There is a narrative that the downfall of Window started with the missteps that lead to Windows Vista being delivered late but the origins of the problem existed well before Windows Vista – it actually goes back to Windows NT 4 and the decision in Windows NT 4 to dump large amounts of code from user space into the kernel for the sake of speed (context switching on the x86 platform is expensive) resulting in a downward slide from Window NT 4 onwards of quick and dirty solutions rather than well thought out long term permanent solutions. As noted by Jim Allchin, during the heady days of 90s, features were thrown at Windows with no consideration as to their impact on the wider system as so far as security, long term manageability, good coding practices to ensure that a ‘spaghetti code’ mess wouldn’t result from the changes being made etc.
Then rolled along Windows 2000 which was originally geared as the grand unifier of the 9x and NT lines into a single line that would span both the consumer and enterprise space but as the delivery was being pushed back sacrifices had to be made – the priority of the enterprise feature over adding the shims required for backwards compatibility meant that it was pushed off to Windows XP which was released in October 2001 so the original Windows 2000 was behind schedule which then necessitated Windows XP which meant the successor to Windows XP (which should have been the successor to Windows 2000) is then delivered late and half baked resulting in the successor spending it’s time to clean up the loose ends of the previous release thus dropping the development of Windows further behind.
What was the net result of this continuous cycle of dropping balls? The net result was that with each release Windows fell behind, what should been released in the form of Windows 2000 should have been the great unifier, when Windows Vista was released it took another 3 years to deliver Windows 7 to clean up the mess – what was released as Windows Vista should have occurred 3 years after Windows 2000 was released so then, what was Windows Vista, was ultimately released would be Windows 10 meaning that by the time 2007 rolled around with the iPhone that Microsoft would have had Windows 10 Mobile as a competitor to iPhone straight off the bat and the UWP framework that would have acted as the grand unifier across its platforms. In total around 8 1/2 years was lost fixing up monumental cockups resulting in more damage done to the Windows platform than any DOJ anti-monopoly settlement could do in a prosecutors wildest dreams.
We’ve seen recently in the latest Windows 10 build that the phone related components have been removed resulting in a branch being maintained based on an old code base providing updates to existing Windows 10 Mobile customers but no new build based on the current line that 16241 is based off of. One has to ask whether Microsoft has more or less thrown in the towel and decided that they’re better off on playing to their strengths and provide middleware for Android and iOS along with pitching their services to businesses including their MDM cloud based capabilities rather than trying to spend hundreds of millions on a failed attempt to win what would be minuscule marketshare with minimal benefit but at a great cost in terms of time, resources and money. One aspect I like about Nadella is the fact that he isn’t sentimental – it’s all about the hard facts of life and whether it makes sense fighting battles that were lost long ago or whether the money and resources are better spent on seeking out new opportunities and ensuring that Microsoft is at the forefront, being the ones setting the agenda and shaping that future instead of following merrily along as a ‘me too’ player in the industry.