After months of resistance I set up a news aggregator today: FeedDemon. I also checked out sharpreader, newzcrawler, and awasu. I picked FeedDemon because its UI was the easiest to use after 3 minutes’ evaluation.

Some essential features:

• Correctly load relative image links. I was amazed at how many aggregators don’t work without a hostname in the <img> tag.
• Support for gzip and conditional GET.
• Atom support. It’s not hard.
• Ability to deal with poorly formed XML. Yeah, I read the debates. Any developer who thinks their user wants to know about a badly encoded & in line 23 column 8 is plain wrong. Hassle the developers, not the users.
• OPML import and export. The actual format may be dumb, but the ability to share a list of channels is important when evaluating software.
• Tranquility.

I like a few things about FeedDemon:

• Browser window on the side: tall rather than wide.
• Clever newspaper view. I don’t like the available styles, but it’s customizable.
• Tabbed browsing.
• Single key to read through all unread articles.
• Customizable blog-this command

FeedDemon isn’t perfect, but it’s pretty good. The beta is significantly improved over 1.0.

One thing I’ve learned is finding syndication feeds is harder than it should be. Autodiscovery helps but NewzCrawler was the only aggregator with it built into MSIE where I need it. The worst thing is folks who publish 9 different syndication formats. I don’t want to have to choose, I just want good bits.

I’m still unhappy with one thing: the visual blandness of an aggregator. Everyone’s blog looks the same! It’s no better than eight months ago. We need a way to rationally apply CSS to Atom. It’s a hard problem.