[asterisk-users] where is 1.4.12?

[Stephen Bosch]

Jared Smith wrote:
> On Thu, 2007-08-30 at 08:02 -0500, Eric "ManxPower" Wieling wrote:
>> As I understand it, Digium does NO formal QA testing before the free 
>> Asterisk/Zaptel/libPRI releases.  Asterisk Business Edition is a 
>> different story and gets extensive QA testing.
> 
> As I understand it, that's simply due to a lack of resources.  At the
> Asterisk Developer's Conference earlier this year, the Asterisk
> Developers were all pretty much in agreement that more needed to be done
> in this area, but that it would have to be a combined effort between the
> Asterisk community and Digium, as Digium simply doesn't have the
> resources at this point to do it all itself.
> 
>> On IRC I have been a "vocal user from hell" about the QA issues of 
>> Digium open source products.  
> 
> I've tried to be vocal about this too.  And now that I'm working for
> Digium, I'd be happy to try to coordinate an effort between the
> community and Digium to try to come up with a framework where we can all
> work together to make this happen.  

That's the spirit.

I just wanted to throw in something.

That a product is commercial is never an assurance that it is or will be
stable. It doesn't matter if the product is from IBM (which spends a
billion dollars on R&D annually) or Cisco or Nortel.

Commercial products break, no matter how much they cost.

Most vendors do a careful job of obfuscating the instability in their
own product. Depending on the depth in their technical staff, they will
solve the problem quickly or slowly, or offer you some limp workaround.
That is somewhat correlated with the cost and class of the product. If
you believe, however, that paying for a product means that it will work
reliably or as promised, you are living in a Madison Avenue-induced haze.

Either way... it shouldn't be an excuse for us or Digium to accept less.
Take the Linux kernel -- there is a community project with a rigorous
vetting process, and I would say the Linux kernel is extremely stable.

We can and should introduce a similar rigor for Asterisk. A big step in
that direction is patience and focus. The "creeping featurism" could
make way for an increased concentration on reliability. That's a
development roadmap thing.

Cheers,

-Stephen-