Registered: Nov, 2003
Re: CORBA's place in architectural direction
Posted: Apr 7, 2006 9:47 PM
CORBA is widely considered an obsolete middleware technology for most applications. The exception being high performance real-time distributed applications where there are not really any good alternatives.
CORBA suffers from a number of defects relative to more modern approaches. These include:
1) An ugly programming model with data structures that do not map cleanly into all language bindings.
2) Diffferent binding models for each language.
3) Client code must compile in "stub" code.
4) The standard is vague in too many places so no two ORB implementations behave exactly the same way. There is also no "reference" implementation.
> This question is being reposted here as the Design Forum
> sadly seems rather inactive.
> Reading this weblog on Software Architecture brought
> thoughts about CORBA and where it stands today as an
> enterprise middleware technology. It seems CORBA has
> gained more popularity in the academia, and very little
> these days in the industry. On the contrary, Jini which
> was slow to pickup, seems to be gaining some momentum in
> the enterprise.
> The question is, for those involved in architectural and
> technological direction and those interested in it, where
> does CORBA, once the solution to all middleware problems,
> a technology that is platform, vendor and language
> independed, stand today when compared to more newer
> technologies like J2EE, .Net and Web Services (which don't
> seem to offer some of the advantages)? Does it still play
> a role in the industry and do people ever consider CORBA
> as the technology of choice today when strategizing their
> new architecture? Or is it considered legacy?