If the group would allow one more message on tokens...
The Washington State Library is using a harvesting technique to extract
metadata. Our server robots visit Internet sites, pursue links, and
read/parse document information. If metadata is present (e.g., HTML
metatags or XML/RDF descriptions) the robot will also index these values
against our preset element schema.
It would be best if the robot did not have to refer to an external map
('Schema') to convert document metadata elements to our index terms. We
would like the document to carry both local terms and universal tokens
value within its metadata.
This is why we are promoting 'internally mapped' metadata elements.
We are very interested in a solution that places the token with the
metadata content (without reference to external schema or tables) for
efficiency of robot extraction.
Are changes needed to HTML / XML standards to support a well-formed
syntax that includes the absolute token value in the metadata itself?
Thank you for your assistance.
_______________________________
Philip Coombs, Project Director, GILS-IMLS Project
Washington State Library 360.704.5279
NEW EMAIL after 6/1/99: [log in to unmask]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|