[asterisk-dev] Asterisk 1.6 Realtime Database must use ', ' not '|' in appdata field?
michael.grigoni at cybertheque.org
Fri May 30 11:51:40 CDT 2008
Daniel Hazelbaker wrote:
> 1.6 will be better than 1.4. If I didn't believe that the next
> version of a piece of software I am using (Asterisk, Mac, Mantis,
> etc.) was going to be worth upgrading to then I would start looking
> for a new piece of software. Because sooner or later I would be
> forced to upgrade.
It's too bad that there aren't more voices from the embedded space
speaking up here; embedded system development in general expects
conservative and very long lived APIs, ABIs, hardware architecture
and lifecycle. I have shuddered over the years watching Asterisk
development as it seems to be driven by a so-called 'leet' factor
pushing feature and bandwidth bloat. For those of us working with
the constraints of embedded platforms, freezing an architecture
(or framework if you like) and maintaining that is paramount.
I know nothing of the commercial asterisk offerings, but I ask,
is a commercial customer of an embedded asterisk system (from
Digium -- does Digium even do embedded systems anymore?),
forced to upgrade on each incremental release? Does that
include the expense of hardware upgrades as well? Like some
previous posters stated, a phone system is expected to run for
decades without major upgrades and expenses of application and
configuration fixes. I for example will depend on IXJ support
for the forseeable future. Seeing the deprecated support for
certain Zaptel hardware is discouraging, after all, MS operating
systems supported the ST506 disk controller for over 20 years!
As for the trend in Macintosh O/S and hardware support as cited
by the previous poster, there is a _lot_ of controversy and public
dissatisfaction in that space.
Asterisk and Zaptel on BSD are also suffering mightily again; just read
[Asterisk-BSD] for a while ;)
Expect to see more forks down the road (as it were) unless a roadmap
is adopted that assures embedded developers and others who depend
on longevity a stable architecture and framework. I still develop
in Win 3.1 and under various x86 DOS flavors because it is the
correct environment for certain hardware and embedded applications.
I use an RTOS or event loop without an O/S where appropriate, and
some of my best tools are over 20 years old. Perhaps stable
branches of Asterisk as separate 'products', with continued
'community' support and without pressure from Digium or
core developers to migrate, would be in order?
More information about the asterisk-dev