Future of the Component Object Model?

Question

Folks,

Please has anyone heard anything about Microsoft [not] supporting COM on future version of Windows?

I imagine that Microsofts engineers are stuck with it (even if they would prefer to only develop and support the .NET framework) simply because of the enourmous backlash from there huge customer vase. There must be literally billions of lines of COM-based stuff out there in the wild. I know I'm NOT looking forward to being pushed into a mass-migration simply because some smarty-pants over-achieving architectural astronaut has to prove he's The Greatest... (sheesh, why don't they just take up boxing?)

I suspect I'm just being paranoid ;-) but can anyone provender any authorative links (google finds "not much")? Ideally, a White Paper from Microsoft saying "COM Stays in Vienna, or else!" would settle my nerves greatly.

Cheers all. Keith.

Answer 1

Well the reason they called it .NET is because COM3 was taken as a serial port name. .NET is the new COM. From .Net Common Language Runtime Unleashed:

The name change from COM3 to COR to COM+ 2.0...to NGWS and finally to .NET.

The reason the primary assembly is called mscorlib is that it stood for Microsoft Common Object Runtime Library.

Answer 2

COM is still THE technology for inter-process-communication. Look at the way you can control Word, Excel, etc from another application. There is no way .Net can replace this.

Answer 3

I can't see COM disappearing anytime soon as there is an enormous amount of legacy COM-based code. For the same reason I expect to see Win32 hanging around for the forseeable future. No-one cares about Windows as such as a platform, they care about their application software.

COBOL on a S/360, S/370, S/390, zSeries and C/unix remain very widely used platforms for the same reason.

Answer 4

There are no any official statements that COM support is being dropped from future versions of Windows. it is used extensively throught the OS internals. Baring full rewrite of the OS from scratch (which I can't envision any time soon), it is safe to assume COM will be there for a while.

Answer 5

COM and .Net serve different needs. As long as there is native code there will be a need for a binary componentization standard, i.e. COM. Even if the OS were rewritten from scratch (which it won't be, nor should it) it would primarily be native code for reasons like performance and versioning. You would quickly need to invent something just like COM, so why not keep the one that has been tested and works?

Answer 6

So to summarize:

COM will stay because: 1. It fundamentally works 2. It handles interprocess communication, as distinct from intermachine communication. 3. .NET inherited much from COM, but not everything. 4. Even Microsoft themselves still rely on it.

And I take the point that there's still bazillions of lines of FORTRAN and COBOL in production systems today... because they fundamentally work.

Thank you all for your answers... Mods, feel free to clean up this conversational response, but (IMHO) this site would do well to allow posters to publicly thank those who take the time to respond.

Cheers All. Keith.

Answer 7

I know its easy to think that the Microsoft world is .NET only from now one, thanks to the over-excellent work their marketing department has done, but Microsoft always supports their old stuff, they have no choice really.

Look at MFC, they've released new packages, and the RibbonBar only works on MFC (as office is developed in it). Sure, they will start writing more and more .NET code over time, and less and less COM stuff, but they will still support it.

The COM development features in Visual Studio will remain in place, when they go, that's the time you know they no longer want us to use it.

I read a blog entry from the Visual Studio project manager where he said that they had received a lot of complaints from developers concerning the focus on C# recently. He agreed and said the next version of Visual Studio would be focussed primarily on unmanaged C++ development.