savas parastatidis

“IDL is as IDL does”

Totally agree with David Orchard's comments in his "IDL is as IDL does" entry, which refers to our "Why WSDL Is Not Yet Another Object IDL" article. As I said in my reply to Mark Little's comment, Jim and I are talking about the use of IDL in object-oriented distributed environments.

One response to ““IDL is as IDL does””

  1. Mark Little
    You’re right that there is a distinction between IDL and object-IDL (didn’t I say that several posting ago?) However, if you think about what actually happens at the level of receiving and dispatching those messages (dispatchers, listeners, whatever you want to call them), then those things are typically multiplexors, taking an incoming message and sending it to an object. But what’s the object do when it receives the (message) invocation? Hmmm, well it could dispatch it to a different object or method or service for each operation, and even (shock horror) on a per operand basis for each operation. And what’s worse 😉 the things that it dispatches to could even be on different machines, in different parts of the world and in entirely different languages, maybe interacted with using different comms. protocols. My point is that you can make “higher up” the protocol stack behave in the same way as “lower down” and achieve the same benefits of loose-coupling. I think your point is better made by saying that the traditional programming paradigm used by object-IDLs is strongly-coupled, whereas that used by WSDL isn’t. But it’s wrong to impy that object-IDLs are not flexible enough.