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

[Andrei Pelinescu-Onciul]

On Jan 25, 2007 at 18:40, Dragos Vingarzan <vingarzan at fokus.fraunhofer.de> wrote:
> Andrei Pelinescu-Onciul wrote:
> >
> > modules were there almost from the beginning. It's true however that ser
> > outgrew the current module interface and it needs a change.
> > Unfortunately we cannot make this change overnight and that's why we
> > have to live with numerous hacks.
> > [...]
> > All this stuff will take a lot of time, but it will eventually happen.
> >
> > Andrei
> >
> given the whole inheritance thing... could they live in parallel for a
> while? Like a parallel module interface with it's own new script? Even
> if you won't be capable of mixing usage of old with new modules... or is
> it just too much to push at once?


Yes they could. The new module interface structure will have a different
name so that the module loader in core could make the distinction
between old stuff and new interface. The same module could even have
both old functions and new ones.

The problem here is time. IMHO the module interface change is linked to
the others (especially script engine and avps) and it should all come
more of less together. This will take lot of time and there are a lot of
more pressing problems (like small rfc conformance problems, the need to 
 rewamp the locks, more dns stuff a.s.o) so it gets continuosly delayed.


Andrei