On Sat, 21 Sep 2024 19:36:45 +0100, The Natural Philosopher wrote:

> On 21/09/2024 11:34, Lester Thorpe wrote:
>> On Sat, 21 Sep 2024 09:17:22 +0100, The Natural Philosopher wrote:
>>
>>>
>>> Ah, Just another troll
>>>
>>
>> Ah, just another self-proclaimed, stuck-up moderator.
>>
>>
> Sorry, I was just talking to myself.
>

That's a bad habit.

>
> I didn't mean to offend your precious sensibilities.
>

Me thinks you got that backwards.

>
> It was merely a note to self to plonk the aforesaid wanker
>

It's the usual recourse of a failed self-proclaimed moderator.

--
Systemd: solving all the problems that you never knew you had.

In article <pan$9c21$def65fdf$ff0cb13f$76105de1@gnu.rocks>, lt@gnu.rocks
(Lester Thorpe) wrote:

> I would love to see Hurd revitalized and to assume the role of
> the GNU kernel, replacing Linux.

That seems hard to achieve now. If a BSD kernel had been adopted
originally, Linux might well never have become popular. Getting the Hurd
to catch up to what the Linux kernel can do now, but doing it better,
would be an immense task.

Relatively few people and organisations want to replace chunks of their
operating systems, and hence the attractiveness of a system design which
makes that easy is quite limited.

John

On Sat, 21 Sep 2024 21:08 +0100 (BST), John Dallman wrote:

> Getting the Hurd to catch up to what the Linux kernel can do now, but
> doing it better, would be an immense task.

Building it as a microkernel was supposed to make it so much easier to
develop, more reliable, less buggy, more secure and all that. But none of
that has come to pass.

On Fri, 20 Sep 2024 18:15:59 -0700, Bobbie Sellers wrote:

> The AmigaOS ran on the hardware and lots of gamers and
> demo-coders would exploit the hardware ignoring the OS. The OS had
> everything but memory management.

It didn’t have a very good graphics abstraction layer. In short, it wasn’t
a very good OS, which is why the apps ignored it.

On 9/21/24 18:15, Lawrence D'Oliveiro wrote:
> On Fri, 20 Sep 2024 18:15:59 -0700, Bobbie Sellers wrote:
>
>> The AmigaOS ran on the hardware and lots of gamers and
>> demo-coders would exploit the hardware ignoring the OS. The OS had
>> everything but memory management.
>
> It didn’t have a very good graphics abstraction layer. In short, it wasn’t
> a very good OS, which is why the apps ignored it.

The applications I used used the OS or I did not buy them.
It was the best multi-tasking OS aside from Unix at the time of its
creation. Amigas could use Unix as well at least the A2000,
the A3000, and A4000 but mostly the higher end were used as
places to run the Video Toaster hardward on. Sometimes they
were rebadged as Macs or simply as Video Toasters. I guess
no one is still using them but not so many years ago when
a problem happened the old Amiga reboot symbol would show
up on our analog TV broadcasts, sometimes on cable channels.
They were used by the way in another hobby by anime enthusiasts
to subtitle on VHS the Japanese cartoons.

bliss

--
b l i s s - S F 4 e v e r at D S L E x t r e m e dot com

On Sat, 21 Sep 2024 22:35:01 -0700, Bobbie Sellers wrote:

> On 9/21/24 18:15, Lawrence D'Oliveiro wrote:
>
>> On Fri, 20 Sep 2024 18:15:59 -0700, Bobbie Sellers wrote:
>>
>>> The AmigaOS ran on the hardware and lots of gamers and demo-coders
>>> would exploit the hardware ignoring the OS. The OS had everything but
>>> memory management.
>>
>> It didn’t have a very good graphics abstraction layer. In short, it
>> wasn’t a very good OS, which is why the apps ignored it.
>
> The applications I used used the OS or I did not buy them.

You can’t have had much of a choice of apps, then.

> ... mostly the higher end were used as places to
> run the Video Toaster hardward on.

The Video Toaster is a prime example of a product that was so heavily tied
to hardware features of the Amiga, it was an NTSC-only product: there
could be no PAL version because there was no room in the Amiga memory
layout for that to work.

On 21/09/2024 21:08, John Dallman wrote:
> In article <pan$9c21$def65fdf$ff0cb13f$76105de1@gnu.rocks>, lt@gnu.rocks
> (Lester Thorpe) wrote:
>
>> I would love to see Hurd revitalized and to assume the role of
>> the GNU kernel, replacing Linux.
>
> That seems hard to achieve now. If a BSD kernel had been adopted
> originally, Linux might well never have become popular. Getting the Hurd
> to catch up to what the Linux kernel can do now, but doing it better,
> would be an immense task.
>
> Relatively few people and organisations want to replace chunks of their
> operating systems, and hence the attractiveness of a system design which
> makes that easy is quite limited.
>
> John

I think the real issues is that outside of the Linux Gopis, no one gives
a fuck what OS they use as long as its

functional
cheap
supportable
stable
reliable
secure

At the moment that is linux.

--
No Apple devices were knowingly used in the preparation of this post.

On Sat, 21 Sep 2024 19:34:31 +0100, The Natural Philosopher wrote:

[snip]

>> Millions of people bought the 8088: the original IBM PC and its clones
>> used it. Interestingly, the Amstrad PC used the 8086. And Olivetti, but
>> dunno if all models.
>>
>>
> My PC clone had an NEC V20 and a hard drive of a monumental 20MB IIRC.
> It was all I could afford

I had a V20 too, and (at least where I bought it) I could get a 30MB (RLL)
for only $20 more than the 20MB. It was one where you could still do a
low-level format. Enter DEBUG and a command like "G=C000:000C". I first
got Spinrite because it could change the interleave,

--
94 days until the winter celebration (Wednesday, December 25, 2024
12:00:00 AM for 1 day).

Mark Lloyd
http://notstupid.us/

"Theology: The study of elaborate verbal disguises for non-ideas."

Mark Lloyd wrote this copyrighted missive and expects royalties:

> On Sat, 21 Sep 2024 19:34:31 +0100, The Natural Philosopher wrote:
>
> [snip]
>
>>> Millions of people bought the 8088: the original IBM PC and its clones
>>> used it. Interestingly, the Amstrad PC used the 8086. And Olivetti, but
>>> dunno if all models.
>>>
>> My PC clone had an NEC V20 and a hard drive of a monumental 20MB IIRC.
>> It was all I could afford
>
> I had a V20 too, and (at least where I bought it) I could get a 30MB (RLL)
> for only $20 more than the 20MB. It was one where you could still do a
> low-level format. Enter DEBUG and a command like "G=C000:000C". I first
> got Spinrite because it could change the interleave,

I remember the "Twister" format you could use on the Atari ST to speed up
floppy copy.

--
Spouse, n.:
Someone who'll stand by you through all the trouble you
wouldn't have had if you'd stayed single.

On 2024-09-22 20:02, Mark Lloyd wrote:
> On Sat, 21 Sep 2024 19:34:31 +0100, The Natural Philosopher wrote:
>
> [snip]
>
>>> Millions of people bought the 8088: the original IBM PC and its clones
>>> used it. Interestingly, the Amstrad PC used the 8086. And Olivetti, but
>>> dunno if all models.
>>>
>>>
>> My PC clone had an NEC V20 and a hard drive of a monumental 20MB IIRC.
>> It was all I could afford
>
> I had a V20 too, and (at least where I bought it) I could get a 30MB (RLL)
> for only $20 more than the 20MB. It was one where you could still do a
> low-level format. Enter DEBUG and a command like "G=C000:000C". I first
> got Spinrite because it could change the interleave,

I had an Amstrad PC 1512 DD, ie, dual floppy. I later bought a HD
mounted on a card, 32 megs which apparently was the maximum that could
be done. Friends of mine had a 10MB HD and complained it was full. So I
went for bigger.

Yes, I remember the low-level format thing. The program in my case asked
for an interleave factor, and there was a recommendation for 3. I tested
several values, and something like 12 got double speed. Amazing at the
time. There was also a table of bad sectors to enter manually, but it
did not work right because the format found bad sectors and marked them
as such in the FAT.

I still have that machine, should still run. Walk.

--
Cheers, Carlos.

On 2024-09-20 10:41, The Natural Philosopher wrote:
> On 20/09/2024 01:23, clinker wrote:
>> Not to derail this thread, which is a good one, but it is possible an
>> operating system will be developed that threaten both Linux and
>> Windows existence as it is known today.
>>
> most things are *possible*
>
>> OS2 had a good run for example. It is still alive and well in some
>> proprietary industrial equipment because it is impervious to hacks as
>> no one remembers it.
>
> DOS is still used in a lot of industrial kit ...When all you want is to
> load a single tasked application that talks to custom hardware...

Makes sense. I worked in that sector in the 90's. The only problem with
MsDOS was the small memory available. We had to use overlays in the
software, or load/download programs with a stub.

--
Cheers, Carlos.

On 25/09/2024 08:57, Carlos E.R. wrote:
> On 2024-09-20 10:41, The Natural Philosopher wrote:
>> On 20/09/2024 01:23, clinker wrote:
>>> Not to derail this thread, which is a good one, but it is possible an
>>> operating system will be developed that threaten both Linux and
>>> Windows existence as it is known today.
>>>
>> most things are *possible*
>>
>>> OS2 had a good run for example. It is still alive and well in some
>>> proprietary industrial equipment because it is impervious to hacks as
>>> no one remembers it.
>>
>> DOS is still used in a lot of industrial kit ...When all you want is
>> to load a single tasked application that talks to custom hardware...
>
> Makes sense. I worked in that sector in the 90's. The only problem with
> MsDOS was the small memory available. We had to use overlays in the
> software, or load/download programs with a stub.
> Well you could in theory kick the processor into 'large' mode and use a
lot more memory

>

--
No Apple devices were knowingly used in the preparation of this post.

Carlos E.R. wrote this copyrighted missive and expects royalties:

> On 2024-09-20 10:41, The Natural Philosopher wrote:
>> On 20/09/2024 01:23, clinker wrote:
>>> Not to derail this thread, which is a good one, but it is possible an
>>> operating system will be developed that threaten both Linux and
>>> Windows existence as it is known today.
>>>
>> most things are *possible*
>>
>>> OS2 had a good run for example. It is still alive and well in some
>>> proprietary industrial equipment because it is impervious to hacks as
>>> no one remembers it.
>>
>> DOS is still used in a lot of industrial kit ...When all you want is to
>> load a single tasked application that talks to custom hardware...
>
> Makes sense. I worked in that sector in the 90's. The only problem with
> MsDOS was the small memory available. We had to use overlays in the
> software, or load/download programs with a stub.

I remember the damned overlays when writing a Turbo C "GUI" (a la ncurses).

I later worked on a large MASM project where you had to load the correct
64k bank when calling a function. Painful.

"640K ought to be enough for anybody." -- attributed to Bill Gates

--
What scoundrel stole the cork from my lunch?
-- J. D. Farley

On Wed, 9/25/2024 3:51 AM, Carlos E.R. wrote:
> On 2024-09-22 20:02, Mark Lloyd wrote:
>> On Sat, 21 Sep 2024 19:34:31 +0100, The Natural Philosopher wrote:
>>
>> [snip]
>>
>>>> Millions of people bought the 8088: the original IBM PC and its clones
>>>> used it. Interestingly, the Amstrad PC used the 8086. And Olivetti, but
>>>> dunno if all models.
>>>>
>>>>
>>> My PC clone had an NEC V20 and a hard drive of a monumental 20MB IIRC.
>>> It was all I could afford
>>
>> I had a V20 too, and (at least where I bought it) I could get a 30MB (RLL)
>> for only $20 more than the 20MB. It was one where you could still do a
>> low-level format. Enter DEBUG and a command like "G=C000:000C". I first
>> got Spinrite because it could change the interleave,
>
> I had an Amstrad PC 1512 DD, ie, dual floppy. I later bought a HD mounted on a card, 32 megs which apparently was the maximum that could be done. Friends of mine had a 10MB HD and complained it was full. So I went for bigger.
>
> Yes, I remember the low-level format thing. The program in my case asked for an interleave factor, and there was a recommendation for 3. I tested several values, and something like 12 got double speed. Amazing at the time. There was also a table of bad sectors to enter manually, but it did not work right because the format found bad sectors and marked them as such in the FAT.
>
> I still have that machine, should still run. Walk.

It's better to leave the unit in the cardboard box :-)

The head on that drive, has probably welded itself
to the surface of the platter. Drives back then
did not have plastic landing ramps, like they do today.

There is nothing wrong with turning it on, but it could
damage or rip the head off it. It doesn't have to end well.

If you turn it on and hear a "strange noise", just put it
back in the box. No one will know.

Paul

On Wed, 25 Sep 2024 09:51:36 +0200, Carlos E.R. wrote:

[snip]

>> I had a V20 too, and (at least where I bought it) I could get a 30MB
>> (RLL)
>> for only $20 more than the 20MB. It was one where you could still do a
>> low-level format. Enter DEBUG and a command like "G=C000:000C". I first
>> got Spinrite because it could change the interleave,
>
> I had an Amstrad PC 1512 DD, ie, dual floppy. I later bought a HD
> mounted on a card, 32 megs which apparently was the maximum that could
> be done. Friends of mine had a 10MB HD and complained it was full. So I
> went for bigger.

32MB was the largest partition size supported by the early version of
FAT16 (16-bit sector count * 512-byte sector size). The disk could be
larger, but then you'd need more partitions.

The drive I had was actually a little larger than the maximum, so I could
create a small second partition.

BTW, there is a later version of FAT16 with a 32-bit sector count.
> Yes, I remember the low-level format thing. The program in my case asked
> for an interleave factor, and there was a recommendation for 3. I tested
> several values, and something like 12 got double speed. Amazing at the
> time. There was also a table of bad sectors to enter manually, but it
> did not work right because the format found bad sectors and marked them
> as such in the FAT.
>
> I still have that machine, should still run. Walk.

--
91 days until the winter celebration (Wednesday, December 25, 2024
12:00:00 AM for 1 day).

Mark Lloyd
http://notstupid.us/

"The Devil...clutched hold of the miserable young man...and flew off
with him through the ceiling, since which time nothing has been heard of
[him]." [Martin Luther]

[snip]

>> I still have that machine, should still run. Walk.
>
> It's better to leave the unit in the cardboard box :-)
>
> The head on that drive, has probably welded itself to the surface of the
> platter. Drives back then did not have plastic landing ramps, like they
> do today.

So you'd rather have it in "somewhat better condition" and completely
useless than take a chance that you can get some use out of it?

[snip]

--
91 days until the winter celebration (Wednesday, December 25, 2024
12:00:00 AM for 1 day).

Mark Lloyd
http://notstupid.us/

"The Devil...clutched hold of the miserable young man...and flew off
with him through the ceiling, since which time nothing has been heard of
[him]." [Martin Luther]

On Wed, 25 Sep 2024 09:57:26 +0200, Carlos E.R. wrote:

> The only problem with MsDOS was the small memory available.

Isn’t that why DOS extenders were invented?

On Wed, 9/25/2024 2:57 PM, Mark Lloyd wrote:
> [snip]
>
>>> I still have that machine, should still run. Walk.
>>
>> It's better to leave the unit in the cardboard box :-)
>>
>> The head on that drive, has probably welded itself to the surface of the
>> platter. Drives back then did not have plastic landing ramps, like they
>> do today.
>
> So you'd rather have it in "somewhat better condition" and completely
> useless than take a chance that you can get some use out of it?
>
> [snip]
>

You need to collect statistical information on the survival rate
of stiction drives from that era. I don't know what the chemistry
possibilities are, for head-to-platter contact for a couple decades.

I have a stiction drive here. It needed a tap on the side of the
drive housing, to release the head and allow the spindle to spin.
You could say, this practice made me <cough> nervous. Because
if you hit it too hard, you could be doing long term damage.
That drive is five feet from me right now, and hasn't been powered
for 25 years. What "chemistry" has been going on in there ? Does
the platter lube suffer breakdown over time ? Dunno. You might not
ever be able to separate the heads from the platter, in a glove box,
without ruining the heads, and needing to replace the head assembly.

Drives with landing ramps, I have no problem with you starting one
of those. I regularly test my old 4GB IDE drive, and it still works
like a champ. I had a 2GB drive fail, a Barracuda 32550N, and the
head lock failed on that drive, destroying the heads and gouging
the platter. They didn't use the head lock idea (solenoid design)
for very long. At a guess then, anything 4GB or more capacity,
is likely perfectly safe to test at your convenience. But a
150MB-250MB drive ? Primed failure material. Already needed to be
whacked to make it work, on a daily basis. Hardly good gambling
odds. Like betting on the pony with three legs, at the track :-)

Paul

On 2024-09-20 10:41, The Natural Philosopher wrote:
>>> DOS is still used in a lot of industrial kit ...When all you want is
>>> to load a single tasked application that talks to custom hardware...

On 25/09/2024 08:57, Carlos E.R. wrote:
>> Makes sense. I worked in that sector in the 90's. The only problem
>> with MsDOS was the small memory available. We had to use overlays in
>> the software, or load/download programs with a stub.
>> Well you could in theory kick the processor into 'large' mode and use a
>> lot more memory

In the mid 1980s, my group was building a small communications
appliance. We built a single-board PC and used a DOS clone as our
bootloader. But booting from floppy was slow as molasses, until I wrote
a small TSR (terminate-and-stay-resident) extension that hooked the BIOS
read call and buffered a track at a time. Simple and efficient.

Later, we built at larger, modular system. We found it more economical
to buy a Taiwanese PC motherboard than to roll our own system package.
We put a DOS Extender to good use (80386 in 80286 protected mode). The
multitasking part of KA9Q was in there somewhere.

On 2024-09-25, Paul <nospam@needed.invalid> wrote:

> I have a stiction drive here. It needed a tap on the side of the
> drive housing, to release the head and allow the spindle to spin.

I had one of those. The housing was pretty solid, so I needed
to give it a whack with a screwdriver handle.

> You could say, this practice made me <cough> nervous. Because
> if you hit it too hard, you could be doing long term damage.

I certainly would consider replacing such a drive, or at least
not storing anything on it that wasn't thoroughly backed up.

One day I visited my wife's office. One of her cow orkers
has having trouble bringing up her Mac. I recognized the
problem as stiction, popped the case open, and tapped the
cover of the hard drive with a fingertip. The cow orker's
initial state of panic was quickly replaced by relief when
the drive spun up. I strongly suggested she take a backup
and consider replacing the drive. I don't know whether she
ever did - people rarely do...

--
/~\ Charlie Gibbs | We'll go down in history as the
\ / <cgibbs@kltpzyxm.invalid> | first society that wouldn't save
X I'm really at ac.dekanfrus | itself because it wasn't cost-
/ \ if you read it the right way. | effective. -- Kurt Vonnegut

On 2024-09-25 20:54, Mark Lloyd wrote:
> On Wed, 25 Sep 2024 09:51:36 +0200, Carlos E.R. wrote:
>
> [snip]
>
>>> I had a V20 too, and (at least where I bought it) I could get a 30MB
>>> (RLL)
>>> for only $20 more than the 20MB. It was one where you could still do a
>>> low-level format. Enter DEBUG and a command like "G=C000:000C". I first
>>> got Spinrite because it could change the interleave,
>>
>> I had an Amstrad PC 1512 DD, ie, dual floppy. I later bought a HD
>> mounted on a card, 32 megs which apparently was the maximum that could
>> be done. Friends of mine had a 10MB HD and complained it was full. So I
>> went for bigger.
>
> 32MB was the largest partition size supported by the early version of
> FAT16 (16-bit sector count * 512-byte sector size). The disk could be
> larger, but then you'd need more partitions.

Which must be the reason, when I applied a compression driver, to use
two partitions.

....

--
Cheers, Carlos.

On 2024-09-25 23:20, Lawrence D'Oliveiro wrote:
> On Wed, 25 Sep 2024 09:57:26 +0200, Carlos E.R. wrote:
>
>> The only problem with MsDOS was the small memory available.
>
> Isn’t that why DOS extenders were invented?

They had to be compatible with the set of libraries you used, and your
application. Most of ours used LabWindows, probably version 1 or 2.
Maybe 3? I don't remember.

I used a stub program that run the menu (and initial data taking), then
offloaded itself and called another exe with the next section.

I don't remember using a dos extender with it. Wikipedia says there was
one in version 2.

https://en.wikipedia.org/wiki/LabWindows/CVI

I remember it used basic or C. I switched to C with Borland C.

--
Cheers, Carlos.

On Thu, 26 Sep 2024 21:17:13 +0200, Carlos E.R. wrote:

> Most of ours used LabWindows, probably version 1 or 2.
> Maybe 3? I don't remember.

All I can recall of LabWindows is that it was a poor attempt to recreate
the Macintosh-only LabView on Microsoft-compatible PCs.

Remember the Mac II and successors had NuBus for their expansion bus at
the time, which offered higher performance for connecting instrumentation
than anything available in the Microsoft-compatible world, until PCI came
along.

On 26/09/2024 20:17, Carlos E.R. wrote:
> On 2024-09-25 23:20, Lawrence D'Oliveiro wrote:
>> On Wed, 25 Sep 2024 09:57:26 +0200, Carlos E.R. wrote:
>>
>>> The only problem with MsDOS was the small memory available.
>>
>> Isn’t that why DOS extenders were invented?
>
> They had to be compatible with the set of libraries you used, and your
> application. Most of ours used LabWindows, probably version 1 or 2.
> Maybe 3? I don't remember.
>
> I used a stub program that run the menu (and initial data taking), then
> offloaded itself and called another exe with the next section.
>
> I don't remember using a dos extender with it. Wikipedia says there was
> one in version 2.
>
> https://en.wikipedia.org/wiki/LabWindows/CVI
>
> I remember it used basic or C. I switched to C with Borland C.
>

Since MSDOS was less an operating system than a program loader, there
was no problem whatsoever in using all the memory the machine had, if
all it was running was your own code talking straight to the Bios or the
hardware.

--
The lifetime of any political organisation is about three years before
its been subverted by the people it tried to warn you about.

Anon.

On Fri, 27 Sep 2024 02:13:56 +0100, The Natural Philosopher wrote:

> Since MSDOS was less an operating system than a program loader, there
> was no problem whatsoever in using all the memory the machine had ...

The problem was the convoluted x86 addressing architecture. And what
happened when you had more than 640 KiB of memory.