Sourcing sites is how Google Scholar decides on quality, so that's
definitely the quick hack for this.
But ultimately we need provenance for metadata (and content) in order to
decide quality, and that leads us back to RDF...
but we might have to supply that RDF using various heuristics for the time
being.
MacKenzie
At 02:29 PM 5/31/2005 -0400, Erik Hatcher wrote:
>>Hmmm. It might be useful, but we aren't doing that here, and I'm
>>having a hard time figuring out how you would know whether a given
>>item was peer reviewed
>>(without just asking a human that is). It's usually not indicated
>>in the metadata, although you might be able to infer it in certain
>>cases (i.e. came from a publishers
>>website, or contains publication metadata from a journal known to
>>be peer-reviewed). Did you have a mechanism in mind to decide that?
>
>We're in the beginning stage of this and haven't given much thought
>to how this would work. I had figured that an initial restriction
>by domain name would be sufficient to start with. Keying off "I'm
>ok" metadata could lead to it being abused or the truth stretched and
>doesn't seem like the right way to denote this.
>
>Perhaps using some type of service to vet the site would be used?
>Connotea does something similar by pulling additional information
>from sites it knows about: http://www.connotea.org/guide#autocollection
Received on Wed Jun 01 2005 - 19:59:17 EDT