From Newsgroup: comp.misc

From the «nope» department:
Feed: The Verge
Title: Microsoft tries to spark new life into Windows
Author: Tom Warren
Date: Fri, 09 Oct 2026 17:00:00 +0000
Link: https://www.theverge.com/tech/1008801/microsoft-windows-surface-event-hybrid-ai-notepad

[image 1 (link #1)]

When Microsoft released Windows 11 five years ago, it felt like an
operating system that was still being renovated, a work in progress[2].
You'd think by now that those renovations would be complete, with
Microsoft turning its attention to Windows 12. Instead, Windows 11 feels
like it's here to stay, forming the foundation for Microsoft's next big ambition: transforming Windows into an agentic OS.

At a Windows and Surface event this week, Microsoft laid out its plan to
bring AI agents to Windows, including hybrid intelligence that lets
Windows take advantage of free local models instead of expensive cloud
ones. "I think today is the day we'll …

Read the full story at The Verge.[3]

Links:
[1]: https://platform.theverge.com/wp-content/uploads/sites/2/2026/10/gettyimages-2299110674.jpg?quality=90&strip=all&crop=0,0,100,100 (image)
[2]: https://www.theverge.com/22708762/microsoft-windows-11-review (link)
[3]: https://www.theverge.com/tech/1008801/microsoft-windows-surface-event-hybrid-ai-notepad (link)

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.misc

On 2026-10-10, Retrograde wrote:

From the «nope» department:
Feed: The Verge
Title: Microsoft tries to spark new life into Windows

Open source Windows 4.* and Windows NT 5.*.

Here, solved, new life sparkled. I'll send Microsoft the consulting
bill.

(It's for 0 USD, the only compensation defined is: "bring back Links".)
--
Nuno Silva
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.misc

Nuno Silva <[email protected]d> wrote:

From the «nope» department:
Feed: The Verge
Title: Microsoft tries to spark new life into Windows

Open source Windows 4.* and Windows NT 5.*.

Except that Windows 4.0 was unreliable crap. Windows 7 was the first
version that met what I would consider 1970s production quality OS design.
NT 5.* I assume means Windows Server 2000 which might have been a good
thing, I don't know.

7 was maybe the last actually usable Windows OS before they
started making everything network-dependent.

ReactOS has a future... there isn't much there even after so many years
of development, but the worse Windows gets the more argument there is to
put support into ReactOS development.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.misc

Scott Dorsey <[email protected]> wrote:

Nuno Silva <[email protected]d> wrote:

From the ??nope?? department:
Feed: The Verge
Title: Microsoft tries to spark new life into Windows

Open source Windows 4.* and Windows NT 5.*.

Except that Windows 4.0 was unreliable crap. Windows 7 was the first version that met what I would consider 1970s production quality OS
design. NT 5.* I assume means Windows Server 2000 which might have
been a good thing, I don't know.

Supposedly NT 3.51 was a very bulletproof version from way back in the
day. Of course 3.51 was back when it was "server only" and basically a reimplementation of VMS by Dave Cutter for MS. Then BillG interveaned,
it (the core) became the foundation for "all windows" going forward,
and all kinds of problems resulted from meshing the bullet proof VMS underpinnings with the wild-west free-for-all that was user facing
MSWin at the time. And it then took them till 7 before they got much
of the mess cleaned up.

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.misc

On 2026-10-10, Scott Dorsey wrote:

Nuno Silva <[email protected]d> wrote:

From the «nope» department:
Feed: The Verge
Title: Microsoft tries to spark new life into Windows

Open source Windows 4.* and Windows NT 5.*.

Except that Windows 4.0 was unreliable crap.

*That* unreliable?

Well, at least Windows 4.10 tends to be quite well regarded.

Windows 7 was the first
version that met what I would consider 1970s production quality OS design.
NT 5.* I assume means Windows Server 2000 which might have been a good
thing, I don't know.

It means Windows 2000 (NT 5.0 not just "Server"), and Windows XP (NT 5.1
and NT 5.2).

7 was maybe the last actually usable Windows OS before they
started making everything network-dependent.

ReactOS has a future... there isn't much there even after so many years
of development, but the worse Windows gets the more argument there is to
put support into ReactOS development.
--scott

The worse Windows NT gets might actually be an argument to opt by
e.g. GNU/Linux or a BSDs (please don't read this paragraph, Lawrence).
--
Nuno Silva
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.misc

Nuno Silva wrote:

Well, at least Windows 4.10 tends to be quite well regarded

By Windows 4.10, you mean one of the DOS based versions? Perhaps Win98? Imagine getting that to work with UEFI and multi-core, even simple SMP.
Dave
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.misc

Scott Dorsey wrote:

NT 5.* I assume means Windows Server 2000 which might have been a good
thing, I don't know.

Somewhere I'll have a stack of Windows 5.0 Workstation beta CDs, from
before they renamed it to Win2K
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.misc

Dave Yeo wrote:

Nuno Silva wrote:

Well, at least Windows 4.10 tends to be quite well regarded

By Windows 4.10, you mean one of the DOS based versions? Perhaps Win98? Imagine getting that to work with UEFI and multi-core

Back then I was running WinNT 3.51 with "Shell technology preview" to
get the start menu look ...

even simple SMP.

I'd upgraded to WinNT 4.0 Workstation before getting the dual-celeron
BP6 motherboard.
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.misc

On 2026-10-11, Dave Yeo wrote:

Nuno Silva wrote:

Well, at least Windows 4.10 tends to be quite well regarded

By Windows 4.10, you mean one of the DOS based versions?

Was there any Windows version that wasn't DOS-based to some extent?
Perhaps 4.90?

Perhaps Win98?

4.10.* is 98, indeed, both first and second editions.

Imagine getting that to work with UEFI and multi-core,
even simple SMP.

If UEFI can't somehow boot 98, then that'd be a serious limitation of
UEFI. I doubt there is such a limitation, but if there is, then that's
quite the major roadblock, and would be an issue with UEFI, not with
Windows 98.

As for multiple processors and cores, wouldn't it just transparently use
one of them? Inefficient, yes. But not an insurmountable problem.

But maybe by open-sourcing something could be added to handle that to
some extent, if it's possible. Also, given I also mentioned
open-sourcing NT 5.*, why not use that then if multiple processor and/or
core support is necessary?

Dave

--
Nuno Silva
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.misc

Nuno Silva wrote:

On 2026-10-11, Dave Yeo wrote:

Nuno Silva wrote:

Well, at least Windows 4.10 tends to be quite well regarded

By Windows 4.10, you mean one of the DOS based versions?

Was there any Windows version that wasn't DOS-based to some extent?
Perhaps 4.90?

At work, all the company's mainframes reside in Windows 11, in the
"DosBox" Emulator.
--
Do you want a friendship with chaos?
Wanna party with Perses?
Consequences mean nothing to me
Morning is your problem, I don't stick around
The damage is done, it’s all you from here
And don’t come a looking 'cause I simply do not care
Like you’d even find me
Well, maybe after a few pints
And I only make an appearance when the time is right
You see, I'm all about the here and now
Live in the moment and all that, carpe diem
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.misc

Nuno Silva wrote:

On 2026-10-11, Dave Yeo wrote:

Nuno Silva wrote:

Well, at least Windows 4.10 tends to be quite well regarded

By Windows 4.10, you mean one of the DOS based versions?

Was there any Windows version that wasn't DOS-based to some extent?
Perhaps 4.90?

Well, NT. Even Win ME, which I guess was 4.90, still depended on an
under laying DOS boot. Maybe the DOS boot part could be rewritten. I'd
guess MS considered it and decided against it.
At least it supported the NT driver environment as well as DOS based
drivers, which should be ripped out. My understanding was a lot of
instability came from mixing the drivers and if only NT style drivers
were used, it was actually stable

Perhaps Win98?

4.10.* is 98, indeed, both first and second editions.

Imagine getting that to work with UEFI and multi-core,
even simple SMP.

If UEFI can't somehow boot 98, then that'd be a serious limitation of
UEFI. I doubt there is such a limitation, but if there is, then that's
quite the major roadblock, and would be an issue with UEFI, not with
Windows 98.

Well, there's UEFI with CSM support, which might be able to boot a shim
that boots Win98. Without CSM support, I'm not sure but think it would
take a much bigger shim that basically creates an old BIOS environment.
CSM support is going away. The lower 4GB is also getting ignored as a
safe space for 32 bit operating systems. PCI address expanding into it,
some systems putting the frame buffer above 4GB with no way to move it
down. Of course a GB or 2 of accessible memory might be fine in a Win98 environment, though it seems all modern software now expects a 64 bit
address space.

As for multiple processors and cores, wouldn't it just transparently use
one of them? Inefficient, yes. But not an insurmountable problem.

That is true.

But maybe by open-sourcing something could be added to handle that to
some extent, if it's possible. Also, given I also mentioned
open-sourcing NT 5.*, why not use that then if multiple processor and/or
core support is necessary?

NT 5.x would be much more doable. Shame that some of the code is likely
still being used and what is unknown with all old versions of Windows is
how much is actually owned by MS and how much is only licensed. Figuring
that out could be a major project which I doubt MS would be interested
in under taking.
Dave


--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.misc

On Sun, 11 Oct 2026 10:39:16 -0700
Dave Yeo <[email protected]> wrote:

Nuno Silva wrote:

On 2026-10-11, Dave Yeo wrote:

Nuno Silva wrote:

Well, at least Windows 4.10 tends to be quite well regarded

By Windows 4.10, you mean one of the DOS based versions?

Was there any Windows version that wasn't DOS-based to some extent?
Perhaps 4.90?

Well, NT. Even Win ME, which I guess was 4.90, still depended on an
under laying DOS boot. Maybe the DOS boot part could be rewritten. I'd
guess MS considered it and decided against it.
At least it supported the NT driver environment as well as DOS based drivers, which should be ripped out. My understanding was a lot of instability came from mixing the drivers and if only NT style drivers
were used, it was actually stable

Perhaps Win98?

4.10.* is 98, indeed, both first and second editions.

Imagine getting that to work with UEFI and multi-core,
even simple SMP.

If UEFI can't somehow boot 98, then that'd be a serious limitation of
UEFI. I doubt there is such a limitation, but if there is, then that's quite the major roadblock, and would be an issue with UEFI, not with Windows 98.

Well, there's UEFI with CSM support, which might be able to boot a shim
that boots Win98. Without CSM support, I'm not sure but think it would
take a much bigger shim that basically creates an old BIOS environment.
CSM support is going away. The lower 4GB is also getting ignored as a
safe space for 32 bit operating systems. PCI address expanding into it,
some systems putting the frame buffer above 4GB with no way to move it
down. Of course a GB or 2 of accessible memory might be fine in a Win98 environment, though it seems all modern software now expects a 64 bit address space.

As for multiple processors and cores, wouldn't it just transparently use one of them? Inefficient, yes. But not an insurmountable problem.

That is true.

But maybe by open-sourcing something could be added to handle that to
some extent, if it's possible. Also, given I also mentioned
open-sourcing NT 5.*, why not use that then if multiple processor and/or core support is necessary?

NT 5.x would be much more doable. Shame that some of the code is likely still being used and what is unknown with all old versions of Windows is
how much is actually owned by MS and how much is only licensed. Figuring that out could be a major project which I doubt MS would be interested
in under taking.
Dave

Use ReactOS?
--
Bah, and indeed Humbug.
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.misc

Kerr-Mudd, John wrote:

NT 5.x would be much more doable. Shame that some of the code is likely
still being used and what is unknown with all old versions of Windows is
how much is actually owned by MS and how much is only licensed. Figuring
that out could be a major project which I doubt MS would be interested
in under taking.
Dave

Use ReactOS?

All these ideas have one shortage, device drivers. Not sure of the
current state of ReactOS but at least Linux+WINE has the advantage of supporting a lot of hardware
Dave
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.misc

Dave Yeo wrote:

Kerr-Mudd, John wrote:

NT 5.x would be much more doable. Shame that some of the code is likely
still being used and what is unknown with all old versions of Windows is >>> how much is actually owned by MS and how much is only licensed. Figuring >>> that out could be a major project which I doubt MS would be interested
in under taking.
Dave

Use ReactOS?

All these ideas have one shortage, device drivers. Not sure of the
current state of ReactOS but at least Linux+WINE has the advantage of supporting a lot of hardware
Dave

Windows 11 supports more software than Linux or WINE plus both combined.
You didn't know that? It's a national standard.
--
Do you want a friendship with chaos?
Wanna party with Perses?
Consequences mean nothing to me
Morning is your problem, I don't stick around
The damage is done, it’s all you from here
And don’t come a looking 'cause I simply do not care
Like you’d even find me
Well, maybe after a few pints
And I only make an appearance when the time is right
You see, I'm all about the here and now
Live in the moment and all that, carpe diem
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.misc

On Sun, 11 Oct 2026 10:39:16 -0700, Dave Yeo wrote:

Well, NT. Even Win ME, which I guess was 4.90, still depended on an
under laying DOS boot.

Remember that NT was an entirely new kernel, but the first version was
numbered “3.1”, specifically to sync up (and be confused with) the
version numbering for the DOS-based Windows.

Maybe the DOS boot part could be rewritten. I'd guess MS considered
it and decided against it. At least it supported the NT driver
environment as well as DOS based drivers, which should be ripped
out. My understanding was a lot of instability came from mixing the
drivers and if only NT style drivers were used, it was actually
stable

I can’t imagine NT-style drivers working at all well on a DOS-based foundation. DOS (and DOS-based Windows) was perfectly capable of being
unstable and flaky all on its own, it didn’t need any extra help from grafting on an extra driver model -- which was supposed to increase,
not reduce, stability anyway.
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.misc

On Sun, 11 Oct 2026 17:53:57 -0600, Blaine wrote:

Windows 11 supports more software than Linux or WINE plus both combined.

Not likely.
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.misc

On Sun, 11 Oct 2026 14:04:38 -0700, Dave Yeo wrote:

All these ideas have one shortage, device drivers. Not sure of the
current state of ReactOS but at least Linux+WINE has the advantage
of supporting a lot of hardware

This is why I think ReactOS really only has museum value. Anything
intended for real-world use would need widespread support for
real-world hardware. This means having a Linux underpinning. If you
want Windows APIs, then that means putting WINE on top.
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.misc

Blaine wrote:

Windows 11 supports more software than Linux or WINE plus both combined.
You didn't know that? It's a national standard.

If you consider Android as Linux (same kernel), Linux wins hands down. otherwise I'd have to ask for a citation
Dave
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.misc

Lawrence D’Oliveiro wrote:

On Sun, 11 Oct 2026 10:39:16 -0700, Dave Yeo wrote:

Well, NT. Even Win ME, which I guess was 4.90, still depended on an
under laying DOS boot.

Remember that NT was an entirely new kernel, but the first version was numbered “3.1”, specifically to sync up (and be confused with) the version numbering for the DOS-based Windows.

Actually the 1st version to boot up was OS/2 v3 NT (or was it OS/2 NT
v3). MS got versions 3+ in the divorce while OS/2 internal versions were
2.x ending with 2.45.
Before the divorce, it was already agreed IBM would pursue Intel 32 bit
OS/2 (ver 2.x) and MS would purse the next generation, ver 3+.
Of course with the success of Win 3.x, MS pivoted to the Win32 API and
luckily the version numbers aligned.

Maybe the DOS boot part could be rewritten. I'd guess MS considered
it and decided against it. At least it supported the NT driver
environment as well as DOS based drivers, which should be ripped
out. My understanding was a lot of instability came from mixing the
drivers and if only NT style drivers were used, it was actually
stable

I can’t imagine NT-style drivers working at all well on a DOS-based foundation. DOS (and DOS-based Windows) was perfectly capable of being unstable and flaky all on its own, it didn’t need any extra help from grafting on an extra driver model -- which was supposed to increase,
not reduce, stability anyway.

I don't know enough, just repeating what I heard at the time.
Dave

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.misc

Dave Yeo wrote:

Blaine wrote:

Windows 11 supports more software than Linux or WINE plus both combined.
You didn't know that? It's a national standard.

If you consider Android as Linux (same kernel), Linux wins hands down. otherwise I'd have to ask for a citation
Dave

Ah, that's probably the rub. I missed the boat on Android & mobile in
general because the guy leading the project wouldn't give me any money
up front. He was one of those overachievers, who walks ten feet in front
of you everywhere you go together. He didn't actually do that, but he
was never much into the concept of keeping me on the same page or
looking in on me, maybe because he just assumed I was very high
functioning and didn't need any babying, not to mention, if I said all
this to him back then, he would have an episode from The Butterfly
Effect when Ashton Kutcher speaks to his dad with the omniscience of a twenty-seven year old.
--
Do you want a friendship with chaos?
Wanna party with Perses?
Consequences mean nothing to me
Morning is your problem, I don't stick around
The damage is done, it’s all you from here
And don’t come a looking 'cause I simply do not care
Like you’d even find me
Well, maybe after a few pints
And I only make an appearance when the time is right
You see, I'm all about the here and now
Live in the moment and all that, carpe diem
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.misc

On 2026-10-12, Dave Yeo wrote:

Lawrence D’Oliveiro wrote:

On Sun, 11 Oct 2026 10:39:16 -0700, Dave Yeo wrote:

Well, NT. Even Win ME, which I guess was 4.90, still depended on an
under laying DOS boot.

Remember that NT was an entirely new kernel, but the first version was
numbered “3.1”, specifically to sync up (and be confused with) the
version numbering for the DOS-based Windows.

Actually the 1st version to boot up was OS/2 v3 NT (or was it OS/2 NT
v3). MS got versions 3+ in the divorce while OS/2 internal versions
were 2.x ending with 2.45.
Before the divorce, it was already agreed IBM would pursue Intel 32
bit OS/2 (ver 2.x) and MS would purse the next generation, ver 3+.
Of course with the success of Win 3.x, MS pivoted to the Win32 API and luckily the version numbers aligned.

Maybe the DOS boot part could be rewritten. I'd guess MS considered
it and decided against it. At least it supported the NT driver
environment as well as DOS based drivers, which should be ripped
out. My understanding was a lot of instability came from mixing the
drivers and if only NT style drivers were used, it was actually
stable

I can’t imagine NT-style drivers working at all well on a DOS-based
foundation. DOS (and DOS-based Windows) was perfectly capable of being
unstable and flaky all on its own, it didn’t need any extra help from
grafting on an extra driver model -- which was supposed to increase,
not reduce, stability anyway.

I don't know enough, just repeating what I heard at the time.

I think you conflated two things there (but it could always be that I'm mistaking some things myself).

The "portable" driver part is the Windows Driver Model, that has been
present in non-NT at least since 4.10.1998. In theory, at least some
drivers would be usable both for 98 and XP, for example.

(Wikipedia says, at a quick glance, that it was introduced in 4.10.1998
and NT 5.0, and that it's intended to be forward-compatible [1].)

[1] <https://enwp.org/Windows_Driver_Model>

Now the instability part, maybe it was some other driver code, or possibly networking or networking drivers, being added to Windows from windows
NT, or being somehow merged. I heard something along these lines many
years ago, but it was about the first MS attempt to depart from the DOS-bootstrapped Windows product line: Windows 4.90 "Millenium Edition",
and with the implication that it didn't go very well.

(The second such attempt, at least the second to reach commercialization
to home users, was NT 5.1 "XP", featuring Rover from Microsoft Bob.)

(I'm still surprised Microsoft hasn't brought back MS Agent as a
frontend to GenAI.)

Dave

--
Nuno Silva
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.misc

Lawrence =?iso-8859-13?q?D=FFOliveiro?= <[email protected]d> wrote:

On Sun, 11 Oct 2026 14:04:38 -0700, Dave Yeo wrote:

All these ideas have one shortage, device drivers. Not sure of the
current state of ReactOS but at least Linux+WINE has the advantage
of supporting a lot of hardware

This is why I think ReactOS really only has museum value. Anything
intended for real-world use would need widespread support for
real-world hardware. This means having a Linux underpinning. If you
want Windows APIs, then that means putting WINE on top.

No problem, as long as you have the ability to support the same driver
model as Windows, so that third-party Windows drivers will run on your
OS.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.misc

On Sun, 11 Oct 2026 22:58:21 -0700, Dave Yeo wrote:

Lawrence D’Oliveiro wrote:

Remember that NT was an entirely new kernel, but the first version
was numbered “3.1”, specifically to sync up (and be confused with)
the version numbering for the DOS-based Windows.

Actually the 1st version to boot up was OS/2 v3 NT (or was it OS/2
NT v3). MS got versions 3+ in the divorce while OS/2 internal
versions were 2.x ending with 2.45.

Before the divorce, it was already agreed IBM would pursue Intel 32
bit OS/2 (ver 2.x) and MS would purse the next generation, ver 3+.
Of course with the success of Win 3.x, MS pivoted to the Win32 API
and luckily the version numbers aligned.

That might make sense, if the first version of NT was 3.0, not 3.1.

I don’t think the version numbering for the “OS/2” project had any relevance to what later became Windows NT. Dave Cutler came in,
chucked it all aside and started again, throwing out all the IBM-isms,
to reincarnate DEC’s VMS.
--- Synchronet 3.22a-Linux NewsLink 1.2