C++ from Microsoft???

[Bob Stout]

In an article of <3 Feb 90 07:41:11 GMT>, (Cay Horstmann) writes:

 >And the beta-test of their new C compiler (the one with the -P option which 
 >will compile C++ 2.0) comes a Turbo Debugger that knows about classes and 
 >inheritance. Unlike Zortech, it has a virtual 386 mode driver (or unlike 
 >Microsoft, for that matter--you need a third-party add-on, Magic CodeView, 
 >to run CV in a virtual machine on a 386.) Are you listening, Walter? 
 >
 >I think Borland is really going to beat Microsoft this time around.

  As someone whose hobby is weasling information out of beta testers, I find  
this interesting. The most recent information I have heard on TC 3.0 (if  
they're really confident in its C++ utility, why is it still TC?) is that  
those beta sites with significant applications in C++ have yet to successfully  
compile them. I've even heard this from a rabid Borland apologist/Zortech  
hater, so there must be something to it. MSC 7.0 (the C++ release) was going  
to be ready about the time MS had Programmer's Workbench ready to be able to  
release MSC 6.0, but is being delayed as well - probably until late 3rd qtr at  
the earliest. Borland's undoubtedly going to beat MS to market, but I think  
that reflects less conservatism rather than more advanced technology. As C  
compilers, both new products should be good. It's only their C++ technology  
that remains a big question mark.

  As far as Walter is concerned, it's public knowledge that he's been actively  
researching '386 and '486 code issues recently. Zortech's also said that the  
current 2.06 release of ZTC++ will probably be the definitive release while he  
works on the next major release. As a Zortech beta tester, the secrecy shoe's  
on the other foot here, so all I can say is what's already generally known.

  Finally, the one thing all vendors are doing that worries me is something I  
might not have expected with the adoption of the ANSI standard. Since C  
compiler technology has reached a level where all of the major compilers are  
pretty much competitive, everyone in the PC market seems to be looking for a  
competitive edge by extending the language. When Walter did this with his  
__handle pointers, it seemed like a pretty good idea. As I hear of more and  
more incompatible language extensions being added to MQC 6/QC 2.5, TC 3, JPIC,  
etc., it's starting to disturb me.