Things I (Rothman) like about Scopus RSS:

• The RSS feature is well-integrated in the search results page, and easy
  to find.
• Naming of feeds makes them clear and unambiguous when they appear
  in the aggregator.
• The RSS URL is created quickly and easily.

Things I’d (Rothman) like to see Scopus RSS do next:

• The RSS items should contain, if not the full abstract, at least a few lines of it. Now, all that shows up in the aggregator is the hyperlinked title and a one-line description of the source. More detail in the item would allow the user to do some filtering without having to go to Scopus. This would save the user’s time (Ranganathan's 4th law, anyone?). RSS feed URL produced should be a live hyperlink, not just text. This lets users with many kinds of subscription tools subscribe more easily.
• The RSS feed contains ONLY the Scopus results, not the web results. Ideally, the user should have the option of whether to do one of the two, or both.
• More user documentation about how Scopus RSS feeds will
  work is needed. It isn’t clear, for instance, how often
  the search that generates the feed will be executed. Perhaps the user could
  select from a menu how often he/she would like the search to be executed and new
  results generated for the feed.
• Change the RSS button to the proposed standard.
  

Summary: Scopus is off to a good start with their offering search results as RSS feeds and should be applauded for having them. However, from the perspective of someone working towards SDI goals, its RSS features are not yet caught up to PubMed in usefulness.

Thanks David for the brief insight into Scopus and RSS. I think it is helpful.