resources) and in application resources such as the extensions and controlled vocabularies developed at the GBIF Vocabulary Server. The OWL ontologies would in this scenario, as a best practice guideline, not provide the normative definition of the basic term, but rather an instance of the term based on the normative definition as provided by the flat RDF vocabulary resource. The rationale is that the more elaborate ontology (OWL) resources would be too complex for most users to easily understand and consume (with consume we mean "reuse" terms); and that using OWL resources for the source declaration of terms would provide a bottleneck for the "reuse" of the terms. The rationale is further that it might prove impossible for the community to agree on ONE semantic ontology description of such concepts - and therefore that multiple context dependent and valid ontology resources might be appropriately developed to describe opinions declaring the richer semantics of the term. These ways of thinking are still only best practices guidelines UNDER DEVELOPMENT by the proposed TDWG Vocabulary Management Task Group (VoMaG). The target is here to maximize the potential for reuse of the term (URI).



References

Harman KT, R Hyam, DP Remsen (2009). Vocabularies - Managing Them. Proceedings of TDWG 2009 Available at http://www.tdwg.org/proceedings/article/view/605

Smith VS, SD Rycroft, KT Harman, B Scott, and D Roberts (2009). Scratchpads: a data-publishing framework to build, share and manage information on the diversity of life. BMC Bioinformatics 10 (Suppl 14) p. S6. DOI:10.1186/1471-2105-10-S14-S6. Available at http://www.biomedcentral.com/1471-2105/10/S14/S6

Drupal (2012). Drupal Open Source CMS. http://drupal.org/

Havlik D (2011). Building Environmental Semantic Web Applications with Drupal. Environmental Software Systems. Frameworks of eEnvironment 359 p. 385-397. DOI:10.1007/978-3-642-22285-6_42. Available at http://www.springerlink.com/content/ktw57451h0155214/