Windows 8/Server 2012 RTM coming

Folks,

In case you haven’t heard the news, RTM and GA dates have been announced for
Windows 8. According to the latest buzz, we’re looking at RTM for the first
week of August (yes, this August) and GA in September/October:

http://blogs.technet.com/b/windowsserver/archive/2012/07/09/windows-server-2012-final-release-timing.aspx

If you haven’t had a chance to try out the new build environment or run your
driver under the latest Verifier now is probably a good time to start.

-scott

–
Scott Noone
Consulting Associate and Chief System Problem Analyst
OSR Open Systems Resources, Inc.
http://www.osronline.com

Brace yourself…

On 09-Jul-2012 19:51, xxxxx@broadcom.com wrote:

Brace yourself…

The sooner is RTM, the sooner comes SP1…
– pa

> If you haven’t had a chance to try out the new build

environment or run your driver under the latest Verifier

… or checked build,

now is probably a good time to start.

“Pavel A” wrote in message news:xxxxx@ntdev…
> On 09-Jul-2012 19:51, xxxxx@broadcom.com wrote:
>> Brace yourself…
>
> The sooner is RTM, the sooner comes SP1…
> – pa

I don’t always say what I think, but what I would like to say is that I
think I do admire them for the great courage they are showing by introducing
certain changes, both at the UI and the kernel level.

//Daniel

> certain changes, both at the UI and the kernel level.

From what I know about WinRT, it is - for now - the world’s best OO programming framework. The reason is that such frameworks either suffer from

• issues of accessing the underlying non-OO stuff (like Java and .NET do)
  or suffer from
• issues like “you need to use 2 programming languages, not 1, to develop a component, and also the 1st of them must be augmented with special library of hackery which makes it the real OO language” (like COM does).

But what is really seriously new in Win8 at the kernel level?

–
Maxim S. Shatskih
Windows DDK MVP
xxxxx@storagecraft.com
http://www.storagecraft.com

“Maxim S. Shatskih” wrote in message
news:xxxxx@ntdev…
>But what is really seriously new in Win8 at the kernel level?

Dynamic clock tick.I find it hard to understand how this is not going to
break a lot of things.But they probably know what they are doing, time will
tell.

//Daniel

Daniel, we certainly do hope we know what we’re doing. And I have a lot of
faith in the people who worked on dynamic clock ticking. But the truth is
that, to be good a very low power machines, we have little choice. We have
to make the changes that result in much lower power consumption, even if
they cause a few changes in the operational environment. We do, of course,
strive to minimize those changes.

Jake Oshins
Windows Kernel Team

This message offers no warranties and confers no rights.

wrote in message news:xxxxx@ntdev…

“Maxim S. Shatskih” wrote in message
news:xxxxx@ntdev…
>But what is really seriously new in Win8 at the kernel level?

Dynamic clock tick.I find it hard to understand how this is not going to
break a lot of things.But they probably know what they are doing, time will
tell.

//Daniel

There’s a *lot* new in the world of power. Tons. And, as I’ve held-forth on many occasions, the whole issue of a function driver having a relationship to multiple buses.

Now, if you could JUST send an IOCTL from WinRT you could actually use it to do real work with a custom device. Yes, yes… Device Access API blah blah Device Metadata blah blah blah. All a lot of annoying nonsense, says I. I personally can’t call it the “best… framework” for anything if I can’t use it to send IOCTLs to arbitrary drivers.

Peter
OSR

> Now, if you could JUST send an IOCTL from WinRT

IIRC WinRT code is just C++ code (well, C++ with MS’s additions), so you can call any C function from it, am I wrong?

–
Maxim S. Shatskih
Windows DDK MVP
xxxxx@storagecraft.com
http://www.storagecraft.com

Well… any C function? Like, mix WinRT with the Win32 API so you can call DeviceIoControl? I don’t think that’s possible.

In any case, my point – in answer to yours – is that I’m not nearly as enamoured of WinRT as you seem to be. It’s a new, restrictive, environment that rather strikes me as a PITA. But, you know, we get to write Metro apps with it, so I’m sure it’s worth it.

Peter
OSR

>It’s a new, restrictive, environment that rather strikes me as a PITA.

On the other hand, for writing client apps it’s better to have easier to use less universal API, which would give less opportunities for programming errors, and would be free of legacy ugliness.

xxxxx@osr.com wrote:

Well… any C function? Like, mix WinRT with the Win32 API so you can call DeviceIoControl? I don’t think that’s possible.

In any case, my point – in answer to yours – is that I’m not nearly as enamoured of WinRT as you seem to be. It’s a new, restrictive, environment that rather strikes me as a PITA. But, you know, we get to write Metro apps with it, so I’m sure it’s worth it.

There’s a spirited discussion on the C++ MVP list right now about this
topic. One well-respected writer is claiming WinRT is to Win32 GUIs
what Win32 GUIs were to DOS apps. That is, those of us who were happily
writing DOS apps were dragged kicking and screaming into the Win32 world
(and many of us continue to cling to our command lines), but users
wanted Win32 GUI apps. He thinks the same thing will happen with
WinRT. We developers will be dragged kicking and screaming into that
world, but users will demand Metro apps and we’ll eventually all be
assimilated.

–
Tim Roberts, xxxxx@probo.com
Providenza & Boekelheide, Inc.

Sigh; I was *so* hoping that resistance was NOT going to turn out to be
futile this time around…

–mkj

On 7/13/2012 1:01 PM, Tim Roberts wrote:

xxxxx@osr.com wrote:
> Well… any C function? Like, mix WinRT with the Win32 API so you can call DeviceIoControl? I don’t think that’s possible.
>
> In any case, my point – in answer to yours – is that I’m not nearly as enamoured of WinRT as you seem to be. It’s a new, restrictive, environment that rather strikes me as a PITA. But, you know, we get to write Metro apps with it, so I’m sure it’s worth it.

There’s a spirited discussion on the C++ MVP list right now about this
topic. One well-respected writer is claiming WinRT is to Win32 GUIs
what Win32 GUIs were to DOS apps. That is, those of us who were happily
writing DOS apps were dragged kicking and screaming into the Win32 world
(and many of us continue to cling to our command lines), but users
wanted Win32 GUI apps. He thinks the same thing will happen with
WinRT. We developers will be dragged kicking and screaming into that
world, but users will demand Metro apps and we’ll eventually all be
assimilated.

–

//
// Michael K. Jones
// Stone Hill Consulting, LLC
// http://www.stonehill.com
//_______________________________________________

Maxim S. Shatskih wrote:

> Now, if you could JUST send an IOCTL from WinRT
IIRC WinRT code is just C++ code (well, C++ with MS’s additions), so you can call any C function from it, am I wrong?

Yes, you are. When you link with WinRT, you are prohibited from using
standard APIs, other than those that are re-exposed through WinRT. I
don’t know how this is enforced. Maybe the WinRT runtime monitors the
DLLs that are imported. A Metro app has to be “packaged” before it will
run; perhaps the packaging process does this check.

It is my understanding that WinRT is not just a layer on top of the
Win32 API. It is a replacement for the Win32 API. Has anyone on the
list played with a Metro app enough to provide tales from the trenches?

–
Tim Roberts, xxxxx@probo.com
Providenza & Boekelheide, Inc.

I resistance is not futile that likely means Apple won, and then we
are all out of work. Plus you will never ever get to build another
system from parts ordered on newegg.

If MSFT manages to get metro accepted then at least for a while longer
there will be pcs.

Mark Roddy

On Fri, Jul 13, 2012 at 1:03 PM, Michael Jones wrote:
> Sigh; I was so hoping that resistance was NOT going to turn out to be
> futile this time around…
>
> --mkj
>
> On 7/13/2012 1:01 PM, Tim Roberts wrote:
>>
>> xxxxx@osr.com wrote:
>>>
>>> Well… any C function? Like, mix WinRT with the Win32 API so you can
>>> call DeviceIoControl? I don’t think that’s possible.
>>>
>>> In any case, my point – in answer to yours – is that I’m not nearly as
>>> enamoured of WinRT as you seem to be. It’s a new, restrictive, environment
>>> that rather strikes me as a PITA. But, you know, we get to write Metro apps
>>> with it, so I’m sure it’s worth it.
>>
>>
>> There’s a spirited discussion on the C++ MVP list right now about this
>> topic. One well-respected writer is claiming WinRT is to Win32 GUIs
>> what Win32 GUIs were to DOS apps. That is, those of us who were happily
>> writing DOS apps were dragged kicking and screaming into the Win32 world
>> (and many of us continue to cling to our command lines), but users
>> wanted Win32 GUI apps. He thinks the same thing will happen with
>> WinRT. We developers will be dragged kicking and screaming into that
>> world, but users will demand Metro apps and we’ll eventually all be
>> assimilated.
>>
>
> –
>
> //
> // Michael K. Jones
> // Stone Hill Consulting, LLC
> // http://www.stonehill.com
> //
>
>
>
>
> —
> NTDEV is sponsored by OSR
>
> For our schedule of WDF, WDM, debugging and other seminars visit:
> http://www.osr.com/seminars
>
> To unsubscribe, visit the List Server section of OSR Online at
> http://www.osronline.com/page.cfm?name=ListServer

Of SOME type. With tiles and touch screens, perhaps.

But not with ARM processors, because you WON’T be able to build those from scratch. Or add custom USB peripherals to them. Raise your hand if you’re having fun building the ARM version of your driver. Oh, wait…

Peter
OSR

wrote in message news:xxxxx@ntdev…
…
> But not with ARM processors, because you WON’T be able to build those from
> scratch.

Hopefully not - thanks to recent progress made by Intel. Few weeks ago I’ve
read a review of some
leaked prototype Android tablet powered by a new Intel SoC.
They were going to make good laugh of it, but instead were very impressed.
So if these new chips could run Win8, MS could throw the ARM fork of Windows
to the same bin where the poor Courier rests…
However I suspect that the WinRT or “Metro subsystem” is, unfortunately, the
writing on the wall for Win32.

Regards,
- - pa

Pavel A. wrote:

Hopefully not - thanks to recent progress made by Intel. Few weeks ago I’ve
read a review of some leaked prototype Android tablet powered by a new Intel SoC.
They were going to make good laugh of it, but instead were very impressed.
So if these new chips could run Win8, MS could throw the ARM fork of Windows
to the same bin where the poor Courier rests…

The purpose of Windows 8 ARM is to win the phone market. In order for
your plan to work, you are assuming that Intel would be able to convince
the phone world to switch from (multisourced) ARM to its (single
sourced) SoC. That is probably an even bigger battle than switching
from Android to Windows.

However I suspect that the WinRT or “Metro subsystem” is, unfortunately, the
writing on the wall for Win32.

You have more confidence in the Microsoft marketing machine than I do.
It is possible users will clamor for Metro apps for their desktop, but I
find that unlikely. I think we’ll see a dichotomy (Win32 on the
desktop, Metro on phones and tablets) that survives for the rest of the
decade.

–
Tim Roberts, xxxxx@probo.com
Providenza & Boekelheide, Inc.

Metro/WinRT is ONLY for desktops/tablets/phones, not servers. Microsoft’s server market share is pretty healthy, so Win32 is in no danger of going away.

Actually, I JUST read on Forbes the quarterly year to year sales of PC’s (87.5 million for 2Q12), and they were down 0.1% worldwide. Considering that smartphones and tablets are not part of this number, it seems like a LOT of people still like desktop/laptop systems. This also doesn’t count servers.

My biggest concern is Windows 8 is being designed for some market OTHER than these 300 million systems/year. If you ask me, the metro interface is just plain silly on my 30" monitor. It’s not uncommon for me to have a dozen+ windows open, and I’m going to want to shift to a system that only shows a maximum of 2? I think not. The number of Metro apps I plan to buy for my desktop system is essentially zero, and I plan to continue to spend money on normal Windows desktop apps (like the Adobe CS Update, and Microsoft dev tools). I personally use the start menu a LOT, and having my whole 30" screen flip to a tile based menu just run a program is silly. I’m trying to let it grow on me, and will see what the RTM is like. This may be the first time in a LOT of years that I don’t immediately shift my primary desktop to the latest Windows version. I’m one of those people who ran Windows NT 4.0 as a primary workstation, shortly after it was released.

It does seem like Apple has proven there are PROFITS in little screens, so perhaps all the Metro hoopla has nothing to do with what kind of computers people want and everything to do with what corporations think will be profitable. For a bunch of years now, data center products (servers and related systems) have paid for my salary, so have not worried much about client system profitability. My understanding is profits on desktop/laptop PC’s are generally terrible (except perhaps for Apple).

A quick examination of the Android market shows what happens if you don’t have a standard core HARDWARE base for an OS to run on. So why doesn’t Android have a BIOS/UEFI layer, and pluggable drivers, so you just install the a new version of Android and add the drivers for your specific device. Perhaps it’s because Linux does not have a binary driver API? If you’ve ever shipped a production Linux product in binary form, you pretty much know you need a build for each version of each distribution you support. Imagine what OUR lives would be like if to ship a Windows driver you needed the HP build, the Dell build, the Lenovo build, the Acer build. I know some device OEM’s DO have this, because their customer (HP/Dell/Lenovo) wants some sort of feature customization.

I think the standard definition of the WinTel platform has been so transparent to many people, they have no clue what it would be like if things were different. My guess is nearly all systems running Windows 7 will run Windows 8, with essentially no vendor engineering required. My guess is essentially ALL Android based systems will require vendor engineering to run the Jelly Bean release, and a lot of those vendors have no interest in expending engineering effort on products that don’t contribute to revenue. Microsoft and hardware vendors I think have an opportunity to create a NEW standard hardware platform, where OS updates don’t require vendor engineering.

Jan

-----Original Message-----
From: xxxxx@lists.osr.com [mailto:xxxxx@lists.osr.com] On Behalf Of Tim Roberts
Sent: Friday, July 13, 2012 4:37 PM
To: Windows System Software Devs Interest List
Subject: Re: [ntdev] Windows 8/Server 2012 RTM coming

Pavel A. wrote:

Hopefully not - thanks to recent progress made by Intel. Few weeks ago
I’ve read a review of some leaked prototype Android tablet powered by a new Intel SoC.
They were going to make good laugh of it, but instead were very impressed.
So if these new chips could run Win8, MS could throw the ARM fork of
Windows to the same bin where the poor Courier rests…

The purpose of Windows 8 ARM is to win the phone market. In order for your plan to work, you are assuming that Intel would be able to convince the phone world to switch from (multisourced) ARM to its (single
sourced) SoC. That is probably an even bigger battle than switching from Android to Windows.

However I suspect that the WinRT or “Metro subsystem” is,
unfortunately, the writing on the wall for Win32.

You have more confidence in the Microsoft marketing machine than I do.
It is possible users will clamor for Metro apps for their desktop, but I find that unlikely. I think we’ll see a dichotomy (Win32 on the desktop, Metro on phones and tablets) that survives for the rest of the decade.

–
Tim Roberts, xxxxx@probo.com
Providenza & Boekelheide, Inc.

NTDEV is sponsored by OSR

For our schedule of WDF, WDM, debugging and other seminars visit:
http://www.osr.com/seminars

To unsubscribe, visit the List Server section of OSR Online at http://www.osronline.com/page.cfm?name=ListServer