Fog Creek Software
Discussion Board




Borland supporting Microsoft effort to mangle C++

http://www.eweek.com/article2/0,4149,1330422,00.asp

====== quote =====

JP LeBlanc, vice president and general manager of the mobile and C++ solutions group at Borland Software Corp., Scotts Valley, Calif., said, Borland is very supportive of Microsoft's effort to push a C++ CLI binding standard through ECMA.

"We are a big believer in standards and by Microsoft taking this step it will help us better support their platform for C++ developers," LeBlanc said. "The significance for C++ developers is big. A huge problem of C++ on Windows in the past has not only been conformance to the C++ language specification but also binary compatibility and interoperability between the compilers and their associated C++ runtime environments."

==== end quote ====

I guess Microsoft's strategy is to embrace-and-extend C++ to make it a proprietary standard. It's not enough that they have C# to screw around with, why must they attempt to destroy an open standard like C++? I can't believe Borland is helping them do it too.

gcc
Friday, October 10, 2003


Interesting.

Extensions pushed by scientists who happen to work for a corporation are evil?

Extensions pushed by scientists who work in "open software" are altruistic and good?

Aren't today's standards yesterday's extensions? Whether their adopted/assimilated depends on a critical mass getting behind it, no?

What are you afraid of?

anon
Friday, October 10, 2003

similar view as anon. do you think the guys who has the knowledge to influnce c++ want's to destroy c++? they probably put a lot of efforts into learning that.

and they are people too. open standards doesn't mean complete democracy, and doesn't mean all features you propose are good ones.

na
Friday, October 10, 2003

You'd think that someone who uses "gcc" as a nickname would know the standard well enough to know that it allows for extensions.  The typical response of the standards committee or Bjarne to proposals such as metadata, garbage collection, binary interoperability, etc. has been that these would be best taken care of by third party extensions, which is exactly what Microsoft is doing (apparently with Borland's support).  The destructor for type_info is virtual specifically to support third party extensions (side note: does anyone know of any compilers that actually take advantage of this?).  This isn't to say that the use of third party extensions is encouraged but when necessary they certainly aren't discouraged.

Microsoft has done a tremendous amount of work on making their C++ compiler standard compliant.  Standard compliance and extensions are not mutually exclusive.   

SomeBody
Friday, October 10, 2003

Also, I find it interesting that you felt the need to chop off the important last paragraph in the Borland's quote.

SomeBody
Friday, October 10, 2003

The funny thing is that Microsofts claims that .Net supports multiple languages.  The truth of the matter is it supports 1 language and multple syntaxes.

My major gripe with the .net platorm is it puts too much emphasis on inheritence and not enough on aggregation.  Herb Sutter, who is now at Microsoft , has been saying for years that inheritence should be used sparringly, but it seems those responsible for the platform didn't get the memo.


Friday, October 10, 2003

Herb Sutter is a frigging genius (guru really).  Geez, I didn't know he worked at MSFT.  Well, they certainly have the right guy to do the job...

nat ersoz
Friday, October 10, 2003

Somebody,
you spoiled the troll.

Baitbuster
Friday, October 10, 2003

Herb Sutter is the best of the C++ "gurus" I believe.  He is the most pragmatic.  Although I've found implementing his "strong exception" garentee in practice nearly impossible.  That's a discussion for another day.


Saturday, October 11, 2003

*  Recent Topics

*  Fog Creek Home