Re: Shepherding large enhancements

[email protected] wrote:
> I'm going to shout also... Seems like an acepted way.
Actually, I feel much better now, and am now ready for some clear
discussion :-) I hope it makes you feel better too.

> > He, I couldn't help myself, sorry for the shouting. ;-) But I really
> > mean it. We agreed on the plan. And one of the things we wanted to do
> > is to have faster release cycles than before - we are aiming for 3
> > releases per year IIRC.
> 
> We are not talking about the plan. We're talking WHAT to do in the case
> (like now) that the plan CAN'T be respected.
What do you mean that the plan can't be respected? I don't see anything
stopping us.

> We had choosed Date and Features (an error in itself).  
No, we chose Date. The Plan was always meant as an attempt at a self
fulfilling prophecy - by talking about something, you bring it to
attention, and maybe it gets help and becomes ready for inclusion.
Though this isn't explicitly said in the plan, you may note that the
language around anything that requires the commitment of anyone other
than Guides is very tentative.

> Some of us want to
> respect the date, other (including me) want to respect the Features.
I don't think it is about what we want - the plan is more or less *the
most we can do* about specifying features. Time we can sort of pretend
to control. Any attempt to pretend to control features
(people+direction), will have far less pleasant results. Specifying
requirements you can't control in a plan doesn't make sense to me.

> > There will ALWAYS be stuff to harvest. There will ALWAYS be cool new
> > stuff to add. But we need a proper beta and gamma period.
> What is more important? the "proper periods" or the results?
In the long term, what determines the content of Squeak is, and always
will be, the amount of effort and creativity people like you put into
it. The community can help with reviews, but us Harvesters can only
affect this in limited ways. What we can do, and try to *using the
release cycle* is -
A. Make sure that the Squeak newbies download is more or less up to date
(requires somewhat frequent releases, and 3 a year doesn't seem madly
over-frequent)
B. Make sure that quality goes up, not down, over time, which requires
some sort of quality process, including reviews, betas and gamma.

Also note that the results are not on the line - NewLook is not being
prevented from going in permanently - just delayed by the 3 week beta
period. I hope in this time it gets reviewed and approved.

> IMO, the SM1.1 is a MUST to avoid a fork in the community.
I agree, in the long term. Though if you mean that SM1.1 is a MUST *in
3.6*, I disagree. I think what we'll have to do is make only one
official release image, the Full image, until we have SM1.1, enabling us
to manage more. Hopefully this will be early in 3.7. Yes, this will
require some arguing about what packages to include in Full. Price we
pay, nothing to do about it.

Daniel