Re: [xml-dev] Aggregated content, fact checking, PICS, Atom/RSS (was Rigged Aggregators)

["Ken North" <kennorth@sbcglobal.net>]

> > No matter how you model it, the very biggest thing would be to have a way
> > for people to very easily add to meta data about how accurate a given
> > article is because relying on the publisher of the article to do this for
> > you has obvious flaws.
> > [snip]
> > But to my mind the usability and convenience of the
> > mechanism by which "accuracy" meta data gets created and reviewed by an
> > active audience is far more critical.

Michael Gorman, President of the American Library Association, recently wrote a
Los Angeles Times op-ed piece about digitizing books for Google:

"Hailed as the ultimate example of information retrieval, Google is, in fact,
the device that gives you thousands of "hits" (which may or may not be relevant)
in no very useful order. Those characteristics are ignored and excused by those
who think that Google is the creation of "God's mind," because it gives the
searcher its heaps of irrelevance in nanoseconds. Speed is of the essence to the
Google boosters, just as it is to consumers of fast "food," but, as with fast
food, rubbish is rubbish, no matter how speedily it is delivered."

In a followup piece, he also commented about the quality and
accuracy of blogs:
"A blog is a species of interactive electronic diary by means of which the
unpublishable, untrammeled by editors or the rules of grammar, can communicate
their thoughts via the web."

Attention.XML, del.icio.us, XFN, and menow are interesting, but "who's reading
what" isn't really a measure of accuracy and quality.

Media organizations, although imperfect, have staff people to do fact checking.
Academic papers are often peer reviewed.
We know there's a difference in the credibility of those sources versus blogs
and fringe web sites. So we need a solution for filtering out the junk.

Search engines, aggregators and semantic web technologies would benefit by
having some type of qualifiers for expressing accuracy and quality.

"Was this article useful?" appears on many web articles. Perhaps we need to
identifying fact-checked articles and start carrying quality rating information
in RSS and Atom feeds. An RSS channel can carry a PICS rating, but we need that
information on a per link basis, not per channel.

Chris Armstrong wrote an interesting paper about PICS as a quality filter and a
proposal by labeling the quality of information sources:
http://www.ariadne.ac.uk/issue9/pics/

CIQM developed a labeling system for bibliographic databases that included best
practices or QA policies such as using a second indexer, using authority files
for name validation, and so on:
http://www.i-a-l.co.uk/db_Qual/label.gif

With 8 billion web pages indexed by Google and 5.8 million feeds indexed by
Feedster, a quality rating system is looking more attractive every day.