Windows 7 To Be Released Next Year?

KrispySausage writes "A recently-released roadmap for the next major Window release — Windows 7 — indicates that Microsoft is planning to release the new operating system in the second half of 2009, rather than the anticipated release date of some time in 2010. This quickly-approaching release date would seem to be at least partially verified by news of a milestone build available for review by an anonymous third party." We've previously discussed the upcoming new OS version, as well as its danger to Vista.

Development Structure

[ozmanjusri]

It will be interesting to see if the new focus on a "clean, lean" Windows 7 can be sustained, given Microsoft's deeply bureaucratic [blogspot.com] development structure.

Each team was separated by 6 layers of management from the leads, so let's add them in too, giving us 24 + (6 * 3) + 1 (the shared manager) 43 total people with a voice in this feature. Twenty-four of them were connected sorta closely to the code, and of those twenty four there were exactly zero with final say in how the feature worked.

The quote is from one of the people in the Vista shutdown menu team. It will be hard to winnow the cruft in that sort of environment.

A bit more of the review...

[Sirch]

... can be found here [istartedsomething.com].

Re:Marketing Slogan

[timster]

That's a terrible count.

Windows 1 - 3 (though the picture here was sort of confused in the first place, but never mind)

Windows 95 (4)
Windows 98 (4.1)
Windows ME (4.2)

The above three being sort of concurrent with:

Windows NT 3.5
Windows NT 4.0
Windows 2000 (NT 5)

Then the line was unified as:

Windows XP (5.1)

So Windows Vista is 6 and now we are talking about Windows 7. Got it now?

Re:Low memory requirements from ms...

[ledow]

Holy cow. I can remember when my HARD DISK was 480Mb. And that was 10 times bigger than the first hard disk I bought. And even THAT was an upgrade that cost nearly 25% of the computer again.

And, as you point out, that's BEFORE you do anything but actually turn the computer on and wait ten minutes. God knows what happens when you actually WANT to work. XP can boot fairly comfortable for low-to-mid-end users in 256Mb - it ain't fast, it'll swap, but on network managed machines without the usual startup cruft you'll get work done without in-app pauses and for a basic Office suite you won't even notice (I tend to find silent-hard-disk computers are percieved as "faster" by users, even when they are swapping more). 512Mb makes for a nice XP system and anything more is a bonus - I've run networks with hundreds of machines on XP and none of them ever needed more than 512Mb for adequate performance, unless they were doing high-end stuff like CAD - more important is to keep your startup entries clear than put more than 512Mb into an "office" XP machine. But having to have 512Mb before you can even boot the thing up?

total used free shared buffers cached
Mem: 254296 249912 4384 0 1288 75964
-/+ buffers/cache: 172660 81636
Swap: 473908 41000 432908

170Mb used out of 256. That's with a full KDE GUI (commonly referred to as bloated by a lot of people who obviously don't get out into retail stores and buy Windows much), an Opera process collecting mail from dozens of accounts and browsing hundreds of webpages each day with memory caching, and that's been running for about 26 days now.

total used free shared buffers cached
Mem: 222712 218960 3752 0 126832 40760
-/+ buffers/cache: 51368 171344
Swap: 1140604 0 1140604

And THAT's a proxy/filter/cache for a school, with transparent bridging that hasn't rebooted in months. 50Mb in use, admittedly no X-Windows running at the moment. Even most of that is Samba, Squid memory cache, Apache and other miscellaneous programs running on it, not all of which are critical to its operation but provide nice web or GUI interfaces to the admins.

Seriously, I know that things move on and you can't stay on a 386 but what benefit does the actual end-user get for all that bloat? What can you do on a 512Mb "Windows 7" machine that you can't on a 512Mb Vista machine, 512Mb XP machine, 512Mb Linux machine? Can you even BOOT with just 512Mb on this new version? More worrying, how many Gigs of rubbish that load on startup does it come with to fill up 480 Mb before you get into the machine? And what does that do to your minimum installation size and baseline CPU use?

I switched, personally, to Linux at home, Linux in work where appropriate (i.e. everything but network-managed desktops, because of the amount of legacy Windows software required) at around the same time that a Linux machine with 256Mb could do the same things as an XP machine with 512Mb, all other things being equal.

I've got a salesman coming tomorrow to try to sell the school Vista, two months after we put in a brand new XP network replacing the previous XP network. They aren't even going to be able to sell us that because I've done my research, which they don't expect smaller schools to do. Too high requirements, too many unnecessary features, too much rubbish, no practical advantage. How are MS going to sell an OS that's going to need literally Gigs of RAM once it's combined with Office and all the usual bundled offering?

This same salesman will be selling Windows 7 in a few years, of course he will, but what do you get for your money? I've seen people selling Windows Vista "digital signage" (i.e. scrol

Re:Vista's missing features

[Keeper Of Keys]

3 - Vista was a total bomb. There is no denying it at all. So why bother? Admit your mistake and move on quickly. All in all, this sounds like a surprisingly smart move on their part.

Has Microsoft ever admitted to making a mistake?

Pretty much [gizmodo.com]

Re:Such optimism?

[morgan_greywolf]

Too many people misread the whole Windows ME thing. Microsoft's goal since the days of Windows NT 3.1 was always to eventually migrate people from the old DOS/Windows codebase to the new NT codebase. In order to do that, they had to get the APIs synched.

Windows NT 3.1 had Win32 and Windows 3.1 had the older 'Win16' API. So they released Win32s for the older DOS/Windows platform, then Windows NT 4.0 with the new user interface. With the Chicago project -- Windows 95 (based on the new UI for NT4) -- was to be the first of the old codebase with the a full version of the new (NT) API, Win32. With that in hand, they had planned to do one more update to each version -- Nashville became Windows 98, and Daytona became Windows 2000. There was supposed to be a combined release of an operating system called 'Cairo' after that, where they finally dropped the whole DOS/Windows thing, but they got sidetracked because they were under pressure to produce a desktop OS for the low-end of the market. So the result was Windows ME, which was rushed out the door at the last minute and annointed as the last of the DOS/Windows line.

Cairo, which was promised to be totally 'object oriented' -- files would be stored as objects in a big database (sound familiar?), but it never happened. So instead, we get, as the first OS of the newly merged OS lines, Windows XP. And yes, XP looks like the greek letter "Chi" and "Ro", of course XP doesn't end up having anything promised in Cairo.

The Cairo promises were to be fullfilled with Vista, but that never happened because the schedule got pushed more and more and they were under pressure to do SOMETHING since competition from Apple and Linux stepped in to fill the void of 5+ years with no new Microsoft OS. So they pushed Vista out the door with none of the promised features and a bunch glitz stolen from Apple. (The last time they stole from Apple, it went exceedingly well, so what the heck, right?)

Windows 7 -- if it's true -- sounds like it could be what Vista was supposed to be. Of course, by now no one will care. It'll be too little, too late, IMHO.

Re:Such optimism?

[jam244]

Drivers - People say an OS is only as good as the software for it, and I'd argue an OS is only as good as the drivers. If you can't support your hardware, then software isn't even an issue. Now all drivers MUST be signed, yet many signed drivers don't work very well, if at all. I think it would be a good idea to have all drivers in one central repository (like the Linux kernel) so you won't have to worry about tracking down drivers for old hardware, but make sure the drivers work. And here is an idea, make the drivers modular. Drivers cause more BSODs and crashes than anything else. Don't let a single driver bring down a system. This is just basic common sense.

First of all, you do know that the vast majority of drivers for Windows are written by third parties, right? (nVidia, ATi, 3com, etc.)

Second, the reason they cause BSODs is because most driver code runs in Kernel mode. When the OS executes driver code in kernel mode, it is basically handing control over to the driver for a short period of time. If the driver makes a booboo, there's typically no recovery. In user mode, the OS would be able to catch the error and halt the process, but in kernel mode it is the CPU itself that often throws the error.

When you're in kernel mode, even small things can BSOD the box. For example, going into dispatch mode (disabling interrupts) and then waiting for a message from another process. This results in deadlock, and the CPU can detect this. The result is a BSOD (technically the CPU is calling the code to generate the screen). When you see the stop error IRQL_NOT_LESS_OR_EQUAL, that's usually what the problem is.

Theoretically, the OS could load the drivers in an emulated mode, to make the drivers think they are running in kernel mode when they are actually in user mode. That would slow current drivers to a crawl, though. Forget about games, gigabit ethernet, and Fibre channel.

A true solution would be to rearchitect the whole driver model to allow performant drivers operating entirely in user mode, and Microsoft actually pondered this, but it received so much pushback from third parties that it was scrapped.

In short, you have the right idea here, but it's just not that simple.

Re:Such optimism?

[Richard W.M. Jones]

And here is an idea, make the drivers modular. Drivers cause more BSODs and crashes than anything else. Don't let a single driver bring down a system. This is just basic common sense.

It is common sense, but PC hardware currently makes this hard to achieve. Give a device driver access to I/O memory and it can hang the PC in numerous ways - eg. writing to another device, accidentally performing DMA to a random bit of memory, putting the device into some state where it grabs the PCI bus and never releases it (effectively causing the PC to hang).

Intel's VT-d which effectively virtualizes the hardware, DMA and PCI bus may one day fix this, or at least allow you to write safe device drivers. But VT-d is experimental and hardware which supports it is very rare at the moment.

Conversely you can not give device drivers full access to the hardware. Exokernels work like this -- all hardware requests go through a small, trusted hypervisor. But performance tends to suck and some things just don't virtualize well, 3D graphics cards in particular.

Rich.

Re:This is fake, mod article down (*sigh*)

[DigitlDud]

Uhh, Windows 7 is just a codename, the version number always refers to the version of the NT kernel. Vista was 6.0, 2003 was 5.2, XP was 5.1, 2000 was 5.0. Apparently Windows "Seven" will be NT 6.2 signifying that it contains minor kernel changes.

Re:windows7

[Altus]

I don't know about flame bait but if you want to make a claim that apple breaks backwards compatibility you really need to back it up. Apple has done a lot to keep backwards compatibility despite several chip architecture changes over the years. Yes, they have deprecated APIs in OS X over the years but they generally give a lot of warning and the migration path is fairly trivial. Sure MS went out of their way to keep windows 16 apps running under win 32 but as I recall it didnt always work very well.

Vistas "XP Compatibility mode" doesn't work that well either. Ive been playing around with Knights of the old republic II. It came out a bit over 3 years ago and it runs reasonably well on my Mac book pro under parallels with windows XP. On my girlfriends vista machine (which is a lot beefier than my virtual windows box) it crashes after a few seconds even in compatibility mode (although the crashes are different).

The grandparent poster made a blanket statement without and evidence. Its a statement that cant be refuted because its has almost no content. The only thing you can do is write back and say "oh yea, says you." Its a statement that will be more likely to lead to an argument than a discussion. If the poster had backed up his statement I would agree with you but as it is, flame bait isn't all that far off. If I said the same exact thing about MS in the same way, I would expect to be modded similarly.

Re:Low memory requirements from ms...

[martinmcc]

I disagree - I had Vista on my Laptop for a while (modern spec, came with Vista pre installed)- I could have been waiting over 10 minutes for it to kick back up after going to sleep. It was well over a minute just to boot. Plenty of times it would spend time 'thinking' and ignoring me. YMMV, but I think the fact that a modern system & OS _CAN_ be slower than one 20 years ago (a long time in IT land) is a sad indication of how badly OS development can go.

With the 8086, I was booting in maybe 15 seconds (time has hazed my memory), loading turbo pascal, and programming away with only some minor waiting times. Sure there was a noticible delay when compiling, but even then it would be less than a program of comparible usefulness on visual studio.

As an aside, I think my most productive system (both in terms of how responsive it was, and how much I got done) was a 40Mhz 386 with, if memory serves, 2MB of RAM. I played with Linux, accelerrated my programming abilities with learning C, Assembly, the intricsies of gcc, gas, ld, make, awk etc. etc. I build my own OS on that system (sure it didn't do anything useful, but it had all the framework (multitasking, mem manger etc.) in place).

 

Re:windows7

[adamstew]

The primary difference between the OEM and the Retail license of windows is portability: The retail license you can move from machine to machine to machine to machine to your heart's content. With the OEM license, the first computer you install it on is the one it's stuck on. You can't move it to another machine.

Now, you _CAN_ move it to another machine, if you call up MS when activation fails and just say you upgraded some hardware, or the HD failed and needed to reinstall...but the extra price you pay for is portability...

Re:windows7

[Locutus]

Windows XP-SP2 was an out of the ordinary release for Microsoft. There were still massive security holes in the OS and the industry was really getting pissed with the holes taking out their networks. IMO, Windows XP-SP2 was, in Microsoft's domain, a new OS release since there so many major changes to the standard Windows XP OS. I don't think you could get Windows XP-SP2 if you just rolled up all the updates for Windows XP and installed them. It was a "new" release.

Windows XP-SP3 is going to be the same and Windows 7 is actually going to be Windows XP-SP4.

Windows 7/XP-SP4 will have the obvious GUI changes to make it look like a sister of Vista, but it will really be Vista's little brother(XP) in drag. IMO.

LoB

Full release

[200_success]

There are actually two versions of Mac OS X. The full release, called Mac OS X Server, is $499. Granted, the feature set of plain Mac OS X is already similar to that of Vista Ultimate.

Re:Such optimism?

[Rufty]

Video wasn't in ring 0 until NT4. Good for a workstation - lousy for a server.

Re:Defragmentation, Windows 2000

[Anonymous Coward]

Yea like XP has so many bells and whistles that it's a problem?

Why, yes, it does! I recently upgraded an old laptop for a friend; win 95, 96 Mbytes RAM, 266M PII and a 6 Gbyte disk. XP ran slowly on it and left not much disk space. Win2k, however, was very snappy and had about a 600 Mbyte footprint on the disk. They can have my copy of win2k corporate when they pry it from my dead, cold fingers.