RDF and the Stanford Metadata Architecture
Combines content and structure: Self-describing data
Human- and machine-readable, integrated into HTML tradition
Cannot accommodate computed attributes
Complexity when scaling up
Clumsy with multiple, like structures
Separates access name from mechanism (e.g. ‘abstract’ could be stored or computed; encapsulation)
Allows use of strong data modeling tools (object abstraction; not limited to parser)
Requires computational infrastructure beyond a parser
Needs serialization process for transport
Services need to run (not just an httpd)
Scope difference: No translators, collection metadata, caching in RDF
May use RDF for InfoBus data transport in next release