[omniORB] COM-omniORB Interface

[Eric Corbett]

Have you considered VisiBroker instead of OmniOrb in this environment?
You can download a version that runs just fine under Borland C++ Builder
3 from Borland's (or is that Inprises's now) Web site.

In message <355cfabf.22759477@smtp.ne.mediaone.net>, David Morgenlender
<dmorgen@alum.mit.edu> writes
>I'm in the process of developing an app, consisting of 2 pieces:
>
>1.  CORBA server app running under ETS-Kernel (a realtime OS supporting the
>Win32 API).  This is currently under development using Visual C++ 5.  I'll
>hopefully know soon how well omniORB works in this environment.
>
>2.  CORBA client app running under Win95 (later NT4).  This is currently under
>development using Borland C++ Builder 3.  I'm running into problem after problem
>getting omniORB working under BCB3 ... I keep running into compiler bugs.  
>
>Until a couple of days ago, this was the entire system.  However, then I learned
>the client would like to be able to write Visual Basic apps to control the CORBA
>server app, e.g. for Q&D diagnostics.  
>
>This left me in a quandry.  At first glance I need to find some way for a VB app
>to use CORBA.  VB handles COM well, so I could acquire & test a COM-CORBA bridge
>product.  But I suspect this is asking for more headaches, both getting it work
>& in ongoing support.
>
>However, there's a solution which solves all my problems (hah, hah <g>).  I can
>write a VC++ exe which uses omniORB on one side to function as a CORBA client,
>and COM on the other side.  This provides support to a VB app.  It also allows
>my BCB app to communicate with it via COM;  voila, no need to get omniORB
>working under BCB!  Theoretically, BCB provides good support for writing a COM
>client.
>
>There are some negatives though, e.g.:
>
>1.  Another layer:  complexity, overhead, etc.
>
>2.  A change to an interface function requires changes in 3, not 2, places:
>CORBA server, COM-CORBA app, BCB app.
>
>3.  Just one more buzzword to support:  COM.
>
>What do you think of this approach?
>
>Can you suggest any better alternatives?
>
>=======================================================
>Dave Morgenlender
>e-mail: dmorgen@alum.mit.edu
>=======================================================
>

-- 
Eric Corbett