A malo proricanja? Kakva je, po vama, buducnost .NET-a? Hoce li ga M$ uspesno izgurati, ili ce Linux+Java(+Oracle) profitirati?
Naleteo sam na clanak (tj post na forumu) na netu o tome i ovde ga prilazem. Dosta zanimljivo misljenje...
The attack of the clones
Posted By: Michael Jouravlev on July 22, 2004 @ 04:34 PM in response to Message #131255 2 replies in this thread
.NET provides incremental improvements on J2EE in my view, building upon J2EE ideas in the main. But .NET cannot really compete with J2EE, because, of course, 'it runs on Windows only'
20 years ago Microsoft was making money selling bare operating system. But this is not the case anymore. Microsoft has a lot more products to sell. If they sell Word for Mac, why would not they do the same for Linux? And if they (and others writing for Windows platform) could port their product easier, would not they do it? Linux market barely existed couple of years ago. There was kernel, there were window managers and desktops, there were even applications, but most of this stuff was free. Now Linux grew into a huge market. Would not MS like a piece of this pie? And with Mono it will be easier.
Now, what is Windows? It is not just krnl386.exe or user.exe, it is the whole bunch of files, executables, registry, DLLs, windowing, messaging, DDE/OLE/COM, etc. It is monolythic thing, hey, Windows console applications did not even exist in old Windows versions, it was so tightly coupled with presentation layer. MS realised this problem. .NET was the first step to decouple apps from kernel and hardware. Now WinFX is the largest makeover in Windows history. New APIs will provide better decoupling of application modules from the system, at the same time making development easier.
What is Linux? Just start whining about this bug or that bug in GNOME or KDE or Konqueror, and Linux zealots would shout that it is not Linux problem, because Linux is just a kernel. Fine. So how you would like this scenario:
* MS developed hardware-independent and language-independent CLR.
* They changing Windows API to decouple new managed apps from obsolete Win32
* New API would be hardware-independent as well and in large scale OS-independent, it would run on top of .NET CLR
* All new Avalon applications would be able to run on top of any OS which provides new API
* They (or Ximian/Novell) will port this new API to other systems, like Linux
Voila! The result is not just MS apps would be running on Linux, but MS presentation API and other APIs would be running on Linux. GNOME and KDE would go fishing, and Linux would shrink back to just the kernel. Basically, the whole difference between Windows and Linux would be in the modules that translate .NET bytecode into machine codes, and low-level stuff like memory management, paging, multiprocessing, thread handling. The actual implementation of these features is not important for most developers.
As the result, Windows and Linux will look and behave alike, and it will not be GNOME or KDE or even Swing, it will be Avalon. The difference will be only in the kernel. Big deal, MS will make money on everything else, from tools to apps.
MS does everything right. I always revered its marketing genius. Sun could not make Java/Swing to become THE API for Windows GUI. But MS got the idea, now watch how Avalon would become the standard API for Linux.
Do you still wonder why MS opened .NET specs and even helped Ximian to port it to Linux?