Brave new (1.0) world.

[Geoffrey Broadwell]

On Fri, 2009-02-20 at 09:55 -0800, Allison Randal wrote:
> Geoffrey Broadwell wrote:
> > 
> > Right.  I think there is growing consensus that n.5 as a milestone
> > number must die...
> 
> Is there any more obvious way than 1.5 to say "This is the public 
> release half-way between 1.0 and 2.0"?

Why do we need a "half" release number?  Instead of numbering the
milestones 1.0, 1.5, 2.0, 2.5, why not 1.0, 2.0, 3.0, 4.0 instead?
Milestones are major, and there is nothing about the second and fourth
milestones that is less major than the first and third.

I have not seen a reason (though I may easily have missed it) why we
need to ever have a 'half release'.  If the rules for what makes a .0
and what makes a .5 are precisely identical except for whether it's the
first or second release in a given year, I think that's not actually
useful information.

If we were using years for major numbers, then the month (or a
first/second designator of some other sort) would be useful.  But given
the desire to just say "we have monotonically increasing version
numbers, and monthlies are smaller than milestones", then just number
the milestones 1, 2, 3, 4, ....  There is no sense important to the code
base or the users in which the second milestone is "half way" between
the first and third milestones, so we shouldn't pretend there is.

(Like it or not, people have had some representation of a number line
pounded into their heads since childhood.  If you use a .5, people will
subconsciously want to ascribe real meaning to that in order to please
their third grade teacher.  When you use whole numbers, you can rely on
the happy innocence of kindergarten.)


-'f