39 lines
2.8 KiB
Plaintext
39 lines
2.8 KiB
Plaintext
%META:TOPICINFO{author="JimGraham" date="1257957022" format="1.1" version="1.7"}%
|
|
%META:TOPICPARENT{name="TDWG2009DataIntegration"}%
|
|
---+ Data Integration Nitty Gritty Page
|
|
|
|
If you want to contribute info to the Data Integration Theme Summary or nominate someone for a Golden T-Shirt edit this page or mail roger(a)hyam.net.
|
|
|
|
This is a wiki. If you see an error correct it don't moan :)
|
|
|
|
---++ Criteria for Good Integration
|
|
* "This is how you use/access our data"
|
|
* "People will be able to extend our system by..."
|
|
* "Anyone will be able to ..."
|
|
|
|
---++ Criteria for Bad/Non Integration comments
|
|
* "This is our system you can click on this and that"
|
|
* "We are saving the planet"
|
|
|
|
---++ Talk Spotting
|
|
|
|
*Andy Jaris* "Value of a coordinate" - good examples of gathering data from many sources and doing something useful with it.
|
|
Expressed need for communicating back but didn't suggest mechanism.
|
|
|
|
"Everything is online & freely available." == all the right things...
|
|
|
|
*Samy Gaiji* - good metaphor of plumbing and stressing of decentralisation.
|
|
|
|
*Michael Browne* - GISIN - very open network. Good use of logo
|
|
|
|
*Dave Roberts* - Cheeky use of logo at the very end. Did he really mention linking to other projects?
|
|
|
|
*Catalog of Life* - We will be able to integrate species occurrences from many sources now that we have web services to resolve scientific names to accepted names.
|
|
|
|
---++ Poster Spotting
|
|
|
|
*Roger Hyam's talk* We have indeed been hearing about the promise of the Semantic Web for a long time, but perhaps it would be worthwhile to become more precise as to what exactly we hope for relative to "data integration". Many use cases in the ecological realm, for example, include not only discovering relevant information, but resolving how these heterogeneous data resources can then be aligned, combined, converted, and ultimately joined or unioned to provide a wider range of co-variate information or enlarged sample size to inform scientific analyses and models. This can require resolving some fairly subtle semantics, that challenge current ontological approaches. I'm very hopeful about the capabilities of ontologies, and these are becoming more and more powerful (e.g. recent OWL 2.0 upgrade). In the meantime, however, just having a controlled vocabulary will immediately provide us with lots of value when discovering and confederating data (especially very stereotyped data). We should be sure that these vocabularies are ontology-ready, using sound informatics structures. Then these nascent ontologies will be able to leverage advanced inferencing capabilities through the interact process of revising/extending our ontologies, along with technical progress in the capabilities of reasoners, specifications (e.g. OWL), and other assistive technologies (e.g. provenance solutions, guids, etc.)
|
|
|
|
-- Main.MarkSchildhauer - 09 Nov 2009
|
|
|
|
-- Main.RogerHyam - 09 Nov 2009 |