On Tue, Dec 2, 2008 at 2:54 PM, Damien Katz <[EMAIL PROTECTED]> wrote:
> > On Dec 2, 2008, at 2:38 PM, Dean Landolt wrote: > > ...and to finally change "count" to "limit" in view query string >>> arguments. >>> >> I'm a big fan of "limit" -- I know couch ain't sql, but no need to eschew >> convention without reason. Also, what about changing "skip" to "offset" >> which is more conventional? And I know this has been asked before, but is >> it >> even remotely feasible to make skip/offset/whatever efficient? It seems to >> me it would make pagination much more logical, rather than having to drop >> down to startkey/startkey_docid semantics. >> > > Sorry, but making skip efficient is hard to do in a partitioned setup, and > considering the stateless HTTP api, the startkey semantics actually work > quite well. Users don't want view rows 50-60, they want the next or the > previous page of results. When using absolute offsets, if the view changes > underneath the user between paging requests, they may not even realize the > rows have shifted and they are now in a completely different place in the > view. > > -Damien I didn't think of it like that (I rarely do). Makes sense -- thanks for the heads up.
