From: Rick DeNatale Date: 2007-10-04T07:05:38+09:00 Subject: Re: Release compatibility/train On 10/3/07, Charles Oliver Nutter wrote: > Rick DeNatale wrote: > > If I understand what Matz said earlier on this thread, not really. > > > > What we used to think of as 2.0 wlll now be 1.9.1 which will be > > considered (feature?) stable. > > > > I interpret this as meaning that the experimental changes will stop > > when 1.9.1 is released and 1.9.1+ releases will be implementation > > improvements only, and any new language changes would be in an > > eventual 2.0 stream. > > > > Did I get that right Matz-sen sei? > > Hmm, that's not what I got. I was still under the impression that 1.9.1 > would include experimental language features that may or may not survive > into future releases, and that the primary focus of getting 1.9.1 out is > to expose the new VM and m17n and a few other big additions to a wider > audience, with the understanding that other areas of the system are > still in flux. I say that's a very sensible idea, and I imagine many > people could start "trying out" 1.9.1, but that in general it would not > be a good release to start doing "mass porting" of 1.8-compatible apps. > > But I eagerly await Matsumoto-san's response! :) My impression came from Matz's response to you >So would the recommendation be that people running production >applications remain on 1.8 series? >> Honestly speaking, 1.9.1 at the end of this year would not be >> "production ready". Despite that production code require months to >> adopt compatibility issues between 1.8 and 1.9. As 1.9 used by many, >> its stability will rise soon (hopefully). Which gives me a picture of production code USING ruby 1.8 being adapted to a stable 1.9.x language definition thus driving defect reduction in the 1.9.1 implementation rather than additional language experimentation. Of course I could well be wrong. -- Rick DeNatale My blog on Ruby http://talklikeaduck.denhaven2.com/