Feed discovery and Flock
I spend a lot of time thinking about bringing feed reading to a wider audience. A sobering statistic is that the overwhelming majority of web users, upwards of 95%, have no knowledge of the technology or how it can help them stay on top of content they're interested in.
One of the larger obstacles to adoption has been the lack of a consistent, user friendly mechanism for identifying and subscribing to feeds on the web. A widely adopted auto-discovery markup convention and browser integration have improved the experience dramatically. Publishers now have a mechanism by which they can provide the user with notification that feeds are available. However, in practice, current implementations can give confusing results. This is due largely to the LINK tag convention suffering from the classic invisible metadata problem: data that is out of sight easily falls into disrepair. Broken links, duplicate feeds, and invalid or poorly written titles are common in LINK tags and lead to ambiguous results like this:
There are two distinct feeds associated with this web page, and no good way to differentiate them using the metadata provided in the LINK tags. This is inconvenient for the advanced user and downright confusing for the novice.
Flock attempts to improve this situation by fetching and parsing feeds mentioned in LINK tags in the background as you browse. This lets us validate the feeds, get authoritative rather than secondhand metadata, and gives us the opportunity to use more advanced deduping techniques to provide the user with a more relevant selection:
The end result is a more consistent feed subscription experience on the wild, wild web.