One thing I've noticed with podcasts, is that my reader will download the full RSS every day, listing hundreds of episodes, just to see if there's a new one.
I get why they do this as a new listener will want to start off with a full list of episodes and may want to pick up a show from the start.
It would be useful if an RSS could specify a separate URL to query that will only list new episodes, returning an empty RSS (or a standardized HTTP status code) until the new episode is available.
The original RSS with the full list would still be the primary URL for the podcast, the one I get when I right-click and select "Copy RSS URL".
IMO, this is not a problem of RSS but how that particular feed endpoint is implemented (assuming your reader will append entries). Usually, popular news sites (that provides RSS/Atom) will have e.g., "/rss/all" endpoint to get all the news for that day, and these entries are refreshed a couple of times per day. Also, they will provide "/rss/latest" refreshed every couple of minutes or so, so it's up to your reader to check that endpoint frequently, parse and populate only the newest entries.
This should be doable just using existing Cache-Control/If-Modified-Since headers, shouldn't it? Basically both sides just need to treat the "age" of the feed document as being equal to the newest entry in it.
Right! This is actually done in blogging platforms, at least in the old ones like Wordpress (last I looked, and I think) and Serendipity. I wouldn't guarantee though that this code still works or that readers do use it...
I get why they do this as a new listener will want to start off with a full list of episodes and may want to pick up a show from the start.
It would be useful if an RSS could specify a separate URL to query that will only list new episodes, returning an empty RSS (or a standardized HTTP status code) until the new episode is available.
The original RSS with the full list would still be the primary URL for the podcast, the one I get when I right-click and select "Copy RSS URL".