April 10, 2008 (Computerworld) Calling the situation "untenable" and describing Windows as "collapsing," a pair of Gartner analysts yesterday said Microsoft Corp. must make radical changes to its operating system or risk becoming a has-been.
OK, that sounds bad. You have my attention...
Nope, the problem with 'Alpha Longhorn' (the version of Longhorn we were working on before the reset back to the Windows Server codebase) was that we were piling everybody's science project (Avalon, WinFS, etc, etc) into Windows with some sloppy project management.
Among Microsoft's problems, the pair said, is Windows' rapidly-expanding code base, which makes it virtually impossible to quickly craft a new version with meaningful changes. That was proved by Vista, they said, when Microsoft -- frustrated by lack of progress during the five-year development effort on the new operating -- hit the "reset" button and dropped back to the more stable code of Windows Server 2003 as the foundation of Vista.
If anything, the fact that we could reset back to a good code base and port over our good Alpha Longhorn features indicates that it actually is doable to 'quickly craft a new version [of Windows] with meaningful changes'.
"This is a large part of the reason [why] Windows Vista delivered primarily incremental improvements," they said. In turn, that became one of the reasons why businesses pushed back Vista deployment plans. "Most users do not understand the benefits of Windows Vista or do not see Vista as being better enough than Windows XP to make incurring the cost and pain of migration worthwhile."
Big corporations, in my experience, want a new version of Windows with 'incremental improvements' - they do not want to retrain thousands of workers every couple of years. There are a lot of companies out there that use Group Policy to make XP (or Vista) look like Windows 2000 (Classic mode, classic start menu, etc) - They want the newer OS, for better perf, hardware support and security, but they don't want their users to notice the change.
It takes too long for Microsoft to build the next version, the company is being beaten by others in the innovation arena, and in the future -- perhaps as soon as the next three years -- it's going to have trouble competing with Web applications and small, specialized devices.Don't think so. Your computer will be your primary residence for your data for the foreseeable future.
Every user has a primary computer where all their data is stored - and your iPhone/smartphone still has to be tethered that computer to get your contacts/music/photos/email accounts. That is not going to change anytime soon.
It will (someday your data will primarily live in the 'cloud', not on a fragile laptop), but not that soon - not for the average American.
"Apple introduced its iPhone running OS X, but Microsoft requires a different product on handhelds because Windows Vista is too large, which makes application development, support and the user experience all more difficult," according to Silver and MacDonald.
iPhone and your Mac laptop do not run the same OS.
Your Windows Mobile phone and your Windows laptop do not run the same OS.
In both cases, a lot of the APIs and libraries look the same. And there is a bunch of code reuse.
It has been this way ever since Longhorn was nothing more than a codename.
This is like saying a company failed because their truck engine could not be scaled down for use in their motorcycles. No duh.
Their advice to Microsoft took several forms, but one road they urged the software giant to take was virtualization. "We envision a very modular and virtualized world," said the researchers, who spelled out a future where virtualization -- specifically a hypervisor -- is standard on client as well as server versions of Windows.
"An OS, in this case Windows, will ride atop the hypervisor, but it will be much thinner, smaller and modular than it is today. Even the Win32 API set should be a module that can be deployed to maintain support for traditional Windows applications on some devices, but other[s] may not have that module installed."
The devil is in the details, isn't it? What makes you think this approach is going the be 'thinner, smaller and [more] modular'?
Backward compatibility with older applications should also be supported via virtualization. "Backward compatibility is a losing proposition for Microsoft; while it keeps people locked into Windows, it also often keeps them from upgrading," said the analysts. "[But] using built-in virtualization, compatibility modules could be layered atop Win32, or not, as needed."Honestly - appcompat takes up a small amount of my time. It isn't very hard to support older APIs/applications. In the spots in Vista where we significantly broke appcompat we could see the problems coming a mile away (The Session 0 security fix, for example.)
Silver and MacDonald also called on Microsoft to make it easier to move to newer versions of Windows, re-think how it licenses Windows and come up with a truly modular operating system that can grow or shrink as needed.?
The SKU story in Windows, in my opinion, is stupid - all you need is a Home SKU, a Business SKU. Done - end of story. (Marketing, are you reading this?)
Wrt grow or shrink: Huh? You can already install/uninstall just about ever usermode part of Windows Vista.
Take Windows Movie Maker: If you don't use it, it doesn't run. Doesn't slow down Vista startup. Doesn't chew up memory when it is not running. If you don't like fact that is taking up (i dunno) 10mb of disk space on your 100GB disk, you can uninstall it. What is the problem here?