FeedOnFeeds (geekery)
Since I last used it, FoF's UI has improved greatly. There's still a few niggling issues (although all the major ones seem to have been fixed; what's left is more issues with certain crappy RSS feeds being broken in certain recoverable ways, and I have the source and can certainly look into fixing it myself), and as a reading experience goes it's way faster than Google Reader, since it doesn't try to act like a fully-featured desktop application.
The main thing that I miss from Google Reader is the "automatically flag as read as you go" thing, but as it turns out, double-clicking the entry checkbox will actually flag all the entries leading up to that one, so you actually can incrementally mark items as read — and if you accidentally mark something read you didn't mean to, recovering from it isn't nearly as much of a chore as it was in Google Reader.
Also, now there's a plugin system which is pretty easy to write plugins for (albeit completely undocumented, like most of FoF). It also finally has actual user accounts; no more having to mess around with .htaccess and being stuck with a single shared collection of entries. I don't know how well it separates the users though; it looks like there's some basic item and feed sharing, but I don't see anything in the db schema which allows for different users to have different sets of feeds and items (or even separate "mark as read" functionality) so it's still not particularly suitable for a multi-user installation. (Also, the way user preferences are stored in the database is horrible and clearly was added on as an afterthought.)
Still, if you have the ability to run FoF (and if you have a webhost with PHP and MySQL, it's absolutely trivial to install), I recommend it as an alternative to Google Reader or Bloglines.
One major downside is that the developer doesn't seem to do any active development, judging by how the bug tracker and mailing list are doing, but it's GPL so it's not like forking it is out of the question.
Comments
(Says the guy who organizes his photos on Flickr, but je suis aware)
I haven't yet looked into backing up the title, description, tag, collection and location data, but you'd better believe I'm going to, because that was a significant time investment.
You might be able to do it through their Atom or RSS APIs (which provide the tag data), although I don't know if that allows you to get historical data (and it definitely wouldn't allow private items).