Link Details

The way some services exposes paging physically ensures each page has to be recomputed every time there's a new addition - new content keeps getting added to page=1, pushing everything else down off the main page causing a ripple effect over the set of pages. As a result, pages easily can't be archived and are tricky to cache - the content of "page=5" changes every time there's new content on the homepage (in Twitter's case new content on the hompage is extrememly frequent), because the paging model is a pushdown stack.