Re. http://mail-index.netbsd.org/netbsd-help/2002/06/09/0004.html

If I recall correctly, 1.4.3 was released *after* 1.5 was released.

The thing to understand is that development isn't linear.  There is on the
one hand fixing bugs (what the 3rd digit represents, essentially).  On the
other hand, there's adding features (which necessarily introduces new
bugs); this is the middle (or maybe even someday the first) digit.

As I understand it, a new ``major'' release represents everything that is
presently in -current.  It gets stabilized, built for many platforms,
tested, fixed up (and maybe some of it pruned, conceivably?), then it is
released.  The ``minor'' (or ``patch'') releases are fixes for a past
release.

So why continue work on 1.5.x when 1.6 is just around the corner?

Because 1.5.x isn't perfect.  Many people will continue to use the
well-shaken-out 1.5.x rather than switch to 1.6 with fresh bugs.

For me (and perhaps you) 1.6 will be a natural choice when it's available
(I may have time to test a snapshot or two, but I am not presently set up
for cleanly installing test versions to play with).  For people running
production servers that have to work reliably and consistantly, a more
conservative 1.5.x makes more sense until the 1.6 release matures a bit
and (inevitably) has a patch release or two to fix the almost-certain bugs
that will be turned up once it's in widespread use.

(As an example, 1.4 involved replacing the virtual memory system.  1.4,
proper, crashed a lot for me.  1.4.1 came out *very* quickly and fixed the
problems that I was seeing.  That's okay (not great, but okay) for a home
system.  For a production server, though, it would be terrible.  Had those
crashes happened during testing, 1.4 would surely have had them fixed,
rather than requiring 1.4.1 so quickly after.)

If I were running a system for other people, I'd be much more cagey about
*any* major new release of *any* product---even NetBSD.  (^&


That, at least, is how I understand the picture.


  ``I probably don't know what I'm talking about.'' --rauch@math.rice.edu