[asterisk-users] SER + Asterisk

[SIP]

Alex Balashov wrote:
> SIP wrote:
>
>   
>> Seriously, though... this seems to be a popular misconception. I hear it 
>> a lot. Where did you come across the information that SER is no longer 
>> developed?
>>     
>
> That seems to be a consequence of looking at the releases.
>
> Anyway, I spoke too soon in saying that there's absolutely nothing going 
> on with the project whatsoever in terms of development.
>
>   
Yes... I'll agree the releases are a bit... odd. SER 0.9.6 (or possibly
0.9.7 -- I'm never sure) was the last actual 'release' labeled stable.
However, SER 2.0 rc1 has been available for over a year now, and hasn't
been granted that 'stable' label, even though I gather it's no more
unstable than 0.9.6/7. All the while, work on SER 2.1 is commencing long
before there's been a release of SER 2.0. It's incredibly difficult to
follow.

But this is where the OpenSER (now OpenSIPS) and SER projects differed
in their ideology most often -- that of releases and documentation. SER
was always a bit sparse on both, preferring to make up for it by way of
solid innovations in the core code.

Unfortunately, it's a bit like the tale of Seymour Cray. Here was a man
who was convinced that if you built a supercomputer, people would buy it
because it's the fastest thing out there, and building peripherals
and/or software for it as part of the business plan was a waste of time
and money. This ideological difference is why he left Control Data. This
is why he was encouraged out of Cray Research. And this is why his final
company, Cray Computer Corp failed -- that sort of missed idea that
people will buy technology simply for the sake of having better technology.

I see a lot of parellels there with OpenSIPS and SER. OpenSIPS is a
stable plaform that has dozens of modules and documentation galore on
how to mesh the system with this, that, and the other.  SER has
rock-solid, incredibly innovative core code, but prefers to leave the
writing of modules and documentation as an exercise for the user,
thereby making it perhaps overly difficult for anyone to implement or
integrate.

N.