NYT article says we should throw away Windows
A couple months ago I posted about an article by some guy at Business Week, that made all sorts of rubbish claims about Windows and OS X.
Not to be outdone, Randall Stross at the NY times decided he could use some TechMeme love and wrote basically the same piece.
He says of Windows:
Painfully visible are the inherent design deficiencies of a foundation that was never intended to support such weight.
Yet he fails to mention what any of these deficiencies might be.
He then says the the best solution to any problems with Windows is to “start over.” You know, because that worked so well for Intel when they tried it.
Stross has a point when he says that the time between XP and Vista was too long. He probably even has a point when he says that Vista doesn’t look like a product that was in development for 6 years.
Guess what? It wasn’t. You see, back in 2001 the Windows division at Microsoft came up with the hair-brained idea to change pretty much everything, as Stross is suggesting now. Only he’s too late, and Microsoft has already learned that throwing out everything you know about Windows and rocketing into a brave new managed-code-centric world just doesn’t work all that well.
Stross also uses some funny math and says that Vista is the equivalent of Windows “version 12.” It’s as if he’s trying to say that somewhere under the pretty UI, the core of Windows hasn’t really changed since Windows 1.0.
Of course that couldn’t be further from the truth. Windows NT was a completely new OS. Windows 2000 was nearly a complete rewrite of that. Server 2003 and XP SP2 saw more major changes under the hood, as did Vista itself.
That is to say, this isn’t your older brother’s Windows (“grandfather” didn’t quite seem appropriate given the time scale).
Even then, I’m still not sure why anyone thinks this “start over” idea has any basis in reality. Do you really think it would only take a couple of years to write an entirely new OS with all the capabilities of Windows Vista?
Stross also repeats the dubious claim that Windows is too “monolithic.” With its NT microkernel, layered and massively componentized architecture, and hardware portability – he can’t be talking about the same Windows that is sold today.
Nobody’s OS is perfect and I’ll gladly accept that Windows has its flaws. But if you want to get on someone’s back about being monolithic and having a hairy, crufty architecture – perhaps you should direct your attention elsewhere. But at least Linux doesn’t have bugs or security holes, right?
Lastly, Stross and others seem to be under the mistaken impression that Microsoft is somehow unable to change the existing Windows codebase. These guys present two options:
1) Build stuff on top of the last version of Windows.
2) Start over.
Why pretend that these are the only two options? Especially when historically Microsoft has always chosen door number 3:
Take what you have and make it better.
Replace the parts that need replacing.
Don’t break something without a good reason.
Trackbacks & Pingbacks
- Apparently, the NY Times fact-checkers took the weekend off | Ed Bott’s Windows Expertise |
- microsoft windows 7 build
- A couple months ago I posted about an article by some guy at Business Week, that made all sorts of rubbish claims about Windows and OS X. Not to be outdone, Randall Stross at the NY times decided… | Riaz Kanani
Comments are closed.
Brandon,
I haven’t seen the codebase so I can’t say with certainty what the problems with Windows are internally. I can see the symptoms which lead me to the opinion that something needs to change (probably not starting over).
Go to the Fonts folder and try to install a font. A folder and file selection dialog comes up that is from Win 3.1. I’m sure there simply wasn’t time to change it but the question I would ask is why does that dialog still exist in the first place? When you have so many different ways of selecting a file, how can the dependencies be manageable? We all know that over time patches get added to various pieces of code. How can the 3.1 dialogs be maintained even and more importantly, why? I don’t know if the newer dialogs are based in any way on the 3.1 dialogs but if they are, the dependencies must be insanely difficult to understand and code for. With the next Windows version, Microsoft should say that if an API was used in 3.1 but superceded by something else in 95 (for example a newer dialog) that the 3.1 stuff will be removed, guaranteed. Over each new version they would increase the “minimum version” for compatability by one version. This would seem to simplify the code base, at least that’s my guess.
When you consider a cost of Windows Vista that is in the multiple billions of dollars, I wonder how much longer the current development methods will work for Windows.
Hi Brandon
See here
http://community.winsupersite.com/blogs/paul/archive/2008/06/29/randall-stross-jumps-the-shark.aspx
Eric, the font dialog has been done ad nauseum. The code works, so it hasn’t been discarded, but it’s quite complex, so it hasn’t been updated, either. See one take on it at http://blogs.msdn.com/michkap/archive/2006/07/02/654110.aspx.
“With the next Windows version, Microsoft should say that if an API was used in 3.1 but superceded by something else in 95 (for example a newer dialog) that the 3.1 stuff will be removed, guaranteed.”
And in so doing, break all Windows 3.1 apps. I still have a few at work. You’d be surprised how many installers still have a 16-bit bootstrapper – Microsoft had to specifically work around this for 64-bit versions, because the 64-bit processors don’t provide the mode that 16-bit Windows apps need to work.
Removing the APIs that have been superseded doesn’t gain much either – they generally just map onto a more feature-rich API that was added later. You also break binary compatibility for anyone who ported a working bit of code from Win16. Some pretty much obsolete APIs only exist in Win32 at all for that reason. To take one as an example, see the documentation for OpenFile at http://msdn.microsoft.com/en-us/library/aa365430.aspx. Microsoft deliberately made Win32 as close to Win16 as possible, to ease porting for existing 16-bit apps. The documentation still details what they did (at http://msdn.microsoft.com/en-us/library/aa383674(VS.85).aspx).
Again, these APIs cost no Windows 7 developer any sleep. They work, they’re unlikely to be broken, and they therefore don’t need fixing. Removing them is more effort than leaving them alone. A little effort is required to ensure they keep working, but probably no more so than ensuring that the APIs added in Windows NT 3.1, and still used all the time by ‘modern’ Windows applications, do.
Brandon seems right that a person like you can ask for a comparison with Stross, precisely to put an end to the negative voices against Microsoft,
I believe that the first reaction to understand that Vista is an OS Overtime must start from you, nailed these journalists botany with a face to face.
These people dovrebbere learn a lot before writing for a major newspaper.