Can we do VTable short circuiting at all?

[Phil Dawes <philipd parallax co uk>]

It is now looking pretty certain that we can't do inproc vtable
short-circuiting within the POA and still remain conformant. (e.g.
section 9.2.10)

But this is still something I'd like to be able to do with orbit. This
is because orbit is the gnome orb, and as such is likely to be used
primarily for local and in-proc objects.

Could this be done through another object adaptor which deals only in
local proc shared library objects? 
An entry in the IMR could state whether the object exists in a shared
library or as an exe-server and the orb could then use this information
to activate the object throught the appropriate OA (I think).

Would it be feasible to write this behavior as an extension to the POA
so that component object implementations are still written as
POA_Servants, and can thus also be connected to a poa as well (if
invocations come in remotely).?

The reason I'm so interested about inproc fast objects is that I wrote a
sound-tracker program called OpenTracker which takes plugins to handle
the rendering of sound data (e.g. samples, constructive waves) and
allows the addition of special effects. 
Because this is a real time application the invocation mechanism must be
very fast - there are lots of invocations a second. Obviously straight
CORBA is not a good fit here, but I'm keen to have a standard
architecture for gnome so that every application that requires a plugin
architecture doesn't have to re-invent the wheel.

Also, if we do embedded gui components (ala activeX and opendoc) we'll
also want fast invocations since there's a lot of traffic between
objects in negotiating for space etc..


Just some thoughts,

Cheers,

Phil.

-- 
_______________________________________________________________________
 Phil Dawes                               |   My opinions are my own
 WWW:    err.. temporarily non-existant   |   and nothing to do with
 Email:  philipd@parallax.co.uk           |      my employer.