hardware abstraction

[muta...@gmail.com]

Scott pointed out that the IBM PC is the exception for
using a BIOS.

But the existence of the BIOS allows me to boot from a
USB stick as a hard disk. That is fantastic.

There is a hardware device that converts RS232 into
Wifi.

But surely the proper solution is to put not just this but
also the printer into the BIOS and abstract the hardware?

BFN. Paul.

[wolfgang kern]

On 12/12/2021 10:35, muta...@gmail.com wrote:
> Scott pointed out that the IBM PC is the exception for
> using a BIOS.
>
> But the existence of the BIOS allows me to boot from a
> USB stick as a hard disk. That is fantastic.
>
> There is a hardware device that converts RS232 into
> Wifi.

interesting, but I don't see any RS232 connector on my 12 PCs.

> But surely the proper solution is to put not just this but
> also the printer into the BIOS and abstract the hardware?

while some printers may work wireless, which brand and type would you
insert into your BIOS ?

My OS supports only a few printers from HP (PCL3 and part of PCL4)
either Centronics or USB connected.
and my oldest(1980) OS-variant supported LPT in 8255 mode.
__
wolfgang

[Grant Taylor]

On 12/12/21 2:35 AM, muta...@gmail.com wrote:
> There is a hardware device that converts RS232 into Wifi.

Sort of, but not really.

There is a device, WiFi-232, that speaks to the host computer using
RS-232. But it doesn't /convert/ the RS-232 to WiFi.

You could easily do the same with any RS-232 (or RS-422 / RS-489 with
proper converter) client device and a null modem cable to any other WiFi
connected computer.

The WiFi-232 just does it in a convenient package while pretending to be
a standard Hays compatible modem to the host that it's cabled to.

> But surely the proper solution is to put not just this but also the
> printer into the BIOS and abstract the hardware?

I get the impression that you have never written data to an LPT / COM
port directly while it was connected to a printer that understood ASCII
text.

Printer drivers come into play when you want to do fancier things than
ASCII text. Think of printer drivers as a software converter that
converts from one thing to another thing that the printer understands.
They are also printer (family) specific.

Also, given the recent Print-Nightmare that Microsoft has been having,
do you /really/ want that in the BIOS (er firmware)? I don't.



--
Grant. . . .
unix || die

[muta...@gmail.com]

On Monday, December 13, 2021 at 3:48:42 AM UTC+11, wolfgang kern wrote:

> interesting, but I don't see any RS232 connector on my 12 PCs.

That's the exact problem. They took away the physical
connector without updating the firmware to redirect it
to something else (USB/Bluetooth/Wifi).

They did allow hard disks to be redirected to the USB port.

Anyway, I'm pretty sure you can buy PCs with a serial port,
which is the donkey solution I will probably have to resort
to if no BIOS manufacturer does this internally.

> > But surely the proper solution is to put not just this but
> > also the printer into the BIOS and abstract the hardware?

> while some printers may work wireless, which brand and type would you
> insert into your BIOS ?

I seem to be missing something (Grant asked something
similar). When you do an fopen of PRN, there is no data
sent to the printer, it's just a raw connection. That's all I
expect the BIOS to do. You still need some sort of driver
to send the right format of data via the PRN connection.

BFN. Paul.

[Grant Taylor]

On 12/12/21 1:58 PM, muta...@gmail.com wrote:
> That's the exact problem. They took away the physical connector
> without updating the firmware to redirect it to something else
> (USB/Bluetooth/Wifi).

They added a new physical connector that was in demand. They removed
the old physical connector that had multiple orders of magnitude less
demand to make room for the new connector.

> They did allow hard disks to be redirected to the USB port.

No.

USB has something that looks like a hard disk. But it's decidedly
different than a hard disk attached to an IDE / SCSI controller.

USB /also/ has something that looks like a serial port.

Neither of the USB counterparts work as well or have the same properties
as the non-USB things.

> Anyway, I'm pretty sure you can buy PCs with a serial port, which
> is the donkey solution I will probably have to resort to if no BIOS
> manufacturer does this internally.

You can buy PCIe add-on cards that have serial ports on them. Honest to
$DEITY serial ports, not emulated things.

[Rod Pemberton]

1980s - not enough processing power
2010s - why not ...


In the 1980s, you needed an additional micro-processor just to add a
printer buffer, or even to control disk drives, e.g., CBM 1541. There
simply wasn't enough processing capacity. The Amiga used a few
coprocessors to perform what a processor can do by itself nowadays.
WinModems (early 2000's) were an example of when x86 PCs became fast
enough for the x86 processor to take on additional tasks.

In the late 1990s, they used a Forth-based interpreter called
OpenFirmware as a BIOS replacement for many non-AT class computers.
So, if they can add a Forth-based interpreter, why not a PostScript
interpreter for a printer? ... So, sure, why not.


--

[Rod Pemberton]

On Sun, 12 Dec 2021 14:32:15 -0700
Grant Taylor <gta...@tnetconsulting.net> wrote:

> On 12/12/21 1:58 PM, muta...@gmail.com wrote:

> > They did allow hard disks to be redirected to the USB port.
>
> No.
>
> USB has something that looks like a hard disk. But it's decidedly
> different than a hard disk attached to an IDE / SCSI controller.

Are you talking about a USB storage device, such as a USB stick or a
USB hard disk?

Or, are you talking about legacy BIOS drive emulation which is a
combination of BIOS support, a USB storage device, and a bootable image
stored on the USB device? i.e., the USB-FDD, USB-HD, USB-ZIP etc boot
options on the BBS Boot Menu

I think Paul is referring to the latter when he said, "they [allowed]

hard disks to be redirected to the USB port."

FYI, the BIOS Boot Specification is what requires the bootable device
type to be reported as either usb, floppy, hard disk, cd-rom, pcmcia,
or network. This standard defines how execution is transferred from
the BIOS (non-UEFI) to the boot code, either a procedure for IPL
devices, or installable Int 13h drive support routines for BCV devices.
BCV devices have an option ROM which has extra Int 13h routines, e.g.,
SCSI. This spec also specifies the BIOS Boot Menu pop-up etc as well as
various changes to Int 18h, Int 19h, where Int 13h gets copied to Int
40h, the BDA updated with installed drives, and so forth.


--

[wolfgang kern]

with PRN you probably mean LPT...
there once were a BIOS function for this old standard,
IIRC it was INT_0x14. But all OS including early DOS used the hardware
direct because in old days LPT- and COM-ports were PC/XT&AT standards.
meanwhile serial and parallel connectors disappeared almost completely.
__
wolfgang

[Scott Lurndal]

Rod Pemberton <noe...@basdxcqvbe.com> writes:
>On Sun, 12 Dec 2021 01:35:09 -0800 (PST)
>"muta...@gmail.com" <muta...@gmail.com> wrote:
>
>> Scott pointed out that the IBM PC is the exception for
>> using a BIOS.
>>
>> But the existence of the BIOS allows me to boot from a
>> USB stick as a hard disk. That is fantastic.
>>
>> There is a hardware device that converts RS232 into
>> Wifi.
>>
>> But surely the proper solution is to put not just this but
>> also the printer into the BIOS and abstract the hardware?
>>
>
>1980s - not enough processing power
>2010s - why not ...

Because BIOS, as Basic I/O System, has been obsolete as a concept
for decades - primarily due to the need to update the BIOS to
support every single new device and the inability to extend the
interface in any rational way without burning a new PROM (or
in modern days, updating a flash chip).

To a certain extent, standards such as AHCI and XHCI have created
a common base for a subset of hardware devices, but there is no
equivalent for networking or printers.

[James Harris]

On 13/12/2021 15:32, Scott Lurndal wrote:
> Rod Pemberton <noe...@basdxcqvbe.com> writes:
>> On Sun, 12 Dec 2021 01:35:09 -0800 (PST)
>> "muta...@gmail.com" <muta...@gmail.com> wrote:

...

>>> But surely the proper solution is to put not just this but
>>> also the printer into the BIOS and abstract the hardware?
>>>
>>
>> 1980s - not enough processing power
>> 2010s - why not ...
>
> Because BIOS, as Basic I/O System, has been obsolete as a concept
> for decades

Hardly.

> - primarily due to the need to update the BIOS to
> support every single new device and the inability to extend the
> interface in any rational way without burning a new PROM (or
> in modern days, updating a flash chip).

BIOSes don't need to be updated to support new devices. Instead, it has
been common for add-in cards to come with ROMs which hook their code
into the BIOS at boot time.

>
> To a certain extent, standards such as AHCI and XHCI have created
> a common base for a subset of hardware devices, but there is no
> equivalent for networking or printers.
>

AFAIK there are no BIOS calls for the network.


--
James Harris

[Scott Lurndal]

James Harris <james.h...@gmail.com> writes:
>On 13/12/2021 15:32, Scott Lurndal wrote:
>> Rod Pemberton <noe...@basdxcqvbe.com> writes:
>>> On Sun, 12 Dec 2021 01:35:09 -0800 (PST)
>>> "muta...@gmail.com" <muta...@gmail.com> wrote:
>
>...
>
>>>> But surely the proper solution is to put not just this but
>>>> also the printer into the BIOS and abstract the hardware?
>>>>
>>>
>>> 1980s - not enough processing power
>>> 2010s - why not ...
>>
>> Because BIOS, as Basic I/O System, has been obsolete as a concept
>> for decades
>
>Hardly.

Certainly. No modern operating system uses the BIOS for I/O. And most
current generation machines don't even have a classic BIOS.

>
>> - primarily due to the need to update the BIOS to
>> support every single new device and the inability to extend the
>> interface in any rational way without burning a new PROM (or
>> in modern days, updating a flash chip).
>
>BIOSes don't need to be updated to support new devices. Instead, it has
>been common for add-in cards to come with ROMs which hook their code
>into the BIOS at boot time.

The ROM includes just enough code to enable booting via the card. It is
very much not a "general purpose I/O driver" for the card. And I haven't
seen a modern plug-in PCI express card with an Option ROM for almost a decade.

>
>>
>> To a certain extent, standards such as AHCI and XHCI have created
>> a common base for a subset of hardware devices, but there is no
>> equivalent for networking or printers.
>>
>
>AFAIK there are no BIOS calls for the network.

PXE.

[muta...@gmail.com]

On Monday, December 13, 2021 at 5:40:51 AM UTC+11, Grant Taylor wrote:

> The WiFi-232 just does it in a convenient package while pretending to be
> a standard Hays compatible modem to the host that it's cabled to.

I'm planning on moving to the Philippines soon, so I will
buy a new desktop computer, with these requirements:

1. Legacy BIOS
2. Boot from USB stick
3. Two COM ports
4. Windows 10/11

And then I'll try to source that WiFi-232. When I did a google
search I seemed to find kits rather than something
consumer-friendly.

Note that I'm planning on setting up a so-called "sneakernet"
using the Fidonet software that I already have running under
PDOS/386 as proof of concept (tobruk and msged).

I'm planning on retreating behind my BBS and being a
"techmonk" - instead of sitting on a mountain going
"ummmmm" I'm going to answer C questions.

BFN. Paul.

[James Harris]

On 13/12/2021 18:04, Scott Lurndal wrote:
> James Harris <james.h...@gmail.com> writes:
>> On 13/12/2021 15:32, Scott Lurndal wrote:

...

>>> To a certain extent, standards such as AHCI and XHCI have created
>>> a common base for a subset of hardware devices, but there is no
>>> equivalent for networking or printers.
>>>
>>
>> AFAIK there are no BIOS calls for the network.
>
> PXE.
>

AFAIK that's only for booting via the network and if one boots from hard
disk then the PXE UNDI drivers will not be available (unfortunately).


--
James Harris