Noindex
via Danny Ayers, I see that Bloglines has proposed an extension to prevent indexing of feeds. There's already some prior art in this area. Greg's proposal apparently got no traction 18 months ago. I think the noindex indication is a necessary thing. Currently, NewsGator will not index any feed that's password protected, but otherwise, noindex requests are handled case-by-case. (I think at this point I'm supposed to say that I'm a NewsGator employee, but the opinions expressed herein are entirely my own and do not necessarily reflect any official position of NewsGator.)
From a specification perspective, I'm not sure about the wording in the Bloglines proposal: "To 'allow' access means a feed may be redistributed to other public sources, including search". The wording suggests that access restriction is not limited to indexing for search purposes (as in Greg's proposal) - I interpret that to mean subscribing as well. What's the definition of a public source? A shared aggregator like Bloglines or NewsGator Online? If you say "deny", does that mean that your feed can be accessed only by a standalone aggregator such as FeedDemon or NetNewsWire (with NewsGator synchronization turned off)? And as Alex Barnett points out, this still doesn't take care of the republication scenario. That said, I like the idea of an RSS / Atom extension better than robots.txt - I believe that using robots.txt for this purpose is an abuse of that spec. Simply polling a syndicated feed is not robotic behavior, and actually places a fairly onerous restriction on both sides - the server now has to service two requests (for robots.txt and again for the feed) and the client has to understand the robots.txt protocol, and cache individual server preferences.