We like to think that RSS uses bandwidth efficiently because it only contains content – and it’s all relatively new content at that. But while retrieving the RSS feed may be more efficient than retrieving the corresponding web site, not may people are polling the web site for changes every 30 minutes either. And in that light, the real question is whether it could be more efficient.
My weblog is configured to have seven days of posts on the home page and a maximum of 25 items in my RSS feed. With my posting frequency, that means that my feed will usually have 7 or 8 items. More importantly, only one of those items is likely to be new to one of my regular readers – more than 80% of my RSS feed is wasted. I’ve got a handful of subscribers, so it’s not a big deal. But if I had more subscribers or I ran a weblog hosting service, than I’d try to get some more efficiency out of my feed.
If we’re willing to sacrifice simplicity, then we can improve the efficiency by splitting the feed into several parts. Suppose we split the feed contents into daily chunks (more prolific bloggers would split each day as well) with the main feed pointing towards the relevent chunks. Then subscribers wouldn’t need to download 6 days of old posts to find the single new post.
I doubt that the added efficiency is worth the added complexity. And RSS bandwidth costs are only an issue for the most popular sites. But it doesn’t hurt to have something in our back pocket.