Metadata & RDF: intention and reality

The W3C has done a fine job in creating a robust and flexible metadata convention with RDF.  However, I have been searching the W3C archives in order to find works relevant to improving the functionality of metadata, and have yet to locate any.

It is understood that the main weakness of metadata, and specifically META tags, is that they are not used for their intended purpose, i.e. the precise and accurate description of the content of a resource.  After reviewing the work on RDF, I see that RDF fails to address this problem as well.  (Granted, third party reviewing exists in the form of PICS, but this seems limited to restricting offensive material.)

Describing the subject matter of a resource is the primary function of all metadata.  In the RDF Schema Specification, description of the subject matter is limited to the RDF-ization of a Dublin Core 'subject' META tag.  On inspection we see that "... Typically, subject will be expressed as keywords or phrases that describe the subject or content of the resource. The use of controlled vocabularies and formal classification schemes is encouraged."

I argue that a formal classification scheme should not only be encouraged, but rather made the focus of the entire W3C Metadata Activity.  Otherwise, the work of the activity will be made meaningless by overzealous webmasters.  A classification scheme must be unique, complete and, for lack of a better word, strict.  Of these three, strictness is the most relevant to the success of any metadata schema.  It is the only way to guarantee the accurate description of subject matter.

Conveniently, such schemes already exist in the form of WWW directories like Yahoo! or ODP.  When we reach a specific category in a Web directory, we can be reasonably certain of the content of the resources therein.  Also, if metadata interpreters (e.g. search engines) were to allow only one 'Web Directory Descriptor' per domain, webmaster abuse of the classification scheme would be reduced drastically.

IMEXIS has recently published a proposal for the 'Web Directory Descriptor' at http://imexis.com/projects/wdd/HEMd-WDD.html (Note:  WDD is a work in progress).  If you are interested, please review our proposal and offer any comments or suggestions.

Cheers,

Angus D Madden  (adm@imexis.com)
IMEXIS internet Market Research and Exploration
http://imexis.com

Received on Monday, 28 June 1999 11:17:37 UTC