Bill de h�ra wrote:
Thomas Broyer wrote:
Bill de h�ra wrote:
Thomas Broyer wrote:
Although this doesn't solve the DOS potential, I'd rather go for a new atom:source child element indicating the source feed's Id or URI. When PlanetBar and PlanetFoo republish a feed from Baz, they should "move" the entry metadata into an atom:source element.
One step further: how does that work for PlanetPlanet?
As I told in the phrase you stripped out?
And will I be able to reverse engineer the order the atom:source annotations were added? I want to determine if this operation is robust and/or reversable beyond a single aggregation step.
Well, depends what you mean by "reverse engineer" and "reversable", I'm not sure to understand...
IMHO, only the first re-publisher changes the entry to move the metadata from atom:entry to atom:source (and adds the atom:feedId or whatever its name). If an entry already has an atom:source child, you just republish it as-is. If you move the atom:source children back into the atom:entry element (and strip the atom:feedId), you should (and it is a should since the use of atom:source is currently a should in the spec) have the same entry as in the "primary" feed. Additionally, the content of the atom:source/atom:feedId tells you the URI of that "primary" feed at the time the entry was republished.
Sorry if I didn't answer you question as I might have misinterpreted it...
-- Thomas Broyer
