Neowin.net

Well, I think they mean more like.. lets just announce what we have decided to actually do and not release every concept and idea like which was the case for Longhorn in 2003 and 2004.

Windows architectural changes to a minimum. makes me think that the general structure won't be much different. Thats a bold statement coming from the team that has it's head change every aspect of navigation within Office, and one of the reasons I thought she was included!

Well, I think this could be a good thing if they do it how I think. Rollout smaller changes over a shorter peirod of time insted of alot of big changes in one go. Which is what they did with Vista to a dagree.

That said, I think with the changes done to Vista already, from here on, making the OS even more modular will mean that future changes and updates can be done without effecting a large group of users/apps etc. You get the idea.

well it had to have been done someday... and vista was it

It performs better yes but read this before you decide one way or the other.

I think the kernel is quite up to date and doesn't need a complete rewrite.
Why do you think it does?

I don't think he meant the kernel (which is why he/she said Windows). He's probably talking about some of the other things of Windows such as the Registry, DLL, API's etc..

I know with the move from OS9 to OS10 that it also included a change of kernel, however the biggest changes were things like moving over to the new nextstep API's, new UI, practically new everything.

I think windows could do with a rewrite on a lot of different things. I think that removing all of the win32 stuff and replacing it with .net would help. I think slimming down the memory requirements and making the system more efficent is another great step (minwin), i sure for all legacy stuff a virtual system could be implemented which would be transparent to the end user, but would allow them to run there XP/Vista apps. (I think WOW is an excellent example of this).

I would make the OS 64bit only, yes it requires more memory than 32bit (which condradics what i said above) but 32bit is reaching the end of it's useful life. Microsoft sees this in the Server space, im suprised they didn't push this more aggressivly in the desktop space. The argument of driver problems would be settled if the manufactures didn't have a choice (similar to when MS introduced Windows 95).

Apple made a leap of faith wth the move to OSX but i think it has paid off, they now have a platform which is mostly free of legacy (only carbon remains, but that has been stopped, with only maintence being performed), Microsoft needs to be daring and release a version of windows as dramatic as windows 95 was to 3.11.

The apathy of Vista is really not because of the problems, but because people can't see any point of upgrading. What does Vista do that XP doesn't do now, is the common question. The bugs etc.. are not really a massive obstacle as people understand that new big releases (1.0) always come with there fair share of bugs, it's the price of being cutting edge.

It's a shame in some ways as Microsoft have some of the smartest people ive ever spoken with, seen and heard. Project singularity and the mountain of others on Channel9, 10 MSDN and Technet forums is incredible with no other company having such as rich ecosystem of experts and users.

At the end of the day i've always believed Microsoft only really shines when they are challenged, when they are allowed to drift or coast then they sort of give up, the server area is a brillant example of this Windows 2003 R2 was a really good product, i installed and ran it the day it was released and it is the most stable and fully featured product ive used in a long time. Windows 2008 looks to be even better.

Sorry for the long blabbering.

But how are you going to fulfill both these at once:
1. Replacing the low-level Windows API with code managed by a virtual machine.
2. Slimming down the memory requirements and making the system more efficient.

.NET was intended by Microsoft to be a RAD development framework on par with the old Visual Basic 6, but modernized for a true object-oriented model, and also somewhat bridging the gap between a RAD environment and C/C++. It was *not* intended as a kernel-close framework and I think would struggle tremendously to not cause system-wide performance problems. What about the unpredictable garbage collector, for one thing? Or the VM and object marshalling overhead for another?

.NET is a high level, abstract, platform for developing applications and e.g. web services, and what's most important here, and what every programmer should know is: every programming language has its niche. I don't think kernel-close development is it as for this one, as little as C++ is usually suitable for web development. Merely the overhead to constantly marshal the .NET objects in this post-Win32 API to something digestible by the kernel gives me the creeps as a programmer. Unless you're saying the kernel in all its real-time system nature should also be written in .NET? ;-)

Last edited by Jugalator on 15 Nov 2007 - 13:05

Because LTD is a software engineer, has access to Windows source code and is certainly qualified to make such a grandiose claim.

You can not remove Win32 stuff and have .NET work... .NET is a wrapper for win32 and user32 it does not replace them!

You are aware that this is the thing that people are ****ed off about in Vista... Right?

Microsoft did changes in Vista that were rewrites. And people are POed that now their stuff doesn't always work. People say that they want Windows completely redone. Yet, when Microsoft starts doing that... It is a "Big mistake".

The issue with Vista is what Microsoft has talked about. If they make a change that doesn't follow every other Windows release... People bitch.

Removing things like win32 would be stupid.

Pushing the newer frameworks is the right way to go. It's a shame we haven't seen more people using WPF....It kicks GDI's ass.

No... people are annoyed by the new features and partial rewrites because Microsoft are making old work with new (trying to support legacy code).

It is, and always has been, a fact that making old work with new ends in tears... Vista is a prime example of this. We are seeing a new OS hampered by old flaws from 98/2000 days. The registry, dll's etc are an antiquated way of working and provide more and more angles of attack on an aged OS. This shows to me that it is time for a full rewrite as LTD said. New source = unknown quantity for hackers, and means the new fatures work correctly with a new kernel and with SDKs and APIs designed from the outset to work with the new said features.

We haven't cause it's still too early to adopt it; but we'll see good things come to surface in 1-2 years; Look at silverlight, that's one good WPF example, and let me tell you, it does kick Flash's butt.

I bet Windows 7 will have a lot of WPF for the pleasing of our eyes and the ease of our day by day computer tasks.

I don't think a Windows re-write will ever be possible... then' it simply wont be Windows, but a new OS... Plus rewriting something that's been updated for decades now... un-hu... only in the mind of those who have no idea what re-writing an OS means;

that's how it goes; everybody asks for what they want and thats it... that's the easiest part of the cake.

Right on the money. MS needs to state that they are doing a rewrite and support Vista while people transition to Windows 7. Apple did it right when they went from OS 9 to OS X.

You will never see fully managed code. even MS doesn't want to write it. Managed code is slow compared to native calls. .NET is just a large wrapper for native calls (best example is winform controls, they are just standard win32 controls wrapped in an interop that converts them to managed .net)

because managed is JIT'ed it will always be a lot slower then native code... and you dont write OS's on JIT'ed code

Hopefully Microsoft's Singularity project will lead to advances in Windows that make this a negligable tradeoff.

You've got a lot of great things in .NET 3.0 that you simply can't do with unmanaged code: Every WxF.

I agree, I like the changes they've done. Many of them are needed. And at some point you'll have to break compatibility with really old stuff and get that legacy code out of the way (maybe into a VM/hypervisor).

But I think what they're talking about here is making smaller changes over a short time period, insted of alot of big ones in one big bang. Or something like that. Whatever the case the changes made with Vista now make it easier to get Windows into a even more modular design. And at that point any changes wouldn't have the same impact as they do now.

You're absolutely right.

But in the end, 99% of the Vista bashing goes on can be boiled down to how they changed some key thing from XP.

I happen to like the visual changes of Vista. However there is far more to vista than a new GUI.

Personally I think changes is what windows need. I'd like to see Microsoft scrap windows and come out with a completely new OS. You can only build on the old code for so long.

Really? seems to sell OSX pretty well... their biggest push is "Look how good we made it look!" it seems like anymore

Double really?

Anybody into the know, that I know, usually disables all the overblown eye candy junk. Half my reason for despising Vista!!

Definitely hope MS keeps all the crap to themselves this time too, so I don't have to bear reading/hearing all that overblown hype about it being so great!! Just about the other half of my reasoning for almost abandoning Windows in favor of Linux too! Wasn't for the fact my 2 scanners don't work in Linux (stupid Visioneer), Windows would be gone!

OpenGL was fixed to work on Vista a long time ago...

I agree, it's not bugs or changes to the os that make Vista less than desirable, it's the cluster of bloated applications and services that hog resources and give you no option of easily disabling or removing along with a complete lack of customizability.

If by never recover meaning "having one of the best quarters in years" then hey they'd love that

That's funny. People said that about previously failed Windows OS's (ex: ME) but they were able to recover nicely with XP. Vista might be a "mistake" in your eyes but you can't imply that every OS after it will be a failure.

I'd love to see this as well. It does make installing windows more difficult to the average user but then again pretty much every program has the "express" or "typical" or "full" setting for those who don't want to mess with it. The "Turn Windows features off" control panel is a step in the right direction except it only contains a handful of options.

Microsoft should really hire better usability testers and UI designers though. Or listen to their beta testers more. I mean just look at the effects timing in Vista for example. All those nice fading in windows etc are useless because they make the OS feel less responsive. With a better timing (or better yet, user customizable timing) we could have the eye candy without it slowing us down. Vista is a bit better in this respect compared to XP so one can only hope that the next version will finally get it right.