hc-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Oleg Kalnichevski <ol...@apache.org>
Subject Re: Time line for 5.0; was Re: Publicising methods in a GA release
Date Sat, 22 Feb 2014 14:20:41 GMT
On Fri, 2014-02-21 at 23:13 +0000, sebb wrote:
> On 21 February 2014 23:05, Gary Gregory <garydgregory@gmail.com> wrote:
> > On Fri, Feb 21, 2014 at 5:26 PM, Oleg Kalnichevski <olegk@apache.org> wrote:
> >
> >> On Fri, 2014-02-21 at 15:54 -0500, Gary Gregory wrote:
> >> > Why not move trunk to 5.0 instead of 4.4 so we can drop all this
> >> deprecated
> >> > code?
> >> >
> >>
> >> Major releases tend to be very disruptive. They are supposed to be. We
> >> should consider doing 5.0 after 4.4 and taking it as an opportunity to
> >> introduce some major features such as HTTP/2.0 support. However,
> >> speaking from 4.0 experience it may take years to complete.
> >>
> >> I would not do a major release just to drop deprecated code.
> 
> +1.
> We should not cause unnecessary grief for end-users.
> 
> >>
> >
> > Perhaps, but you could start there, release early release often. 5.0 could
> > drop code, 5.1 could start adding features.
> 
> Release early release often (RERO) does not make it easier for end-users.
> 
> RERO may work well if the API is not being changed - i.e. only fixing bugs
> However the "release early" mantra often means that the API is
> half-baked - and then cannot be changed without causing grief to
> end-users
> 

+1

This is precisely the problem. RERO works well as long as the APIs are
either very stable or do not need to be stable. 

Once 5.0 APIs are frozen it will be quite difficult to introduce big
changes in 5.1. Let's keep 4.x line going for one or two more feature
releases and then consider doing 5.0.

Oleg



---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@hc.apache.org
For additional commands, e-mail: dev-help@hc.apache.org


Mime
View raw message