Monday, November 7, 2005, 3:59:56 AM, James M Snell wrote:
> Tim Bray said: >>> * http://www.intertwingly.net/wiki/pie/ >>> PaceOrderCollectionsByAppModified (broyer) >> >> >> No consensus, rejected. > I'd say that this is still an open question. Eric and Thomas are making > a strong case for it (tho I'm still not convinced) and should probably > be given a bit more time to try to convince folks before it's closed. I agree. I think actual modified dates are a more useful sort key for the protocol than the partially subjective atom:updated. The only drawback I can see is that it adds some burden for implementors who don't require it (although to support conditional GET, the servers must retain the actual modified date anyway). -- Dave
