[asterisk-users] extensions.conf vs. AEL

[Tilghman Lesher]

On Thursday 04 October 2007 07:07:47 Barzilai Spinak wrote:
> All this discussion is pointless. As pointless as the discussion of
> assembly versus high-level languages decades ago.

As one of the main architects, I don't find this discussion pointless.  My
personal opinion of AEL is that it's coming along nicely, but it's still not
up to the point where I would consider using it for most dialplans.  That day
will come, and I'm working with Steve Murphy to ensure that it does.  One
thing that you did not see in the language wars of yesteryear was of the
assembly language changing in subtle ways, to make development in the
higher level language easier or more consistent, as is the case with AEL and
extensions.conf.

> Except most people rooting for "extension.conf" don't even have the
> technical and conceptual amplitude to understand what they are talking
> about... they just want some telephony system to make a quick buck, or
> save in their LD calls...

This seems like a rather harsh indictment, when it really comes down to the
fact that writing in extensions.conf works today, and while AEL does work to
a certain extent, many people would rather not have to rewrite their dialplans
every time an architectural flaw is found in AEL that limits what they can do;
ergo, they write their stuff in extensions.conf until the point where AEL
becomes more trusted.

> A lot of Asterisk is technically and architecturally twisted, and
> "spaghettied", and with many redundant ways of doing the same thing (in
> different stages of obsolescence, incompleteness, and (un)documented).

As a maintainer and architect, I would very much like to hear specific
criticisms on how you think this could be improved.  We try to deprecate
specific functionality that doesn't work correctly or which could be expressed
in better ways, which allows users of the system to transition away from those
expressions to better methods over a period of time, instead of immediately at
an upgrade; we believe this facilitates adoptions and upgrade processes.

> At least AEL is a step in the right direction (even though it has to
> adapt itself to the ugliness that exists below..)

All high level languages have to adapt themselves to the ugliness below.  That
is part of what makes them high-level languages.

-- 
Tilghman