semantic annotation

To whom it may concern:

We would like to present a new approach to annotation that spans the 
range from associating differing kinds of information to ontology-based 
semantic annotation.
A preview version of the proof-of-concept Java tool, which is called 
TUAM (tool for universal annotation and mapping) can be found here: 
http://www.scai.fraunhofer.de/tuam0.0.html?&L=1

It is based on the following principles:
- annotations are of the RDF subject - predicate - object kind
- annotations and mappings can be n:m with an arbitrary number of 
different relationships (predicates)
- annotation is done non-invasively, only pointers to data sources are 
stored
- data sources can be plain text, HTML, bitmap pictures, relational 
database tables, CSVs, PDF, RDF, ...
-- other data formats can be included by implementation of a simple 
interface
- sources can be accessed through URLs either locally or remotely
- annotations between various data sources result in semantic networks 
that can serve as the basis for new ontologies
- annotations are stored in a proprietary format in an RDB repository 
but can be exported as RDF

The principle limitation so far (apart from the sometimes apparent 
absence of bug-freeness) is that annotations can not be inlined in 
either one of the data sources; thus, it is not yet possible to write 
back e.g. RDF into HTML.

To our knowledge, the versatility offered by TUAM with respect to data 
formats has up to this point not been offered by any other tool.

Feel free to download TUAM and send back comments, criticisms or 
suggestions.


Best regards

Kai Kumpf

Received on Tuesday, 25 July 2006 20:31:11 UTC