TDWG Life Sciences Identifiers Applicability Statement Ben Richardson
TDWG Life Sciences Identifiers Applicability Statement Ben Richardson Review Manager, LSID Applicability Statement Western Australian Herbarium Department of Environment and Conservation urn: lsid: biocol. org: col: 15701
Public Review • Details at: – http: //bit. ly/p 7 Hwp • Comments open until 26 November
What is an Applicability Statement? • An advisory document rather than a specification • The “glue” that connects technical specifications • Documents how to best apply relevant technical specifications • Specifies what is to be done to conform to a specification
Review Process • Prior to April 2009 – Submissions on the Applicability Statement (AS) received from 4 peer reviewers • Late April 2009 – I agreed to be Review Manager for the LSID AS • May 2009 – I requested a revision based on reviewers comments and discussion on the Technical Architecture Group mailing list • 22 October – Revision accepted for Public Review • 26 October – Public Review begins
Applicability Statement in two sections • GUID Applicability Statement • LSID Applicability Statement
Part 1. GUID Applicability Statement • The GUID AS provides an overview – Documents the recommendations that apply to all GUID technologies • There are 6 types of GUID technology – HTTP URI – Life Sciences Identifiers (LSID) – Digital Object Identifiers (DOI) a type of Handle – Permanent URL (PURL) a type of HTTP URI – Universally Unique Identifier (UUID) – Handle System
GUID Recommendation Summary • One object must have only one GUID • Only assign GUIDs to objects for which you are the authority • Support the Semantic Web – HTTP GET resolution must be provided – Default metadata response format = TDWG Ontology RDF/XML
Part 2. LSID Applicability Statement • Documents best application of LSIDs for biodiversity domain to – Maximise LSID permanence – Explain LSID concepts of “data” and “metadata” – Support the Semantic Web
LSID Recommendation Summary — Maximising Persistence • Providers should control domain names used in LSIDs • Avoid using your organization’s domain name if it is susceptible to change – If this is problematic, apply to use TDWG’s domain – Case in point: calm. wa. gov. au dec. wa. gov. au • Don’t parse the LSID string (except to resolve it) – E. g. urn: lsid: authority. org: name: 1234: 2 a • “authority. org” may no longer own the data • “name” might be a person rather than a taxon – Get the metadata to be certain
LSID Recommendation Summary — LSID “data” vs “metadata” • LSID “data” must never change • Non-binary encoded objects should be served as LSID “metadata” – Taxon name, concept, occurrence data is LSID “metadata” – Images, audio, video is LSID “data”
LSID Recommendation Summary — Semantic Web support • Provide a HTTP proxy version of the LSID • HTTP GET must retrieve LSID “metadata” by default • LSID “metadata” must be RDF/XML – Makes biodiversity data easily accessible to Semantic Web clients that can’t resolve LSIDs
What next? • With some care, LSIDs are the best fit for biodiversity data – LSID permanence is enhanced by their independence from Internet protocol – Domain Name requirements can be overcome – Semantic Web clients can be supported – However: • HTTP URIs and PURLs are simpler to implement • DOIs already exist for publications, use them if available
What next? • Use the Applicability Statement – Decide whether LSIDs will work for you • If so, implement LSIDs many projects already have • Integrate your GUIDs with everyone else’s data
What next? • Complete the TDWG Ontology in RDF
Acknowledgements • Reviewers • GUID Applicability Statement – Author: Kevin Richards • LSID Applicability Statement – Authors: Ricardo Pereira, Kevin Richards, Donald Hobern, Roger Hyam, Lee Belbin, and Stan Blum • Lee Belbin review process
Public Review http: //bit. ly/p 7 Hwp Comments open until 26 November Merci beaucoup
- Slides: 16