[Nick Danger]

I'm unconvinced. No one in the C++ world would tolerate a new C++ release that couldn't run old C++ code

They would, and they did, at least for Windows programming. Both Microsoft (MFC) and Borland (OWL) totally re-wrote their class libraries, leaving all kinds of stuff in the dust. 'Twas a major pain in the butt. Especially after they told us that the class libraries would isolate us from changes in the Windows API.

[Southack]

I'm unconvinced. No one in the C++ world would tolerate a new C++ release that couldn't run old C++ code

"They would, and they did, at least for Windows programming. Both Microsoft (MFC) and Borland (OWL) totally re-wrote their class libraries, leaving all kinds of stuff in the dust."

So MicroSoft has now broken backwards compatibility twice?! For VB 6 to VB.Net and for C++ when they rewrote their Foundation Classes...

Now you're talking about a corporate *pattern* of breaking backwards compatibility. Certainly MicroSoft is also going to do this again soon with either their VB.Net or with their VBA developers, too.

Ugly. It's just pretty ugly to see this sort of behavior.

How can I recommend MicroSoft solutions to my clients when I can't in good faith depend upon our current work being supported in the future? How many times can I go to my clients and explain that *they* have to pay us for re-writing existing projects...simply because MicroSoft changed the rules in mid-stream?