[asterisk-users] Stable-Stable Asterisk

[Russell Bryant]

Steve Totaro wrote:
> 1.0 is over-ripe or rotten/forgotten and thrown away.  Besides, we all 
> know that 1.0 was just a marketing ploy to legitimize Asterisk.  What 
> serious company is going to install 1.BETA2 or .90?  Maybe a 
> nonessential piece of software but not something as mission critical as 
> a PBX.

A marketing ploy?  Are you serious?  Asterisk 1.0 was the point in time where
there were more than enough people using Asterisk in production to justify
maintaining feature frozen releases.  It meant that those people no longer had
to use the development code on their production machines and could start
focusing on a release that wasn't a moving target.

I maintained Asterisk 1.0 for about a year and a half through my own personal
*volunteer* efforts.  It was absolutely not a Marketing ploy, as I didn't get
paid for those nights that I stayed up all night reviewing the bug fixes that
Mark had been making in the development code so that I could backport them.

> 1.2 is sitting at the fresh fruit market.  It is a nice peach color, 
> soft, sweet and juicy, most of the bad peaches have been discarded such 
> as worm and bug infestations.  It has been aged perfectly.
> 
> 1.4 is still a bunch of peaches on the tree.  It is far from ripe and is 
> still very green.  It is prime lunch for bugs, worms, and other 
> infestation which will not get sorted out until they get ready for the 
> market.

Mmm ... peaches.  Anyway, this transition from Asterisk 1.2 to 1.4 has been a
very interesting learning experience.  We will definitely benefit from all of
this when it comes around to the next time that we do a major release.  I have
really come to understand the different expectations of stability that people
have of their phone system versus other software.

The transition between 1.0 and 1.2 was a different animal.  There were some
really major features added between 1.0 and 1.2 that a lot of people decided it
was worth running the development tree to get as opposed to waiting for 1.2 to
be released.  The realtime configuration architecture is one example.  So, 1.2
got a lot more production use before it was actually released, and there wasn't
quite the same flood of people all starting to use it at once like we have had
with 1.4.

Now, with Asterisk 1.4, I think we have a couple of challenges.  We have the
fact that there are now a *lot* more installations out there than there was at
the time of 1.2 being released as the project is growing rapidly.  Also, I think
a lot more people have been content with the feature set of 1.2 and haven't been
as eager to upgrade.  So, 1.4 didn't receive as much production use before it
got officially released.  It has hurt a bit during the early months of 1.4 as we
started dealing with various major issues.

I would also like to note that the development team did recognize the difference
in the situation we had at hand.  These are the exact reasons we decided to
fully maintain Asterisk 1.2 during the first 6 months of the life of 1.4.  When
Asterisk 1.2 was released, 1.0 was immediately deprecated and only maintained
with security fixes.

I am now feeling very good about Asterisk 1.4.  When we had our developer
conference in May, we talked about a lot of cool things.  However, we also
talked about how it must be a priority that we fix bugs and decided to work
extremely hard on bugs for the Summer.  We lived up to our word.  The past few
months have seen a *ton* of serious issues get resolved, and I am very pleased
with our progress.

To the whole user community, thank you very much for your support and patience
with us as we push Asterisk forward.  Feel confident that we will not leave you
hanging.  We will continue to do whatever we can to make Asterisk stable as we
further improve functionality.  If the changing needs of the user community mean
that Asterisk 1.4 needs to be maintained for a full year after 1.6 is released,
then so be it.

Thanks for reading,

-- 
Russell Bryant
Software Engineer
Digium, Inc.