Tim Bray wrote: >I was chatting with Steven Lees of Microsoft about this during the >interop. One of the thing that makes me nervous about SSE is the >sse:deleted element which says "This element isn't here any more"... >this is going to be mega-confusing to a non-SEE-savvy client.
That concern makes sense to me, given the way that we're used to using feeds and feed readers today. It would be strange if a blog post was gone when viewed from the web page of the blog, but the post was still present in the feed. I think the deleted entry issue will come up more often as people start to use Atom and AtomPub for exchanging and syncing data. There may never be a web page at all in that case, and the feed becomes a way (or the only way?) for the server to hand state changes out to clients. In that case, you must have a way to represent deleted items, otherwise the clients will never know that they're gone. This could be useful even in the typical feed reader case today. It doesn't happen very often, but if someone actually deleted a post, I'd be happy to have my reader keep the cached copy around, but still give me some indication that the post was deleted by the publisher. As we discussed on IRC during the interop, there are some ways to make the two schemes coexist. You could have different URIs depending on whether the client understood deleted items; or you could have the client provide some HTTP header indicating whether or not it wanted deleted items to appear in the feed. Steven
