Currently RS uses Open Annotation Collaboration (OAC) version 201110 for Data Annotation and Image Annotation.
A W3C Annotation Community Group was chartered to merge AO and OAC.
The result is a draft version 201205 of a merged AO and OAC ontology, called Open Annotation (OA) ontology.
- OA does not define classes for Body and Target, i.e. oa:Target and oa:Body are GONE!. Instead, the body and target should have:
- type of content, represented as a DCMI type, eg
- be of any media type, represented as a MIME type, eg
- provenance is now represented with oa:annotator (subProperty of dcterms:creator) and oa:annotated (subProperty of dcterms:created)
- modelVersion: there SHOULD be exactly 1 per Annotation:
- instead of a fragment selector at the URI (#xywh=50,20,200,200), now uses a node oa:FragmentSelector with rdf:value="xywh=50,20,200,200"
- Extension Annotation Classes defines an appropriate class: oax:Bookmark
- oax:Bookmark usually has no body, but could have a Comment body
- oax:Bookmark with a comment body, stored as cnt:chars in cnt:ContentAsText.
Uses oa:annotator, oa:annotated (not DC terms).
- Target resource has rdf:type dt:Text and dc:format "text/html"
- Image is a dcterms:StillImage and has dc:format "image/jp2".
TODO: shouldn't this be dcmitype:StillImage (not dcterms:StillImage nor dcmitype:Image).
Annotated point is represented with xywh oa:FragmentSelector where w=h=1
- List of annotations uses OAI ORE (Object Reuse and Exchange):
a ore:Aggregation that has ore:aggregates links to the individual items
OA uses the following ontologies and prefixes:
|Pref||URI||Get from URL||Local file|
It would be best to update Data Annotation and Image Annotation to use the new version (OA). This won't be hard, see the changes above.
The priority of this task is not determined and it is not scheduled yet.
- W3C Community Draft of merged spec was published in Feb 2013
- A Rollout of this spec is going on in summer 2013
- Robert Sanderson (LANL) has an excellent set of recent presentations: http://www.slideshare.net/azaroth42 including
Data Model tutorial
Uncertaintly and Multiplicity in SharedCanvas
Plans and progress of the spec at iAnnotate workshop