Re: [Fresco-devel] OSG-Prototype

cinap_lenrek wrote:

>> while this is indeed the idea behind the efford to encapsulate the 
>> scene-graph
>> inside the server, I don't see how passing an id is any better than 
>> passing
>> an object reference (in fact, where is the difference ?). 
> 
> 
> CORBA-object-ref's are untrusted. iirc here is no way to check where the 
> object on the other side is
> implemented without calling into such an object. a call to an 
> CORBA-object is dangerous because you don't know
> where you relly call in. what happens if the client instances his own 
> implementation and does a big sleep(100000000000); in
> it?

you are quite right. For get for a moment that we are dealing with CORBA.
We don't want to make the scene graph accessible any more. CORBA here is
really just used for two things (and even they will be replaced one day):

* the transport protocol
* the type system / object model, i.e. the high level communication protocol

If you only exchange IDs you have to do all the type checking yourself.
Whatever replacement there will be for CORBA, it will have to provide
these facilities for us.
And as to the location transparency, yes, we do know that we are calling
an object that lives in the server. The proxy objects are client specific
and not sharable.

> The Idea is to use CORBA only for the public client-interface. If we 
> want to do things on the server-side internally
> (for example connecting nodes), we can bypass CORBA by using nativ C++ 
> method calls on this Impl-Classes.
> we don't need to make the whole interface public to CORBA.

yes, that's exactly the point.

Regards,
		Stefan