On Sat, Sep 27, 2008 at 3:14 PM, Marvin Humphrey <[EMAIL PROTECTED]> wrote: > The hard part about implementing PForDelta is that it involves block > read/write > of integers. KS isn't set up to do that, either at index-time or search-time.
How are you thinking about using PForDelta? The algorithm makes sense, I haven't thought much about the actual integration. At a first glance, I'm not sure at what level it would fit in. If you use it for position data alone, I'd guess that most tokens don't have enough occurrences to make it a win. Would it work for match-only posting with the positions in a separate stream? Nathan Kurz [EMAIL PROTECTED]
