LSID Based Integration Ants Demonstration

Coordinator(s):

Rod Page

Participants:

Description

This is sort of what I've done. What I'm aiming for is something like the diagram below, where diverse sources of data are integrated into a triple store. The two key issues of relevance to TDWG GUID are the use of GUIDs to identify digital objects, and the use of metadata as RDF.

For details on how this was done, please see my blog (I find it a lot easier to blog stuff as it gets done, rather than edit a Wiki).

Technical Information

• URL for prototype user interface: http://linnaeus.zoology.gla.ac.uk/~rpage/ants/
• LSID authority(ies): ncbi.nlm.nih.gov.lsid.gla.ac.uk (served from Glasgow), lsid.gla.ac.uk
• LSID namespace(s): pubmed, taxon, gi, predicates
• Hardware platform: Apple iBook G4 for prototyping, Generic PC server (AMD 1.4 GHz CPU, 512 Mb RAM, 130 Gb hard disk)
• Server platform: Fedora Core 4, Apache 2.0.54, MySQL? 4.1.11, PHP 5.0.4
• LSID Software stack used: Modification of Roger Hyam's PHP code
• RDF/OWL ontology used for metadata: Dublin Core, PRISM, FOAF, uBio, home grown
• Approximate number of LSIDs stored: Around 360,000 triples
• Benchmarchs: Too embarrassing to mention
• Other resources:
  • Google Maps API http://www.google.com/apis/maps/
  • Connotea RSS feed http://www.connotea.org/
  • 3Store 2.2.22 http://sourceforge.net/projects/threestore/
  • GraphViz? http://www.graphviz.org
  • numerous Perl scripts
  • and a C++ program to populate triple store http://iphylo.blogspot.com/2006/05/ants-rdf-and-triple-stores.html

Roadmap, Milestones, Timeline

See http://semant.blogspot.com/.

Discussion, Implementation Issues

For background please see http://semant.blogspot.com/.

The server has two IP's, and this causes Google Maps some confusion -- if you get a warning about the Google Maps API key, just ignore it.

Lessons Learned, Conclusions, Recommendations

Don't try this at home...

Categories