RSS for search engine syndication

["Stephen Tyler" <tuesday@growinglifestyle.com>]

There has been a lot of discussion of the use of RSS for weblog and
news feed syndication.  In other words time ordered results.

But is this the only thing that RSS can be used for?  Well what
information does RSS convey?  Basically, it is a list of links.  What
do search engines return as results?  Basically a list of links.

Why not use RSS as a format for search engine syndication?

Well rather than just talk about it, I went and did it.

At my site http://www.growinglifestyle.com/ , whenever you do a search
there is a link to an RSS version of the search results at the bottom
of the page. The results are sorted by rank, not time.

Example: http://www.growinglifestyle.com/s1/pruning_roses/index.html
Example: http://www.growinglifestyle.com/s1/pruning_roses/index.rss

Elsewhere on the site, navigated by a topic hierarchy, are several
hundred topics each with their own RSS feed (link at the bottom of the
page).  The results are sorted by time.

Example:
http://www.growinglifestyle.com/h/garden/plants/roses/index.html
Example:
http://www.growinglifestyle.com/h/garden/plants/roses/index.rss

Additionally, for each topic, you can get an HTML or RSS version of
the links in each topic from a single source site. The results are
sorted by time.

Example:
http://www.growinglifestyle.com/h18/garden/plants/roses/index.html
Example:
http://www.growinglifestyle.com/h18/garden/plants/roses/index.rss
(in this case articles sourced from DoItYourself.com)


Why would anybody want this?

Well search engine syndication is hot, with all the major portals
setting up affiliate programs and way of providing their results (and
sometimes advertising) to other sites.  For example goto.com (a pay
for results search engine) provides a Javascript version that clients
can embed in their own sites.  Other metasearch engines (like
askjeeves, mamma, gohip) have other methods of syndicating the goto
search results.  In the early days most metasearch engines relied on
scraping, but I hope things are a bit cleaner now.

As clients adopt XML aware (and in particular RSS aware) web hosting
environments like Jetspeed, Zope etc it becomes sensible to return
search results in a compatible format, so search engines can be
handled in the same way that freshmeat/slashdot/zdnet/moreover news
feeds are handled.