That's not a bad question, actually. I've been away for a few weeks, and I didn't even realize that 4.0 had been officially released. I think the general consensus for a long time with the team was that we wanted to port 4.0. We used to (still do?) have a branch for that, and a while ago, we had ideas of skipping and/or working on both simultaneously. If this project was ported only for myself, I'd choose to port 4.0 only.
One thing to consider is the number of people that use Lucene and the amount of backward compatibility changes that have been made between 2.9.4 and 3.0.3 (we're already forcing this on our users) and then from 3.0.3 to 4.0. Counting just from what I see with Java from 3.0.3 to 4.0: Backward compatibility changes: 87 Runtime behavior changes: 36 API Changes: 107 (These include renamed, removed and added api calls, which might not affect everyone) That leaves the question of whether it would alienate our users to abandon the 3.x branch in favor of working solely on the 4.0 release. There are other related questions as well, as to the actual scope of the API changes and to the backward compatibility of the index formats. I don't have the answer to any of those, and I can't make a decision about it by myself, I only one member of the PMC. Thanks, Christopher On Mon, Oct 29, 2012 at 4:31 PM, Michael Mitiaguin <[email protected]> wrote: > I guess, a bit late to ask this question, as I can see 3.6 in svn > Has skipping of respective Java version ever considered ? I presume, it is > easier to migrate from 3.6 to 4 other than from 3.03 to 4 , but isn't it a > double effort ? > > Not clear, for how long we will be waiting for this important improvement : > http://blog.mikemccandless.com/2011/03/lucenes-fuzzyquery-is-100-times-faster.html >
