Peter Keane wrote: > On Sun, 6 Jan 2008, Brian Smith wrote: > > > [snip] > > Could you please give an example of such harm. The rationale for > > James's design is that it is the least harmful of all the > > alternatives, especially the FeedSync design. > > > > Because the "feed abstraction" is that much more complicated.
I disagree that it is a significant complication, and even if it was, I fail to see the actual harm caused. > Why is a bit of metadata as a child of an "entry" element a > non-starter? James and I have explained this in this thread multiple times. If you disagree with our reasoning, that is fine, but let's not keep running in circles. > > Implementations are free to ignore the tombstones and have > > the exact same behavior that they have now. That is the primary > > benefit of this design over the FeedSync one. > > Yes, and then what sort of extension will be proposed for the > folks who want a "deleted" feature, but need/want to keep the > deleted entry AS an entry. I suggest using FeedSync, if that is what you want. It has a lot of nice features besides tombstones. - Brian
