Views from an Red Hat -> Ubuntu -> Fedora migrator

[D. Hugh Redelmeier]

| From: Lennart Sorensen <lsorense-1wCw9BSqJbv44Nm34jS7GywD8/FfD2ys at public.gmane.org>

| On Sat, Oct 20, 2012 at 12:08:19PM -0400, Peter King wrote:
| > This baffles me. Why accept any half-bakedness for any time at all, in your
| > working environment? 
| 
| I certainly don't want half-bakedness under the name of a stable release.
| To me that is a contradiction.

[In defense of Ubuntu, they kind of gave a hint that the 12.04 wasn't
stable.  They recommended waiting for 12.04.1.  Still, I think that is
kind of funny.]

I like the idea of a rolling release.  I'm not sure why that isn't the
norm.

Perhaps because certain changes have cascading effects, ones that
almost require a "flag day".

- "a major transition (perl 5.12 to 5.14 for example)"

- switching to NetworkManager "without a net" (pun intended).

- switching to systemd

Could all those be reasonably captured in accurate dependencies?  I'd hope 
so.  But capturing dependencies outside the distro (eg. in users own code 
or practice) isn't as easy.

Perhaps because users are not ready for continuous change if it has 
observable effects.  Maybe they like punctuated equilibrium 
  <http://en.wikipedia.org/wiki/Punctuated_equilibrium>

At one point, releases were meant to let the vendor resell the product to 
you at intervals.

Debian's releases are so infrequent that they drive a significant number 
of users to unstable.  If unstable were billed as a rolling release, and 
everyone drew the right inferences, it would be pretty neat: the best of 
both worlds.

It's a mystery to me.
--
The Toronto Linux Users Group.      Meetings: http://gtalug.org/
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists