|
Many are familiar with weblogs.com, a website that
lists recently changed weblogs. In order to participate, a weblog 'pings' the
site to indicate that it has updated. The ping serves as a more accurate and
less-resource intensive way to maintain an up-to-date list than a traditional
crawling approach.
RSS aggregators work by using the crawling
approach, fetching RSS feeds at set intervals to look for updates. Most of time,
the majority of the feeds fetched have not updated, but there is no way to be
sure unless each feed is checked. This takes time, and can place a load on sites
that host feeds. A ping-style approach would be more effective and (time)
accurate.
fyuze has been making use of a crawler for some
time, and after each crawl the majority of the feeds register no changes. In
addition to the tried-and-true crawl, there is now a ping-style interface that
allows one to notify fyuze that one's RSS feed has changed, and have it
immediately fetched and indexed. The API for pinging supports XML-RPC as well a
plain and simple HTTP post. Information on making use of this API can be found
at http://fyuze.com/api/ .
Here is a quick example of how the HTTP post method
works:
Simply substitute the URL of your RSS feed. If
fyuze has never seen your feed before it will fetch it, and if its valid, add it
to its list, submit it for categorization in the directory, and index it. If the
feed has pinged before, it will simply fetch and index it. It will only allow
each feed to be fetched once every five minutes. (to prevent abuse)
Please feel free to add functionality to your
weblog to automatically ping fyuze when it updates (using HTTP post or XML-RPC).
|