On Tue, Jul 27, 2010 at 6:45 AM, Cole <[email protected]> wrote: > As I understand it, a subscriber uses a topic url to subscribe to a > feed at a hub. The topic url is a feed alias that resolves to a feed > that is identified by its feed id. I'm trying to understand why some > functionality in the GAE reference implementation is keyed on the > topic url rather than on the feed id. > > For example: > > FeedToFetch uses the topic url as an identifying key. If the hub is > going to retrieve the newest entries from a given feed why would it > want to make the same request again to the same feed via a different > topic url?
Great question Cole. The reason it works this way is: 1. Multiple feeds on different topic URLs have the same feed ID and very different contents. This happens with proxied feeds (FeedBurner, pheedo). You can use feed ID to determine aliases of a feed, but you cannot assume they have the same content. 2. Who do you trust as the canonical owner of a feed ID? If two topic URLs on two different domains provide the same feed ID, which one should you trust?
