Sort help for feed aggregators

As our lists of feeds grow it becomes more difficult to sort through the clutter that greets us as we fire up our news aggregators. The list of publishers of supported feeds continues to grow, as does the appetite for consumption. The next important step in the feed aggregation space will be how you tame the data available within the application. I propose sorting services that would allow developers to offer their own reclassification of a list of feeds or their content.

Pass a list of feeds in OPML and a web service will return the same OPML with an additional “sortorder” attribute for each outline element. The aggregator consumes the OPML, sorts on sortorder ascending for the group, and the user receives a different representation of their subscribed feeds.

Possible sort plug-ins include Google PageRank, Technorati source authority, degrees of separation through a social network, FOAF relations, blogroll, etc.

I sent NewsGator an e-mail asking if they would support sort plug-ins. If you develop a feed aggregator and would be interested in pursuing broad sorting solutions, let’s start building. Do you have ideas about more ways to sort your feeds? Leave a comment or send a TrackBack.

  • Posted
  • Updated at
  • Comments [5]


Commentary on "Sort help for feed aggregators":

  1. Greg Linden on wrote:

    Wouldn’t more sophisticated sort methods require a lot more information to pass between the client and the web service? For example, sorting by social network, FOAF, blogroll, or any type of personalization would require passing the entire social network, FOAF relationships, etc. across to the web service, wouldn’t it?

    Simple sorts might be able to be done in the service layer — sorting by blog authority, popularity, or recency — but more complicated sorts might have to be done at the client, don’t you think?

  2. Niall Kennedy on wrote:

    Information such as a social network or a blogroll would exist on other servers. Assume MySpace and Blogrolling.

    When passing an OPML file complete with RSS information such as link, managingEditor, or webMaster an outside service could match such information to their own subscriber records. Since link is the only required element it would have to be the starting point.

    The service would then process the file and add its own element(s). In this case sortorder would be the minimal addition. The client application does not need to know anything about the social network, blogroll, or FOAF, or even that it just called a service of that type versus a randomization engine. It simply provides the OPML and acts on the result.

  3. /pd on wrote:

    I think here there has to be some sort of meta data element(s). This literally feeds the sort algrothim! Some sort of huerstic pattern matching technic where The user keeps meddling with the meta elements .e.g. lets take 2 meta elements per feed, which indicate (a)quality (b) time. Based on this, I might go thru a ranking process (i.e.quanitifcation mechanics) which indicates to the sort engine that all feeds with (a =1) and where (b=1) needs to be seeded at the very top!!
    So whatever pops of the stack will be the first item to show on the aggregated list (?)

    Did I make sense ??

  4. Niall Kennedy on wrote:

    I do not see what you would like to gain by the extra attributes. I am trying to make the aggregator indifferent as to the source and the importance of the metadata. Sorting is handled remotely in most cases, and only a sortorder is returned. The aggregator has no interest in whether the channel is most recently updated or if you would like to sort by members of your poker club. In some cases you may want a weblog that has not updated in a while to rise to the top of your list since you have given up on the channel.

  5. /pd on wrote:

    Niall, this is extactly the point !! “aggregator indifference”. Why would I use an aggregator which was not smart enough to tell me that these are the channels which you most want ?? The next step towards aggregation is to have a “smart aggregator.” Some amount of smart agent kinda aggregator.
    Anyway that 2 3/c cents, on how I look at smart RSS mechanics. Its for food for thought, literally!!