[asterisk-users] * 1.4 -> 1.6, zaptel -> dahdi

[Greg Woods]

On Wed, 2009-08-19 at 11:49 -0700, Steve Edwards wrote:

> But, when the pain of using packages exceeds the hassle of the source, 
> I'll use the source without hesitation.

Agreed. I use trunk for MythTV, for instance, because there are features
not in the latest packaged version that I really do need (VDPAU and
HD-PVR support, for those who know about such things). And it's a pain
to upgrade, which of course needs to be done relatively often when you
are following the development version. So I do know how to do this and
what's involved. But I'm not convinced I've reached that point yet with
asterisk. It's true that I wouldn't need the development branch, but
it's still more hassle to apply security fixes when compiling from
source. And in this particular case, the conversion I am contemplating
is one that is going to have to be done eventually. When the developers
want to convert the config language, sooner or later they will stop
supporting the old stuff and it won't be possible to get the newest
supported features without converting. So I don't want to fall TOO far
behind. I'd rather convert now when I can do it in a leisurely manner,
before I absolutely HAVE to.

> I don't like being dependent on others 
> for my mission critical stuff.

Unless you are one of the developers, you're still dependent on others
to maintain the source. Granted it does eliminate the folks in the
middle doing the packaging. And in my case, nothing on my home server
could really be described as "mission critical". It's closer to what the
MythTV folks call the WAF (Wife Approval Factor :-) She likes the
features that asterisk offers (separate voice mail boxes, phones in
every room, intercom ability to call between phones, having a backup
VOIP line out so we can make two calls at once, etc.) but none of that
is critical to her use of the phone. We can live without asterisk for
short periods of time, so it's not really "mission critical". I have an
old-fashioned answering machine that will take messages if the computer
fails to answer (it's crashed, asterisk isn't running, we've powered it
off for a vacation, etc.) and I can patch the cordless phone base
station through to the wallplate (although then the VOIP phones don't
work).

As always, in the end, we make a tradeoff between security concerns,
reliability concerns, having the latest and greatest features, and ease
of long-term maintenance. What that means exactly is determined on a
case-by-case basis. In my case, I'll run the old system until I can find
a way to convert to the versions packaged with Fedora 11 relatively
quickly, to minimize my window of vulnerability and down time. Then for
the longer term I'll work on the AEL conversion. I was just hoping to
find something that would aid in that effort, but so far all I have seen
are suggestions on how to avoid having to do it at all.

--Greg

>