[
Lists Home |
Date Index |
Thread Index
]
9/18/2002 8:51:49 AM, "Thomas B. Passin" <tpassin@comcast.net> wrote:
>>Today Google uses mostly links between web pages for it's rankings and
>>assosiations.
>
>But that is because embedded meta data often lies to influence the rankings,
>not because Google cannot get any metadata, and also as a kind of quality or
>importance ranking.
Right!!!
> Using xml (or RDF, or ...) for meta data will not
>change that.
Well, RDF, or XLink perhaps, might be used by *third parties* to describe
resource, or a relationship between resources, in a way that Google or
some other search mechanism might find useful.
Trivial example: Google counts all links to a resource as "votes" to increase
the Page Rank of the target. This means that by linking to something as
an example of something profoundly stupid, one is "voting" for it to have
a higher Page Rank. This leads to the mildly pathological
situation where popular weblogs are treated as "authoritative" by Google
simply because they are the original source of links to resources that
are authoritative.
So, even though Google rightly ignores metadata about a resource within the
resource, it might someday reasonably use metadata about the link between
one resource and another.
This is not to "vote" for any particular link/metadata technology, just
for the notion that this is not an intrinsically useless idea in the age of
Google.
|
