+1 On Feb 27, 2011 6:54 PM, "Martin Atkins" <[email protected]> wrote:
On 02/26/2011 11:51 PM, Jeff Lindsay wrote: >> >> Perhaps what is needed is to split off, or start a... Although as currently written the two issues are intertwined in the spec, the feed-specific parts are actually not difficult to separate from the rest: Fundamentally, PubSubHubbub provides a way to subscribe to changes to a resource. There are many ways that the "changeset" can be represented. As an accident of history, an representation for Atom and RSS feeds was defined first, and the simpler "just send the entire resource" representation was not defined until later. However, I think all it takes is to reorganize the specification a little so that the default case is whole-resource updates and then some specialized, resource-type-specific representations and a mechanism for determining which is desired by the client. I don't see any reason to do anything radical like fork off another spec that would only be slightly different than what we have today in terms of protocol flow.
