My expirement with Dave Winer’s Feedhose protocol started off a week ago with a simple river that auto updated with the latest headlines from the NY Times. As I said at the time, I enjoy pushing out that first draft to help grasp what else I want to do with it.
So far, it’s been really useful. I leave a browser window open all day pulling headlines from the NY Times as they are published.
It only took a few hours before I started wishing for more.
My next step was to branch off a bit. I was able to hook in to The Guardian through their pretty fabulous open platform API. This wasn’t as easy as hooking into Dave’s Feedhose, but it was pretty close. Long polling isn’t an option and I have to keep track of my own cursor. The Guardian’s API keeps track of things by page number and items per page instead of item number, so some duplicate content has to be parsed. The API is snappy though, which is very helpful. Right now I’ve been making a request every two minutes for updated articles and I’ll probably up that to every one minute pretty soon.
Two sources still wasn’t enough, so when I saw Dave mention his AFP feed (coming soon here) in a comment on his post covering Erik Kidd’s Feedhose client, I decided to plugin BBC to see what would happen. Sure enough, it had already been hooked up and has been working smooth all day.
Of course, once I had three separate pages up tracking real time news from these great sources, the next logical “what if” became combining those into one easy format.
So here it is, with a brand new domain, Feed River Wire.
Note: The one caveat with The Guardian is their news traffic drops drastically during the late night hours, London time. NY Times seems to have a decent stream throughout the night. We’ll see about the BBC tonight.