Microsoft unveils the final/product name for "Windows 7"

I can't speak for richard, but the driver model is the same as Vista, the kernel is generally the same as Vista but with some tweaks of course, the network stack is the same as Vista, the audio stack is the same as Vista, WDDM is the same as Vista... The GUI is really the only piece I see changing, and it's less GUI and more Explorer changes if you ask me.

Considering that this is pretty much the same concept of the changes between Win2K and WinXP (which was a minor release, in terms of versioning stamp) then I think Win7 deserves similar treatment.
 
If getting rid of dispatcher lock and adding full VHD support are tweaks in your book then I think I'd be blown away with the kernel changes that are actually significant for you. If you disregard changes just because they are not as visible as GUI improvements than we probably just have to agree to disagree on the subject of changes in Windows 7.
 
In the context that I said it -- the kernel revisions between Windows 2000 and Windows XP and Server 2003 -- this update still registers as a "tweak" to me. Yes, multiple lines of code and I'm sure several dependant modules had to be updated. Yes, I get that, but that doesn't make for a "major" OS overhaul to me.

And I was one of the primary proponents of Vista's "underhood" changes, so don't give me the whole line about "All you care 'bout is GUI" grief. You can do a search for my user name on this forum for numerous examples of how I feel OS improvements under the hood have more merit than the GUI, all other things being considered.

I'm happy with Vista - with or without the GUI changes - and I like what I'm seeing in Windows 7. But that doesn't make it a "new" OS to me...
 
I'm sorry but I will give you "All you care 'bout is GUI" grief as long as you evade to answer simple question: what is a major overhaul? Does it really need breaking changes in the system?

And it doesn't really matter what you think of Vista. My point is not about Vista but about Windows 7 - there are too many significant changes and improvements to call this release minor: from kernel changes to performance improvements to size reduction to streamlined UI to bloatware removal to IE 8 (huge changes there) to better OOTB dev experience with new user controls. The list goes on and we don't even know when is it going to be on the shelves so it's bold to assume we know everything about this release.

So I ask again: what IS a "new" OS for you, Richard, anyone. With all due respect it seems to me that if software is not overhyped, it's minor.
 
And it doesn't really matter what you think of Vista. My point is not about Vista but about Windows 7

Yes but, as I mentioned, I believe how people will view Win7 depends on what version they are using now. Coming from Vista the architectural changes are minor. It's all relative. The fact they aren't breaking compatibility (AND the short turn-around period) tells you right away that Win7 does not bring major architectural changes. If you use XP then all your apps are going to break (of course not, adapting to Vista has fixed most app-compat problems that had a feasible solution).

The list goes on and we don't even know when is it going to be on the shelves so it's bold to assume we know everything about this release.

Correct. That's why I was careful in pointing out that my opinion was based on all that I have read/seen not a final judgement on a product that will ship 1 year from now. This is especially important since it was at PDC 2003 that MS demonstrated longhorn with all new effects and features, most of which were either cut or significantly pared down in the final release.

So I ask again: what IS a "new" OS for you, Richard, anyone. With all due respect it seems to me that if software is not overhyped, it's minor.

I have a very subjective but powerful system of determining whether a Windows release is major/revolutionary/whatever: anytime I go to a store and buy it retail. That happened only twice thus far: Windows 3.1 and Windows 2000. All other versions I got either through OEM or MSDN. This system is not perfect however, since my copy of Win 95 was OEM and that was a major release. Oh well. :p

For me Win2k was a major release but I hadn't used NT4 extensively, instead coming from Win98SE. Actually, I had recently (less than a year) purchased a new PC that came with Win98SE OEM but I still went out and paid the equivalent of 300 euros for Win2000 Workstation. I also still have the Win3.1 floppy diskettes (even though I don't have any floppy drive to put them into :) that thing was really impressive back then.

Contrasting with that, I only moved my home desktop to XP on August 2004, shortly after MS released SP2 and ONLY because Win2000 didn't support logical processors and I had just bought a P4 with Hyper-Threading. So even though the PC came with XP I did a format and installed W2k and then a week later after I found out the reason for the missing logical processors, I had to give in and go with XP. :cry:

Now Windows 7 is interesting. By MS's own admission they're focusing on end-user features but that's a problem because those features are only good if you use them. What do I care that Wordpad and Paint now use the ribbon (that I love in Office) if I have Word and Paint Shop Pro? What do I care that Win7 will come with IE8 if I use Opera? Things I'm liking (a lot - but I'm skeptical they'll carry through to the final release) so far are:

Reduced disk foot print. It's funny seeing people defend MS saying "meh, HDs are so cheap nowadays" only to have MS finally agree with us that said 13GB for an OS install is madness.
Reduced memory foot print, only starting services when they're needed. Ditto.
More focus on the virtual stacks. Finally, after pushing them in Longhorn and paring them back in Vista, I'm glad they're back on top (a-ha!). Hopefully they'll stay there.
Removing filler apps and putting them on Windows Live. Never used Windows Movie Maker, Vista's mail app, whatever.
Thread/Scheduler improvements. Finally. Perhaps MS now realises that without games there's no need to buy new PCs, and if there's no new PCs there can't be more features in Windows/Office and if that can't happen people won't buy new versions of MS software either.
 
I am glad MS is actually pulling their head out of their arse and taking a look at what made Vista so bloated in terms of disk space and memory usage. Its quite refreshing and encouraging to see the Windows 7 keynote being done on a netbook! That means they are truly focused on making it a performant OS. I still think and am sticking by my claim of Vista being slow for no reason.
 
Hmm, that sounds like a bunch of quite nice changes there both under & on the hood :smile:
 
Let me add another feature I'm pumped about: fonts.

As an amateur typesetter and DTP designer I've always had a problem with how Windows manages fonts. Despite being a multi-user OS since NT, fonts were still a system-wide resource. I always have a lot of fonts installed but I don't always need them because it's a hobby, not my job. So the OS has to keep them into memory, searching for a font in Word for your average report requires me to sift through all my "designer" fonts, etc. Why not make them a user-wide resource?

Well, if you follow the link it seems W7 is not going quite as far (why?) but it's getting there. By being able to keep most fonts unloaded and even hide some (say for regional considerations) MS is finally addressing this (admitedly obscure) issue.
 
The fact they aren't breaking compatibility (AND the short turn-around period) tells you right away that Win7 does not bring major architectural changes.

Changing the implementation doesn't have to alter the semantics of the interface.

If someone at MS figures out how to rewrite the memory manager so that it's way faster, but they're really careful to ensure this change doesn't break any existing drivers or APIs, is that a major architectural change or not?

If someone at MS figures out how to rewrite the thread scheduler, so that it can now efficiently scale to 256 processor cores, and they're really careful to ensure this change doesn't break any existing drivers or APIs, is that a major architectural change or not?

Reduced disk foot print. It's funny seeing people defend MS saying "meh, HDs are so cheap nowadays" only to have MS finally agree with us that said 13GB for an OS install is madness.

The reason for Win7 to go on the diskspace diet is the sudden popularity of the netbooks, many of which ship with tiny solid state drives < 16 GB.

Since MS would like to discontinue XP at some point, they need an OS that can run on a netbook reasonably well.

IMHO, for normal systems, Vista's disk foot print doesn't really matter much.

These days you can buy 1 TB hard disks for ~ $120, what does ten or fifteen GB for an OS install matter?
 
And it doesn't really matter what you think of Vista. My point is not about Vista but about Windows 7 - there are too many significant changes and improvements to call this release minor: from kernel changes to performance improvements to size reduction to streamlined UI to bloatware removal to IE 8 (huge changes there) to better OOTB dev experience with new user controls. The list goes on and we don't even know when is it going to be on the shelves so it's bold to assume we know everything about this release.

For the average Joe-user the kernel tweaks are minor. You don't see dispatcher problems on normal quad-core systems. The changes are primarily aimed at servers: Supporting up to 256 cpus and improving dispatching to actually scale peroformance all that way.

The other improvement is in how processes are scheduled in order to power down as many cores as possible to increase battery life for laptops. Those are minor tweaks compared to how the transition from XP to Vista (NT 5.1 -> NT 6.0) completely revamped the kernel and driver model.

The changes are mostly in the UI, which is just an app. (explorer) that sits on top of the OS. The changes in the UI is also what is visible to the end-user and hence the new moniker Windows 7

Cheers
 
For the average Joe-user the kernel tweaks are minor. You don't see dispatcher problems on normal quad-core systems. The changes are primarily aimed at servers: Supporting up to 256 cpus and improving dispatching to actually scale peroformance all that way.

For now these things are targetted towards the server.

But everything points to Intel going for multi-core in a big way in the future. I would speculate that enhancing the kernel means Windows is in a good position to take advantage of future hardware developments.

Besides, even if Joe-user doesn't see a big difference right now, if large chunks of various subsystems had to be rewritten and restructured to enable the things described, isn't that a "major change"?

In fact, isn't that precisely the kind of "major change" you want in the kernel? Something you barely notice, until you actually need it, and then it just works?

Unlike "major change" that goes around breaking everything and generally making your life miserable until drivers and hardware vendors catch up. (Clearly sometimes you can't avoid doing this, but it's not something you want to do often.)

Those are minor tweaks compared to how the transition from XP to Vista (NT 5.1 -> NT 6.0) completely revamped the kernel and driver model.

There are lots of ways to measure how big a change is.

If change is measured by "what code was altered and how fundemental were those changes", then many of the changes in Win7 kernel are pretty big. As in the stated example, the dispatcher lock is a fundemental piece of NT's kernel architecture, and has been there since the very beginning 15+ years ago.

Removing it was an impressive piece of engineering IMHO, especially for what people are calling a "minor release" that uses the "same" kernel as Vista.

This is the type of change that if done incorrectly, can have wide ranging effects and introduce all sorts of bugs that people will be tracking down for ages, and is generally something developers avoid in a "minor" release. The fact that MS did it, and didn't introduce any signficant problems in the pre-beta they released at PDC is quite impressive.

On the other hand, if you want to measure the change by "overall impact to Joe User", I agree that for now, this kernel change might not be noticed. It does however put Windows in a good place to take advantage of future hardware. (And wasn't "getting ready for the future" one of the reasons for Vista's "major changes"?)

Finally, if you want to measure the change by "what got broken", I'm pleased that Microsoft is saying "very little".

The point is, you don't have to break the driver model, regress appcompat, and cause industry upheaval for there to be some good solid significant new stuff in the Win7 kernel.
 
Last edited by a moderator:
Changing the implementation doesn't have to alter the semantics of the interface.

If someone at MS figures out how to rewrite the memory manager so that it's way faster, but they're really careful to ensure this change doesn't break any existing drivers or APIs, is that a major architectural change or not?

Take a look at what I underlined, if they're being careful to maintain compat they're going to take longer to do it. Rewriting Windows memory manager is not something you try every day (or year), that's why I mentioned the short turn-around time. Also, simply making it faster does not make it a major architectural change, considering we're talking about nanoseconds. Rewrite the boot process to go from 45 seconds to 10 or less, now that would be something.

If someone at MS figures out how to rewrite the thread scheduler, so that it can now efficiently scale to 256 processor cores, and they're really careful to ensure this change doesn't break any existing drivers or APIs, is that a major architectural change or not?

That's another issue. How many people have 256 cores? Vista supports them but if you don't have the hardware how do you know if it's efficient or not? If you don't have the hardware, why does it matter to you that Win7 will improve efficiency if you're only going to benefit from this long after Win8 has already shipped? Vista also supports 16TB of address space and 128GB of physical ram and Win7 will probably increase the physical ram limit but, unless you're running a server then what does it matter to you or me? And if you're running a server then you're probably going to be running Windows Server 2008 R2 anyway not the client version of the OS. So, these are good features to have, but are completely irrelevant on my decision to upgrade or not.

The reason for Win7 to go on the diskspace diet is the sudden popularity of the netbooks, many of which ship with tiny solid state drives < 16 GB.

Since MS would like to discontinue XP at some point, they need an OS that can run on a netbook reasonably well.

IMHO, for normal systems, Vista's disk foot print doesn't really matter much.

These days you can buy 1 TB hard disks for ~ $120, what does ten or fifteen GB for an OS install matter?

More disk space for OS means more IO access for the OS (bigger files, more to read), it means your games/documents are put further inwards into the drive where transfer speeds start to drop. And while capacity has increased steadily, speed has not (only as a factor of angular velocity of higher density platters).

But let's forget that for argument's sake. You say 10 or 15 GB doesn't matter. Why those values? Why not 30 or 50 GB? When does it start to matter? Who draws the line between what's reasonable and what's not? No one. The argument is not how much space the OS can use. The argument is, how much space the OS needs for the features it provides. If you write your code to be efficient in the first place, then the disk/memory/cpu it requires is reasonable given the feature it allows (assuming those features are useful of course).
 
It saddens me that people on this board judge technology by version number and not by technical improvements (e.g. those pointed out by aaaaa00). :rolleyes: Ignorance is ok on the Internet. But on the technical board like this one it's just a mindblowing disappointment to see statements like this one.

So, do you think that XP was major version after 2K?
What about SP1 for XP? What about SP2 ?
Since there was consensus that XP was minor revision, Win7 should be (and that is refelected in the version number!) minor too. Why? - Some people already answered you.
My comment was meant to mean "its (relatively) minor tweak version" the way XP way tweak over 2K.
And tech staff at MS knows that, but for PR reasons they had chosen to go with 7.
Which is fine name :)

Don't get blinded by the name, look at details.
I think noone outside MS can say if the changes to the kernel are minor (ie without rewriting big parts of it) or major.
The internal version number is in favour of "minor" :p
 
I hope Windows 7 plays with SSDs better than Vista. I was reading something about Vista doing too much I/O and that resulted in poor performance with SSDs or something like that. I am not caught up with SSDs but would that not mean fragmentation would be a thing of the past? Or would fragmentation of files still matter?
 
I hope Windows 7 plays with SSDs better than Vista. I was reading something about Vista doing too much I/O and that resulted in poor performance with SSDs or something like that.

You mean this?

I am not caught up with SSDs but would that not mean fragmentation would be a thing of the past? Or would fragmentation of files still matter?

Fragmentation will still happen since the file system is the same. However, because SSDs have a very small access/seek time you can make the case that having the drive access two halves of the same file on "opposite" sides of the drive has a neglegible speed hit, unlike HDDs.
 
Fragmentation will still happen since the file system is the same. However, because SSDs have a very small access/seek time you can make the case that having the drive access two halves of the same file on "opposite" sides of the drive has a neglegible speed hit, unlike HDDs.

Thanks Richard. I figured as much but was not totally sure.
 
@Richard: I read your post and what I see still deserves grief. ;) You see, the question wasn't "where you want changes" but "what do you think is the major OS overhaul". I understand and sympathize with your usage scenarios. But they are as irrelevant as mine since this is thread about Win7 as it is, not about what software you need.

I'd love to see lots of features removed from Windows, I want classic skin and nothing more. But that's not what makes major system change. It's my preference not Windows Seven's technical prowess.

The other improvement is in how processes are scheduled in order to power down as many cores as possible to increase battery life for laptops. Those are minor tweaks compared to how the transition from XP to Vista (NT 5.1 -> NT 6.0) completely revamped the kernel and driver model.
Vista's kernel is "just" an improved version of XP's kernel, Win7's kernel is "just" improved Vista kernel. The only difference between those two transitions is that you can't instinctively tell the difference between Vista and Windows 7 because there's no obvious indicator like breaking change of some sort. But once again if e.g. VHD support is a small change than I'd love to hear what's a huge change.

The changes are mostly in the UI, which is just an app. (explorer) that sits on top of the OS.
It's just an app which absent would make it impossible to use any other app. This is flawed reasoning. How about I use it to justify that kernel is just an app that sits on top of bootloader? But that's still fallacy - you talk about product stripping off its parts that you think are irrelevant to prove your point. Windows 7 is Windows 7, not Windows 7 kernel and Windows 7 UI.

The changes in the UI is also what is visible to the end-user and hence the new moniker Windows 7
Then why bother with changing anything else if this is what matters?
 
@Richard: I read your post and what I see still deserves grief. ;) You see, the question wasn't "where you want changes" but "what do you think is the major OS overhaul". I understand and sympathize with your usage scenarios. But they are as irrelevant as mine since this is thread about Win7 as it is, not about what software you need.

Heh, okay, my whole point is that whether an OS update is major is completely subjective. It doesn't matter if we all agree that in theory Win7 is a huge upgrade if none of us then considers it one in practice. Or put another way, I belive the whole discussion of minor/major to be borderline useless. In my book, all that matters is if we like/need the new features. As an example, one of the Vista features I enjoy the most is pretty insignificant in terms of engineering: copy/move dialogue boxes have the est time remaining in the title instead of the % complete.

Vista's kernel is "just" an improved version of XP's kernel, Win7's kernel is "just" improved Vista kernel. The only difference between those two transitions is that you can't instinctively tell the difference between Vista and Windows 7 because there's no obvious indicator like breaking change of some sort. But once again if e.g. VHD support is a small change than I'd love to hear what's a huge change.

That's a great example right there. Right now I have no interest in having native VHD support but I do use Virtual PC. So while that feature doesn't sway me to upgrade that could change.

It's just an app which absent would make it impossible to use any other app. This is flawed reasoning. How about I use it to justify that kernel is just an app that sits on top of bootloader?

Well, if I can interject between you and Gubbi, one is user-replaceable the other isn't.
 
Oh, I agree that minor/major discussion is useless. But that's what people focus on for whatever reason. Instead of system's merits people focus on numbers and labels. As I stated before I find it dishearting to see this kind of talk on technical board. From "6.1.XXX nuf said" to "I don't care about feature X therefore it doesn't matter" to "no one would use it anyway" - this is as far from technical as one can get (well, except for "M$ sux" and "Kill Bill" perhaps).

I mean seriously, do you guys see "scales to 256 cores" and the only thought that crosses your minds is "meh, I don't have 256 cores anyway"? How about: hey, so if it scales like this I bet it's not just a constant value changed in the code; perhaps that's because removing kernel locks boosted multiplier when doubling cores? How does that matter to me? Obviously Intel is not releasing i7 this year...

BTW: what do you mean by "user-replaceable"?
 
Back
Top