Kevin Burton’s NEW FeedBlog

Tim Bray posts a article on the naming and usability of feeds:

The syndication jungle drumbeats are throbbing back and forth over what to call ’em and how to subscribe to ’em. Feeds, I mean. Which is irritating: the important problem—how to make them easy to use—is easy, and we could solve it pretty well right now if we focused on it, instead of on the other problem—what to call them—which doesn’t matter very much, and we can’t do much about it anyhow.

I’ve given up. I don’t think we should call them RSS or Atom but just ‘feeds’. Thus I named this blog ‘feedblog’ and not ‘rssblog’ or ‘atomblog’.

One downside is that RSS is such a good Google term. ‘Atom’ of course is not. Neither is feed.

Here’s the truth: an orange “XML” sticker that produces gibberish when you click on it does not win friends and influence people. The notion that the general public is going to grok that you copy the URI and paste it into your feed-reader is just ridiculous.

Exactly. Safari has done a great job of this in my opinion. They’ve gone above and beyond here. I’ve always wanted to revisit the RSS handling in NewsMonster but never had the time. The user should never see XML in their browser.

RSS should work like this; it never has, but it can, and it won’t be very hard. First, you have to twiddle your server so RSS is served up correctly, for example as application/rss+xml or application/atom+xml. If you don’t know what this means, don’t worry, the person who runs your web server can do it in five minutes.

They can but they don’t. The percentage of webservers configured to return text/plain for XML is just too high. Same with template systems such as php. If a web monkey writes their own RSS output as PHP (rss.php) the content type by default will be text/html.

The key point here is that media types are broken. You should be assuming that they will fail not that they will work. Developers are still making mistakes on this simple stuff. Character encoding, well-formed XML, and media types. Oh my!

Aggregators just need to be smarter. I know web purists will decry the feed: URI and content-type sniffing in the browser but all this framework is necessary if feeds are ever going to take off. Unfortunately its the lesser of two evils. It’s less evil to use the feed: URI than it is to have 100k users seeing the XML of an RSS feed and being confused.

Update:

Danny Ayers blogs on this as well.