TDWG2010 Annotations Session 3: Domain
Thursday PM Session
RAW NOTES TAKEN BY Pete Ruhl, USGS
ALA:Atlas of Living Australia
aggregate from primary sources improve value by linking among
uses open linked model. things, identities, then linking data should be matter of facilitating exposure of data as properties of objects.
linked open datasets
Two sig ways to enrich value 1. someone tells us things we didn't know before 2. identifications for image 3. provide set of additional properties for one of those things
annotations: reasonably well structured snippets of information for things we have an identifier for.
if you know what "agent" provided certain piece of information then you can conceptualize different handling of info from different agents. Trust relationships, ignore, expose to this audience or that, etc.
annotations were most helpful for handling info for major domain objects.
It has turned out to be harder than forseen to utilize the central store of annotations. It seems to be easier to store the annotations inside the databases that feed the actual web interfaces.
Future directions 1. Proposing 2. thumb up, thumb down for value of various data records
Literature: a *lot* of information is contained in literature. Formal descriptions, popular descriptions, illustrations from journals, image databases.
- augmentation of source record
Darwin Core doesn't let you embed uri's
Annotations - Used elswhere?
Astrophysics: uses the annotations model for old plates, etc. They even call it annotations. Physical Oceanography discipline Amazon: annotations collected from customers.
Annotation Class interpretation augmentation correction
Does DarwinCore stuff (schema) need to be extended to annotation.