My understanding is that feed readers should be acting as user-agents, not robots, so this _ought_ not to be a problem unless you want eg search engines to index the feed contents. (Which we probably don't.)

Can you turn up examples of feed readers that are using robots.txt prohibitions which are affected by this?

me wrote:

I ran into this because the Python feedfinder library follows robots.txt and thus won't find the Wikipedia RSS feed.

Yahoo does the same thing: http://jeremy.zawodny.com/blog/archives/001474.html

I'm not convinced the feeds are really search-friendly... each page has a history feed, and the whole site has a number of feeds (RC and otherwise) which tend to change very quickly. In addition, the generation of diffs etc for the feeds may result in nastiness on a spider crawl visit. I'm marking this WONTFIX for now.