I would like to interject the following.
First of all, to the best of my knowledge, the goal of CORBA is distributed computing, i.e. the "program" doesn't really know where the objects are that make it up. I believe, in my ignorance, that there are other ways of doing this. I also don't know, at this point, how successful CORBA is. I do know that Java is an incredibly intuitive and forgiving environment that does many many things for you without a lot of effort.
CORBA has a different goal than Java for distributed computing. I attempts to make it language neutral. This results in compromises that don't have
to be made when one stays in a single language. Web Services seems to have a similar goal to CORBA. CORBA actually is highly successful and is running in a lot of places around the world in many enterprises. It simply works. In addition, effort was taking in the late 90's to create domain interface specifications in CORBA (and UML). These have not been widely adopted although they are more common than you might think. We know of one vendor that implemented the PIDS specification for their own benefit, but did not bother to tell their customer. Basically the healthcare domain specs captured the state of the industry in terms of the least-common-denominator of functionality, which is the beginning of what is required for basic interoperability. The CORBA work can be easily translated to WebServices in a standard way but becomes more than an order of magnitude more complex. Interoperability is the watchword.
Dave
