Paged Feeds

This document explains the reasoning behind our recommendation to create paged feeds by podcasters and to provide support for paged feeds in podcast clients. This is not a format specification. Instead, we are defining best practice here and point to the IETF specification RFC 5005.

The problem with feed size

Podcasts are communicated to the user by the use of podcast feeds. These feeds, usually encoded in RSS 2.0 format, syndicate content by listing news items and are periodically checked by interested parties. This concept is known as polling.

The more parties are interested in the content, the more polling occurs and increases the overall network load on the server. Also, the bigger the feed grows, the more load is being generated both on the server and the client who needs to parse the content of the feed to check for new content. This is especially problematic for mobile equipment that needs to save battery power and has less computing power available compared with desktop machines.

In addition to that, popular feed proxy services like FeedBurner have certain feed size limitations. If a feed grows too big, they stop processing it, eventually shutting down the service. In order to prevent this from happening, the feed size must not grow beyond this limit.

This has lead to the standard practice of just including the most recent items in the feed,  kicking out entries that point to older podcast episodes.

While old feed items are not needed to check for the latest content, people sometimes explicitly want to access older episodes or even go back and review the whole archive. Podcasters sometimes present archive pages on their websites, but most people prefer to use the same tool – their favorite podcast client – to access these episodes as they do it with the latest episodes.

This is where Paged Feeds help.

Using Paged Feeds

The RFC 5005 defines two new methods of how to present feeds to clients:

  • Paged Feeds
  • Archive Feeds

While the usefulness of Archive Feeds for podcast feeds remains to be discussed, we found that Paged Feeds actually provide an appropriate and easy to use solution for the problem of providing access to older episodes without increasing feed size for entities checking just the the latest information.

Once a podcast distributes feed content over multiple pages, using the specification mentioned above, the clients still get small subscription feeds but clients interested in the whole archive can get to this easily by reading additional pages until they find the content they were looking for or get a complete archive.

Interestingly, this is also compatible with feed proxy sites like Feed Burner because the first page still serves as a valid subscription feed whose size can be kept to a minimum and the service leaves the link to the next page in so that informed clients can still access the additional items.

We encourage both developers of podcast CMS and podcast clients to include support for Paged Feeds as specified by RFC 5005. It does not add much complexity to either side but enhances the user experience for podcast consumers significantly.

9 thoughts on “Paged Feeds

  1. Pingback: Paged Feeds for Podcasts | Podlove

  2. Pingback: Neue Formate, neue Feeds, neues Glück | CRE: Technik, Kultur, Gesellschaft

  3. Ich habe den Text bei GitHub nicht gefunden, daher hier ein paar kleine kosmetische Verbesserungsvorschläge:

    Typo: they statt the: “If a feed grows too big, they stop …”
    Typo: ein Komma fehlt: “…, they stop processing it(,) eventually(,) shutting down …”
    Stil: gleiches Schema im nächsten Absatz; ist ‘eventually’ wirklich (bei beiden) nötig? “… in the feed, eventually kicking out …”
    Stilbruch: ‘stuff’ sticht aus dem gehobenen textstil heraus; im ersten Fall möglicherweise “… to parse the content of the feed to check for updates.” oder (je nachdem, wie ‘feed’ definiert ist) “… to parse the feed for new content.”

  4. Pingback: Podlove: Phase 2 | Metaebene Personal Media

  5. waere es nicht sinniger dem server einfach mitzuteilen welche folgen man schon hat? wordpress AFAIK generiert doch eh den feed bei jedem abrufen neu. und das abrufen ist ein http request, eine technologie, die darauf ausgelegt ist daten in den request zu packen. meine gedanken gehen in richtung eines ?since=$uuid get parameters

  6. Pingback: Update – Listening to Podcasts on Android |

  7. Pingback: Dentaku » @unixtippse Es gibt http://t.c…

  8. Pingback: AntennaPod für BlackBerry 10 aktualisiert |

Leave a Reply

Your email address will not be published. Required fields are marked *