Brian Smith wrote:
James M Snell wrote:
Excellent points. The primary issue for me is the fact that the set of entries we're synchronizing has to remain stable throughout the synchronization process or else we risk race conditions and a variety of possible conflicts. If the stability of the pages can be reasonably guaranteed then paging is acceptable.

I'm pretty sure I agree with you, but please define "stable".

Look at how twitter implements paging (it's not using atom, but the concept stands). The content on page=2 of your twitter changes as a function of new posts on the home page, although the entries themselves are uniquely identified. That's not stable.

cheers
Bill

Reply via email to