[Serdev] SER architectural decisions - was: So who/what is SER for, anyway?

[Dragos Vingarzan]

Greger V. Teigre wrote:
>
> I don't really agree with that evaluation. If you look at SER 2.0, you
> will for example see a completely different data model, as well as many
> discontinued modules. There are many things that can be done, but I
> think most people agree that a total redesign is never a smart thing to
> do, it takes forever to finish.
>
I agree with you, so again, it is not my intention to bash SER. I think
that it is a great piece of software. How else could it be if it does
the job 10 times faster than the commercial counterparts?

Yet, personally, I would've expected that over 0.10.99, 2.0 would've
been a completely new thing...

Many things changed for better, of course, it's just that I personally
would've welcomed more deep changes.
>
> To be honest, I also believe in a model where one person has an overall
> responsibility for end-to-end (like Linus for the Linux kernel).
> However, this person should only have the open-source project as
> priority and make final decisions where no consensus can be reached (or
> overrule in some special cases).
>
For the sake of SER, one of the companies behind it should finance this
"Linus" unconditionally. Well, not even FOKUS was perfect at this as
commercial reasons pushed some things more than they should have, though.

Of course, commercial guidance is very important and shouldn't be
discarded as else the product won't be usable in the real case, but too
much might be noxious.
>
> So, Dragos, if you were that lead developer, which difficult decisions
> would you take?
>
I am far from that position and I really don't understand SER that well
as the core developers, but I would sit back a little and just be the
"lazy-developer". That is, figure out what are the things that
developers and users are struggling and spending most time with and fix
them.
>
> And then, which interests do you believe would suffer under those
> decisions?
>
When you do changes, the "old" people would always complain. They have
customers that are currently paying and the risk of loosing them is too big.
>
> If none, that's good :-)
> If there are, how do we organize SER so that good and balanced decisions
> are made?
> g-)
>

This might sound crazy, but if it is so hard and slow to gradually
evolve SER, what do yo say if a new branch would start? One that would
take a lot of code base from the current one, yet problematic parts
being rewritten from scratch. New internal APIs, new module interface,
new scripts, etc. There are a lot of things that are very good now, so
it won't take forever to get it working. I would happily participate and
also move the Open IMS Core to the new project to help testing.

And was there enough watter under the bridge so that this next gen SER
would see a reunification of the SER/OpenSER teams?

-Dragos