ANN of new versions of Mercurial

[Matt Mackall]

On Sat, 2012-12-22 at 10:24 -0800, Adrian Klaver wrote:
> On 12/21/2012 10:19 AM, Matt Mackall wrote:
> > On Fri, 2012-12-21 at 07:00 +0000, Gaunet Sylvain (ASTEK) wrote:
> >> Hello,
> >>
> >> It seems that you doesn't announce the new versions of Mercurial on the mailing list now ?
> 
> >
> > http://markmail.org/message/r3uyv3dxy6vtj6ob
> >
> 
> I can sort of see your point, but in general would have to disagree for 
> the following reasons.
> 
> 1) Bugs are part of writing software. When they get squashed is not 
> important, that they do is.

Nonsense. Fixing bugs sooner is a win. Fixing bugs when the set of
people exposed to it is still small (aka before release) is a BIG win.
The bug Tess finds on day -1 may be fixed before it potentially affects
10000 Norms. The bug Earl finds on day 0 won't be; 100 Norms will
randomly download Mercurial on day 0 and run into the bug without even
knowing its a release day and their bug reports are just as valuable as
Earl's.

Thus, Tess is an essential part of our community. If we have 10 Tesses
and 10000 Norms, we're a lot better off than if we had just 10010 Norms.
But if we have 100 Earls and 10000 Norms, do we produce better software
than with just 10100 Norms? No, because Earl shows up too late to affect
the release. Earl's enthusiasm is wasted. If we traded even half of
those Earls for Tesses however, our release quality would skyrocket.

And that's what the goal of announcing the release candidates instead of
the releases is: encourage Earl to be earlier so that his extra
enthusiasm actually helps the project.

-- 
Mathematics is the supreme nostalgia of our time.