Michi Henning's article The Rise and Fall of CORBA in the last issue of ACM Queue has led to some discussion, in the CORBA community and elsewhere.
Now Steve Vinoski (co-author with Michi of Advanced CORBA Programming with C++) has joined the discussion.
While Steve agrees with Michi on the standardization issues (members of WS-* working groups, pay attention!), he criticizes Michi for changing his opinion, U-Turn style, on the technical merits of CORBA, apparently for business reasons. After all, Michi's company sells a competitor to CORBA. Now, Steve still works for a company selling CORBA products (although he is not working on them anymore), so I think his opinion is not entirely unbiased, either.
Personally, I never liked CORBA that much. The only usable thing in CORBA, the remote method invocation stuff, can be done in a much simpler way (Java RMI, for example), the C++ binding is brain-dead and all the rest is committee-ware in the worst possible form.
I don't know much about Corba or Java RMI but have had some exposure to DCE RPC in the past (~8 years ago). That being said, this post is a question rather than comment:
What's your opinion about DCE RPC, especially in the light of the fact that it is open source now?
I haven't had much exposure to DCE RPC, so I cannot really comment on it. I played around a bit with Sun RPC (which is different from DCE RPC) years ago and it basically worked (as evidenced by all the technology built on top of it - NFS, etc.). Should I find some time (by current estimate, this will be around 2010 ;-) I will look at the DCE RPC open source release.
Posted by: Guenter at June 26, 2006 08:34 AM