Before NT was released for Alpha, Microsoft was doing early 64-bit
test builds with Windows 2000 on Alpha. One of those early 64-bit
experimental builds has now been discovered:

https://www.theregister.com/2023/05/19/first_64bit_windows/

and directly here:

https://virtuallyfun.com/2023/05/15/windows-2000-64-bit-for-alpha-axp/

Once again, I am reminded how much more readable the GUI designs were
back in those days when compared to today.

Simon.

PS: BTW, am I the only one who thinks that Windows 10 font quality is
actually inferior to the font quality on Windows 7 ?

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.

In article <u47q84$lvps$1@dont-email.me>,
clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley) wrote:

> Before NT was released for Alpha, Microsoft was doing early 64-bit
> test builds with Windows 2000 on Alpha.

Presumably, you mean "Before NT was released for Itanium ..."?

> PS: BTW, am I the only one who thinks that Windows 10 font quality
> is actually inferior to the font quality on Windows 7?

I think the Windows 10 fonts are optimised for higher-resolution screens.

John

On 2023-05-19, John Dallman <jgd@cix.co.uk> wrote:
> In article <u47q84$lvps$1@dont-email.me>,
> clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley) wrote:
>
>> Before NT was released for Alpha, Microsoft was doing early 64-bit
>> test builds with Windows 2000 on Alpha.
>
> Presumably, you mean "Before NT was released for Itanium ..."?
>

No. I meant Alpha.

Check out the original design specifications for NT. It was designed as
a very architecture neutral operating system and Cutler brought across
a lot of ideas from Prism/MICA.

In fact about the only major idea he didn't bring across was the use of
Pillar (or something comparable) instead of C as the SIL.

We lost a major opportunity to establish a viable alternative to C because
of that and the security of our computing infrastructure is poorer because
of it.

>> PS: BTW, am I the only one who thinks that Windows 10 font quality
>> is actually inferior to the font quality on Windows 7?
>
> I think the Windows 10 fonts are optimised for higher-resolution screens.
>

Interesting, thanks. What I can tell you is that they don't look as good
as the Windows 7 fonts (to me at least) on normal resolution monitors.

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.

On 2023-05-19, Simon Clubley <clubley@remove_me.eisner.decus.org-Earth.UFP> wrote:
> On 2023-05-19, John Dallman <jgd@cix.co.uk> wrote:
>> In article <u47q84$lvps$1@dont-email.me>,
>> clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley) wrote:
>>
>>> Before NT was released for Alpha, Microsoft was doing early 64-bit
>>> test builds with Windows 2000 on Alpha.
>>
>> Presumably, you mean "Before NT was released for Itanium ..."?
>>
>
> No. I meant Alpha.
>

And before anyone takes great pleasure in pointing it out, :-) I now see
what you meant in that Windows 2000 was way _after_ NT so yes, I got the
timeline _way_ wrong.

That's what happens when you post while feeling under the weather... :-(

Sorry for the confusion,

Simon.

PS: My comment about Pillar still stands however. Cutler missed a real
opportunity there IMHO.

> Check out the original design specifications for NT. It was designed as
> a very architecture neutral operating system and Cutler brought across
> a lot of ideas from Prism/MICA.
>
> In fact about the only major idea he didn't bring across was the use of
> Pillar (or something comparable) instead of C as the SIL.
>
> We lost a major opportunity to establish a viable alternative to C because
> of that and the security of our computing infrastructure is poorer because
> of it.
>
>>> PS: BTW, am I the only one who thinks that Windows 10 font quality
>>> is actually inferior to the font quality on Windows 7?
>>
>> I think the Windows 10 fonts are optimised for higher-resolution screens.
>>
>
> Interesting, thanks. What I can tell you is that they don't look as good
> as the Windows 7 fonts (to me at least) on normal resolution monitors.
>
> Simon.
>

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.

In article <u48n61$pdn1$1@dont-email.me>,
clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley) wrote:

> In fact about the only major idea he didn't bring across was the
> use of Pillar (or something comparable) instead of C as the SIL.
>
> We lost a major opportunity to establish a viable alternative to C
> because of that and the security of our computing infrastructure
> is poorer because of it.

I think it was already too late by 1988. Such an OS would have had to
have supported C and C++ for software portability. Microsoft would have
had to implement any new language, but must have wanted to get on with
Windows NT. Also, their developers already knew C/C++ and they really
weren't very security-conscious in those days.

DEC had languages of its own, like BLISS, and was more willing to use
different languages for different projects. The elaborate VAX calling
standard meant that integrating modules written in different languages
was reasonably straightforward. That was not the case with any Microsoft
environment until .NET appeared in 2003.

John

On 5/19/2023 4:45 PM, Simon Clubley wrote:
> In fact about the only major idea he didn't bring across was the use of
> Pillar (or something comparable) instead of C as the SIL.
>
> We lost a major opportunity to establish a viable alternative to C because
> of that and the security of our computing infrastructure is poorer because
> of it.

I would tend to agree.

But there were a long period from late 80's to mid 10's
where C was considered the only viable language for OS
development.

NT was designed in (the beginning of) that period.
Cutler decided to follow the trend on that one.

Back in time "begin end" languages were not unheard of for
OS development. Multics used PL/I, NOS/VE used Cybil,
MPE used Pascal etc..

And in recent years both the Linux and Windows teams are
embracing Rust.

But in between ...

Arne

On 5/19/2023 7:28 PM, John Dallman wrote:
> In article <u48n61$pdn1$1@dont-email.me>,
> clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley) wrote:
>> In fact about the only major idea he didn't bring across was the
>> use of Pillar (or something comparable) instead of C as the SIL.
>>
>> We lost a major opportunity to establish a viable alternative to C
>> because of that and the security of our computing infrastructure
>> is poorer because of it.
>
> I think it was already too late by 1988. Such an OS would have had to
> have supported C and C++ for software portability.

It would need to support standard C and standard C++ and
possible Posix API, but that does not preclude the OS
being written in something else than C/C++ and also
exposing a non-C/C+ centric API.

> Microsoft would have
> had to implement any new language but must have wanted to get on with > Windows NT.

MS has implemented lots of new languages or at least new flavors
of languages.

> Also, their developers already knew C/C++ and they really
> weren't very security-conscious in those days.

True.

OS/2 was C and C++.

> DEC had languages of its own, like BLISS,

Bliss is a bit low level compared to other languages
mentioned in this thread.

> and was more willing to use
> different languages for different projects. The elaborate VAX calling
> standard meant that integrating modules written in different languages
> was reasonably straightforward.

Yes. DEC was ahead of time on that one.

> That was not the case with any Microsoft
> environment until .NET appeared in 2003.

(1.0 was 2002 and 1.1 was 2003)

..NET CTS was their first let us call it VAX calling standard easy to
use solution.

But COM back from 1993 intended the same "integrating
modules written in different languages" - mixing C++, VB6,
Delphi, VBS etc.. But I am not so sure about the
"reasonably straightforward" though. Using a COM
component from any language but C/C++ is relative
easy. But the rest is more tricky.

Arne

In article <u490md$qipd$1@dont-email.me>,
Arne Vajhøj <arne@vajhoej.dk> wrote:
>On 5/19/2023 4:45 PM, Simon Clubley wrote:
>> In fact about the only major idea he didn't bring across was the use of
>> Pillar (or something comparable) instead of C as the SIL.
>>
>> We lost a major opportunity to establish a viable alternative to C because
>> of that and the security of our computing infrastructure is poorer because
>> of it.
>
>I would tend to agree.
>
>But there were a long period from late 80's to mid 10's
>where C was considered the only viable language for OS
>development.

Nonsense.

>NT was designed in (the beginning of) that period.
>Cutler decided to follow the trend on that one.
>
>Back in time "begin end" languages were not unheard of for
>OS development. Multics used PL/I, NOS/VE used Cybil,
>MPE used Pascal etc..
>
>And in recent years both the Linux and Windows teams are
>embracing Rust.
>
>But in between ...

....but in between lots of systems were implemented using C++
(K42, BeOS, Haiku, L4/Fiasco), Modula-3 (SPIN), Oberon (the
Oberon system), Ada, and on and on. Even C# (Singularity,
Midori) was used.

- Dan C.

On 2023-05-19, John Dallman <jgd@cix.co.uk> wrote:
> In article <u48n61$pdn1$1@dont-email.me>,
> clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley) wrote:
>
>> In fact about the only major idea he didn't bring across was the
>> use of Pillar (or something comparable) instead of C as the SIL.
>>
>> We lost a major opportunity to establish a viable alternative to C
>> because of that and the security of our computing infrastructure
>> is poorer because of it.
>
> I think it was already too late by 1988. Such an OS would have had to
> have supported C and C++ for software portability. Microsoft would have
> had to implement any new language, but must have wanted to get on with
> Windows NT. Also, their developers already knew C/C++ and they really
> weren't very security-conscious in those days.
>

DEC had already thrashed out many of those issues by then for PRISM, if
not fully implemented them before it was cancelled (according to the
public material on PRISM).

In particular, they considered both Pillar and C to be system programming
languages for PRISM. You may find the following memo from the time
interesting:

https://bitsavers.org/pdf/dec/prism/mica/870827_DECwest_Compiler_Project.pdf

Section 3 is the interesting section, where DEC talks about it using Pillar
due to the added software robustness, while customers would be likely to
use C because that is what the customers were familiar with.

BTW, I notice from that memo how Pascal was going to be in the _first_
wave of available user-level compilers for PRISM. :-)

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.

In article <u4fnn3$25v19$1@dont-email.me>,
clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley) wrote:
> On 2023-05-19, John Dallman <jgd@cix.co.uk> wrote:
> > I think it was already too late by 1988. Such an OS would have
> > had to have supported C and C++ for software portability. Microsoft
> > would have had to implement any new language, but must have wanted
> > to get on with Windows NT. Also, their developers already knew C/C++
> > and they really weren't very security-conscious in those days.
>
> DEC had already thrashed out many of those issues by then for
> PRISM, if not fully implemented them before it was cancelled
> (according to the public material on PRISM).
>
> You may find the following memo from the time
> interesting:

Yup. I've seen that. Dave Cutler and his team brought ideas to Microsoft,
but no implementations AFAIK. Microsoft are always very ready to believe
that their way of doing things is correct, and likely would have regarded
PILLAR or any other safer language as a distraction.

A friend of mine was working for IBM on OS/2 at the time, and they felt
that microsoft were a bunch of cowboys, taking no care and accepting
anything that worked. The falling-out was more or less inevitable.

John

Simon Clubley <clubley@remove_me.eisner.decus.org-Earth.UFP> writes:

> Before NT was released for Alpha, Microsoft was doing early 64-bit
> test builds with Windows 2000 on Alpha. One of those early 64-bit
> experimental builds has now been discovered:
>
> https://www.theregister.com/2023/05/19/first_64bit_windows/
>
> and directly here:
>
> https://virtuallyfun.com/2023/05/15/windows-2000-64-bit-for-alpha-axp/

I had been a beta tester for Microsoft and tried the betas of Win 2K for
Alpha. They more or less worked, but Red Hat Linux worked much better on
that hardware (I think I had an Alphastation 200? IIRC, it was a 233MHz
Alpha CPU.)

I'm sure I threw away those CDs long ago, even before I stopped using
Windows.

On 6/5/2023 12:28 PM, Bud Frede wrote:
> Simon Clubley <clubley@remove_me.eisner.decus.org-Earth.UFP> writes:
>
>> Before NT was released for Alpha, Microsoft was doing early 64-bit
>> test builds with Windows 2000 on Alpha. One of those early 64-bit
>> experimental builds has now been discovered:
>>
>> https://www.theregister.com/2023/05/19/first_64bit_windows/
>>
>> and directly here:
>>
>> https://virtuallyfun.com/2023/05/15/windows-2000-64-bit-for-alpha-axp/
>
> I had been a beta tester for Microsoft and tried the betas of Win 2K for
> Alpha. They more or less worked, but Red Hat Linux worked much better on
> that hardware (I think I had an Alphastation 200? IIRC, it was a 233MHz
> Alpha CPU.)
>
> I'm sure I threw away those CDs long ago, even before I stopped using
> Windows.
>
>

We had Windows 2000 running on some Alphas at the University.
They all, eventually, moved to Linux or were just abandoned.
The reason. Something that has been said about VMS here
numerous times. No applications. Users don't buy machines
because of the architecture or operating system. They buy
it to get a job done and if they can't get the applications
to do that the system is worthless. In this day and age where
less and less programming is being done in-house this is an
even more important issue.

bill

On 6/6/2023 8:28 AM, bill wrote:
>> I had been a beta tester for Microsoft and tried the betas of Win 2K for
>> Alpha. They more or less worked, but Red Hat Linux worked much better on
>> that hardware (I think I had an Alphastation 200? IIRC, it was a 233MHz
>> Alpha CPU.)

233 Mhz must have been either a 200 or a 255.

>> I'm sure I threw away those CDs long ago, even before I stopped using
>> Windows.
>
> We had Windows 2000 running on some Alphas at the University.
> They all, eventually, moved to Linux or were just abandoned.
> The reason.  Something that has been said about VMS here
> numerous times.  No applications.  Users don't buy machines
> because of the architecture or operating system.  They buy
> it to get a job done and if they can't get the applications
> to do that the system is worthless.  In this day and age where
> less and less programming is being done in-house this is an
> even more important issue.

Software is important.

I believe the target software for Windows / Alpha was
database software.

Back then databases was one of the few thing that could
benefit from 64 bit.

Arne