While building a feed aggregator into a project of mine, I ran into a usage of a HTTP header unfamiliar to me before today.

From the documentation of the (very excellent) python feedparser project:

ETags and Last-Modified headers are two ways that feed publishers can save bandwidth, but they only work if clients take advantage of them. Universal Feed Parser gives you the ability to take advantage of these features, but you must use them properly.

The basic concept is that a feed publisher may provide a special HTTP header, called an ETag, when it publishes a feed. You should send this ETag back to the server on subsequent requests. If the feed has not changed since the last time you requested it, the server will return a special HTTP status code (304) and no feed data.

Following warning (from the feedparser documentation) should drive the significance of these headers home:

If you do not support ETag and Last-Modified headers, you will repeatedly download feeds that have not changed. This wastes your bandwidth and the publisher’s bandwidth, and the publisher may ban you from accessing their server.