OpenSearch results, returned in a standard format (whether that's RSS, RDF, OAI or whatever), would certainly make things easier and less fragile.
To produce a Flitter page at the moment, search results are collected in a number of ways: an XML-based web service for Amazon; text-based web service for Audioscrobbler; RSS for Flickr, del.icio.us and Blogdigger and screen scraping for Feedster.
It must be difficult to fetch results from all the sources listed on the 'add columns' page of A9, so I can see why they want to standardise things.
Comments
All fields are optional, email address will not be shown; no HTML, URLs are automatically hyperlinked.
Why screen scape flickr when we've got rss feeds?
whups, I meant feedster
add &type=rss to a search
The reason was that the Feedster HTML page had text excerpts already stripped of markup, which is what I wanted, whereas the RSS feeds contained the full posts. I took the easy route, rather than making the excerpts myself, but you're right, the RSS data would probably be best overall.