RDF improving Resource Discovery

I am doing some research into improving resource discovery, by greatly
reducing the volume of data that needs to be indexed and searched by search
engines. RDF seems to be quite good for homogenous content or a small amount
of heterogeneity,  however, how would one go about using RDF to describe a
site that is updated very frequently - such as a news site that may change
hourly.  Obviously adding to an RDF header each time a new story is added
would be very time consuming and it just wouldnt happen.  Maybe if the site
architect was smart, the title of the news story could be automatically
added to the RDF header, but is there another method that is incorporated in
RDF?

Also, if RDF use becomes widespread, why wont it suffer in much the same way
META tags have, ie: by being abused by xxx site publishers? that is, adding
entire dictionaries or completely irrelevant entries to guarantee a large
number of hits.

Thanks,
Thad