>> I don't how things work in VC++, so I can't comment on this.
>
>When C++ was first introduced into the PC development market, the code generators produced C/ASM code that could then be compiled with the C compiler. It wasn't until later that people like Walter Bright of Xorland (later Xortech) produced a native C++ code generator.
>
>I'm guessing that a lot of fundamental changes have to be made to the back end code generator to support OO, just look at the exports generated by a C and C++ compiler and you will see what I mean (name mangling etc).
>
>It is a lot easier to make these changes to P-Code in an interpreted language than to a native X86 compiler.

I'll accept your word on this since it is something that I don't know about. I did read a book about software project management by the guy that was in charge of VC++ 1.0 (sorry, don't remember his name). Borland C++ was number one at the time and MS wanted that position on the first release. VC++ is pretty much written from scratch.