[Serdev] A suggestion for how SER should focus - was: So who/what is SER for, anyway?

[Dragos Vingarzan]

Andrei Pelinescu-Onciul wrote:
>
>
> Writing ser from scratch would be a huge effort and with very little
> benefit.  There are still lots of problems that needs solving in the
> current ser version and you'd want us through everything away and start
> again, waisting a lot of time.
>
no pb, just wanted to get an answer.
>
> >    - design an improved architecture that takes into account the
> > changes in the last years in the usage patterns
>
> Could you detail the changes in the usage patterns?
>
well, I am not sure anymore that stateless proxies are so much required
now. Even statefull, with the latest hardware you can handle the traffic
for a 3rd world country on a single machine. I understand performance,
but this seems that you're trying too hard now.
>
> Ser is not an exclusivist tool (you can start it with a pre-existing good
> config, so even inexperienced users could use it for simple stuff), but
>  if you want special configs, then yes you have to know a little about
>  sip, ser configs and modules.
>
I disagree here a little. To do the standard registrar scenario maybe
you're just fine, but when you want to change the script a little more,
you need to know the many RFCs very well, to read lots of script
examples, keep up with the latest bug/fixes/issue/workarounds, with the
best practices, follow closely the mailing lists, etc. Then if you want
to do more you need pretty good programming skills and LOTS of patience
and time to dig through the code in order to find answers. The learning
curve is steep, long and completely unrewarding once you get high on it.
>
> Actually I'm very curious what the OpenSer developers would say about
> this. It might be that only you see the need for a big change.
>
I don't think that their clients and direction is much different, so I
would also expect that... oh well...
>
> Some of us are not really interested into what you call NGNs and don't
> have a good opinion about IMS in general (on the contrary).
>
me neither, but I feel the need for a change. Everybody assumes that I
would be IMS's advocate. Well, I am not. I am just a guy with a mission:
find out if IMS would practically work. I see a lot of issues everyday
with it and I don't think that it would be _the_ panacea for telcos.

-Dragos