I was planning on writing a longer entry than this originally, based on the feedback I've received via e-mail and the discussions which the original article has spawned all over the internet; instead, I'll just focus on setting a few of the common misconceptions straight and clarifying a few points.
If you actually read the article all the way through (I notice that a large majority of people who complained about the article actually didn't), it doesn't say anywhere that Microsoft are abandoning backwards compatibility; instead, it says quite clearly that they're changing the way that compatibility is provided in Windows 7. That was the overall point of the article, not that Microsoft are breaking support for older applications (they're not, and most likely never will).
Also, the new legacy support won't be provided through the classic form of virtualization (ie, emulated hardware); instead, we'll be looking at very thin, dynamic, on-the-fly API translation, in much the same way WINE provides support for rudimentary Win32 applications on Unix platforms. Microsoft will most likely be providing the functionality through their Windows-on-Windows subsystem, but without the architecture conversion (from 64-32, 32-16, etc); making the process all the more streamlined and seamless for the end user.
I also found it quite funny that most complaints seemed to be in regards to the introduction of the article, rather than the actual body of it. It seems that picking Linux vs Mac OS vs Windows fights is still the most popular activity for users on the internet these days; I had hoped that had changed in the last 5 years, but apparently not. It was quite disappointing to see many people miss the point of the article entirely, purely because they took offence to the aside: "(despite other operating systems such as Mac OS X and Linux apparently being immune from such criticism)."
I've had some feedback from Microsoft after the article as well (in fact, the single domain with the most visits was tideXX.microsoft.com, with approximately 850 uniques in a 24 hour period), some positive and some negative. Some departments clearly know more than others in regards to the 7 architecture; although the most interesting claim I've heard so far is that the information I was provided with, which formed the premise of the article, could have actually referred not to Windows 7, but to an out-of-band project being simultaneously developed alongside Windows 7, possibly set for release some time between 7 and 8. This lends credence to the possibility that this may be intended not for the mass-market but instead for specialised applications such as embedded devices. This actually makes a certainly amount of sense, considering that if the legacy compatibility layers were indeed modular and separate from the main OS, removing them entirely from a distribution would cut down on installation size, and switching to the native APIs for development would yield significant performance gains over previous versions of Windows.
Tomorrow (or possibly later on tonight), I'll be posting a new exclusive, entitled "Looking to the Future: 8 and Beyond", which will include information about Microsoft's vision for the next decade or so in computing, and specifically the direction we'll see Windows take. |
