a few questions and concepts

Kevin Kinsey kdk at daleco.biz
Sat Apr 8 13:15:14 UTC 2006

Robert Huff wrote:

>Kevin Kinsey writes:
>> It's somewhat interesting (to me) that the even numbered releases
>> seem to have had longer and perhaps more successful runs than the
>> odd numbered ones.  5.X, in particular, seemed short to me.
>	Two comments:
>	1) Sample size = 2.

You may need to explain the math for that.  6 code trees, divided
by either the number of odds or evens, equals 3.  Still a small
sample size, to be sure, (but I doubt FreeBSD will ever have a large
enough number of released branches to ever approach anything
like the number of samples that would be required for statistical
"proof" of such a casual observation), and 6.x isn't yet "done" (for
that matter, neither is 5.X, but it's close), but I'm pretty sure that
5.x will be shorter lived than 4, and probably shorter than 6.x,
though I imagine they'll try to get 7.X out "faster" than 5.x was....

>	2) 4 has run long because it took forever to get 5 out the
>door.  While appreciative of the advances made, way too many useful
>things got delayed "waiting for 5.0".
>			Robert Huff

Agreed.  But I still see, casually speaking, an alternating length
of life, with 1.x and 3.x and 5.x lasting notably shorter amounts of time
than their even-numbered successors.  I think that core/releng
or someone has expressed a desire to not delay 7.x as long as
5.x was delayed, so there's a chance that this "trend" could be broken.

However, I take comfort in the fact that whatever branch it is,
it's not gonna be -RELEASEd until it's ready.  (And, of course,
we could bat that one around all day, too, but it's really starting
to feel like overkill --- all I made was a side comment about the
relative lengths of the numbered code branches).

It's a generalization, and a speculation, and merely an observation;
and, I think it's interesting to generalize and observe nonetheless,
but it's not measurable nor quantifiable, and not worthy of such
effort to sustain a discussion.


P.S.  The real hole I see in my "conjecture" is that, in reality,
the 5.X branch has been around a long time, being tagged
at the time that 4.0 was -RELEASED; however, it lived most
of its life as -CURRENT ....

An idealist is one who helps the other fellow to make a profit.
		-- Henry Ford

More information about the freebsd-questions mailing list