Neowin.net

True. some comments sound weird...

Anyway... Do any of you if the "traditional" windows GUI will be changed or have a huge makeover with Win7 ? And, if so, in what time-frame/release we will be able to notice that ?

Thanks.

I hope they don't completely change the looks with Windows 7! I love the Vista look, it's just that Vista hasn't got XP performance to go with it, and I don't fancy using Vista Transformation Pack again.

... because it's M1, very early in the dev schedule. Critical thinking anyone?

^ yes, agreed with the above. Remember Longhorn looking almost exactly like XP in its M1 stage. This is still about alpha builds. But that's not to say I think a "lesser" release wouldn't be in place. That could be exactly what's needed now. Windows 98 was for example a very successful OS, not because it innovated a lot on top of what we got from 95, but because it corrected and polished that OS.

And if too much changes from Vista (even if performance improves) Windows 7 will get hammered, just as Vista has been. I remember when Nashville (what later became IE 4.0) leaked (back when Windows 95 was the OS du jour, and NT 4 hadn't even shipped yet); the howls were rather loud and vociferous. The howling only got louder when a more polished version of Nashville (I cryponymed it "Nashville II" leaked, showing that Microsoft was quite serious. On hte one hand, we want Microsoft to venture out into new frontiers; however, whenever they do, we hammer them by being too different.

I had a nice chat with a head developer from the local MS office, he is the chief of the local windows vista team and future windows systems, I asked if MS has any plans to ditch backward support built into the OS, he said that this is something the developers are anxious to do and constantly push that feedback up the ladder but at the moment it seems no one in MS is brave enough to take that decision.

I asked how "easy" would it be for them to implement visualization as a way for backward support, he said it would be way easier and satisfying work to do.

I guess only Bill Gates can decide that.

Bill Gates is Microsoft Chairman now so less involved in any decision making.

Someone might have to make that hard decision if the backwards compatibility start impacting usability though. You can only take it to a point until the cracks start to form. It happened on an architectural level with Windows Me. ;-) It can happen on a user interface level in other aspects.

they should just build a new Windows and make something like WINE-style libraries for it.
maybe it'll happen with Windows 8.

I suppose everyone missed the part about the hypervisor right? Signs that MS is pushing for virtualisation to replace built in legacy code is right there. Also the 64bit version of Windows runs 32bit code on emulation of sorts WoW64. Most of the pieces are there, still a few things to do, but MS can if they wanted to dump built in legacy support and offload it to an on-the-fly VM system where apps using hardware virtualisation on the CPU side as well, will start up in a VM without the end user even knowing about it.

Performance will be slower than native support but you gain stability and security from it, plus any native apps run better in the long run.

ahem
http://66.102.9.104/search?q=cache:Lzzdx4H...;cd=1&gl=uk

Just use Win95. What's your problem?

You underestimate what Microsoft *could* do with virtualisation.

If they wanted redo the codebase completely, they could do it with some careful planning and thorough work.
When a program starts, Windows checks what APIs it tries to call, then load the program in a virtualised environment with kernel suitable to that version of Windows.
So if a program calls APIs that existed in say Windows 95, the program would run off a kernel that fitted that version of Windows. Same for 98, 2000, ME, XP, Vista.

Since the MinWin kernel is said to be modular, it should be possible to create kernel images (similar to Linux kernels) that could be loaded in by a virtualisation software.
Yes, this would mean the OS would be inflated in size. But these kernel images could be optionally downloadable (with 2000, XP and Vista included by default).
Yes, this would mean the software would take longer to load. But this is a small price to pay for an exceptionally stable OS.

You seem to think Virtualisation means loading up a whole OS within the OS, but it doesn't need to be that way.

It's the 'careful planning and thorough work' which bothers me most.. Vista has had the most resources and the most 'thorough work' put in to the creation of any OS before it, it was so thoroughly and carefully planned to switch control of the OS from the user to the OS itself and to prevent certain user based operations that it was designed from the bottom up with the aid of the FBI (before everyone starts shouting.. please look it up)

You say that your ideas of how it would work would inflate the size of the OS, but everything, including all that's been said so far by Redmond, points to the OS being designed to be smaller and less 'bloated' so it's a strong probability that, although your ideas have some validity, it is very unlikely that the os is being designed that way.. from my point of view 'Modular' can mean a great many things.

You say 'could' and 'should' quite a lot, and in a perfect world I may agree but there is very little that is perfect or even widely accepted as adequate as far as MS OS design is concerned and as for 'exceptionally stable'.. MS haven't EVER made ANYTHING that could be considered exceptionally stable and I hardly think that they are capable of doing so now, especially by completely throwing all os building knowledge, gained in the past 18 years, out the window in order to attempt something entirely brand new (which will obviously require a whole new familiarisation and settling period for everyone involved, end user and programmer alike, plus a whole bunch of issues that 'could have been done better and will be next time.)

As for loading up a whole OS for use in virtualisation, I am of a mind that that is the absolutely only way to go, because if not then a newly programmed virtual space for compatibility will not be guaranteed to be compatible and we will again be drowned in more compatibility issues this time with virtually every programme ever made up to now, and this time we won't have a non virtualised environment to fall back on.

Thank you for taking the time to respond sensibly to my post.

Last edited by The Walker on 23 Apr 2008 - 14:32

From what I gather, though, the reason Vista took so long was that a majority of the code was scrapped mid-development.

Also, you can't deny that Vista is the most secure Windows OS to date.
You may disable UAC because you are computer savvy enough to take care of your computer yourself without the Linux-esque prompt asking you for your admin password (if you run as user, which I've been meaning to swap myself over to doing for a while now :p), but for the common user, this offers a layer of security that's unprecedented in the Windows franchise.
Whether or not users simply OK these things away like they were just another "Are you sure" dialogue, is irrelevant. It's still security; a Linux box would be no more secure if there were users that input their root passwords to every popup in existence.

I was a bit unclear; what I meant was that the folder the OS is in (as we wouldn't put kernel images in /Program Files/, would we?) would be inflated relevant to the number of kernel images installed (again, similar to Linu.

But yeah, I agree with you that they would most likely not code it in this way.

I do wonder, what other definitions of "modular" are there? I'll admit, I'm a newbie when it comes to modular programming, having yet to write my own (web) project that's fully modular.

Well yeah, I can't say 'will', can I
In all seriousness, I realise that something to this effect is an utopian wish, but if nobody suggests anything, nothing will happen

I wouldn't go so far as to call it "throwing all OS build knowledge out the window", rather "taking all OS building knowledge and correct mistakes from the past".
But yeah, I completely understand the issue of software compatibility hell, which is why I suggested the virtualisation utopia.

There will always be "could be done better next time", that's development. It's never finished, only authors die.

I wonder if I was clear enough: When I said "OS within an OS" I meant like a Linux Live CD, where you could have a separate desktop virtualised inside your OS software.
Imagine say your browser window, but instead of the normal title bar, it would be the virtualisation software's title bar with the program's own title bar functionality added to it.
So it would look like a program, work like a program, and crash like an old program, without generating any Windows errors (or in case of, say, a game, that would require working video drivers, not take the entire OS down with it should the graphics drivers BSoD on you).

No problem, and thank you for showing that you're not a run-of-the-mill doomsayer, I originally thought you were simply out to bash their efforts.

+1

OMG ... ** insert some random baseless knee-jerk statement based on what you think my motivations are and where my conclusions come from**

FYI, I haven't even bothered to look at the screenshot, as how something looks on screen gives very little actual evidence of what is under the hood... or is that how you draw conclusions, by looking at it's shinies?

You have just made the dumbest, unsubstantiated reply I have read in a very long time.

-1 for playing 'follow the leader'

Duhhrrr!