Re: [9fans] punched cards live

Re: [9fans] punched cards live

[David Leimbach]

[-- Attachment #1: Type: text/plain, Size: 2179 bytes --]

On Tue, Nov 4, 2008 at 7:25 AM, Eris Discordia <eris.discordia@gmail.com>wrote:

> i think it's a tradition at this point to use 0x20 and not 0x00 to
>> fill a fixed-with signature.  ata identify device uses 0x20 to fill
>> out fixed-width fields like the serial number.  i'd be interested
>> where this tradition popped up.  0 would make more sense.
>>
>
> I risk being wrong--as always--and say it must have popped up in a normal
> ASCII environment. 0x20 = 32, the ASCII code point for a simple whitespace.
> BIOS routines know how to display a whitespace, or any ASCII character, in
> text mode. I remember somewhere back in time I could load AL with an ASCII
> character, call interrupt 0x0A service 0x0E, and have the character printed
> on the screen and the cursor moved one character to the right. This was
> (is?) fairly standard and time-proven. And it worked (works?) everywhere, at
> least in the PC world.


DOS string routines used $ character termination. (AX = 09, DX=(address of $
terminated string) INT 21h, if IIRC).

The print routines in the BIOS I knew of took a length parameter in the CX
register (also IIRC)

Why do I sometimes still yearn for the simplicity of DOS?  Maybe it's Vista
that makes me feel so.  Maybe it was the amount of stuff we could do with so
very little RAM back then.  Running protected mode servers really wasn't all
THAT bad. :-)

Perhaps I'm just getting old.



>
>
> --On Monday, November 03, 2008 7:06 AM -0500 erik quanstrom <
> quanstro@quanstro.net> wrote:
>
>  This courtesy of the ACPI spec: ""RSD PTR " (Notice that this
>>> signature must contain a trailing
>>> blank character.)"
>>>
>>> So where do we get the guys who design this stuff? Can we send them
>>> back? Or put them in an infinite loop in a time machine (oh wait see
>>> the subject).
>>>
>>
>> i think it's a tradition at this point to use 0x20 and not 0x00 to
>> fill a fixed-with signature.  ata identify device uses 0x20 to fill
>> out fixed-width fields like the serial number.  i'd be interested
>> where this tradition popped up.  0 would make more sense.
>>
>> - erik
>>
>>
>>
>
>
>
>
>

[-- Attachment #2: Type: text/html, Size: 3167 bytes --]

Re: [9fans] punched cards live

[Eris Discordia]

> The print routines in the BIOS I knew of took a length parameter in the
> CX register (also IIRC)

These were string routines. Service 0x0E of interrupt 0x0A (now that I
think better perhaps it wasn't 10 = 0x0A, rather 0x10 = 16) provided
character output.

> Running protected mode servers really wasn't all THAT bad. :-)

Did you ever use a DOS "extender" like DOS/4GW? I didn't but whenever I saw
its startup message I knew the program I had run was cool... and prone to
crashes.

--On Tuesday, November 04, 2008 8:25 AM -0800 David Leimbach
<leimy2k@gmail.com> wrote:

>
>
>
> On Tue, Nov 4, 2008 at 7:25 AM, Eris Discordia <eris.discordia@gmail.com>
> wrote:
>
>
> i think it's a tradition at this point to use 0x20 and not 0x00 to
> fill a fixed-with signature.  ata identify device uses 0x20 to fill
> out fixed-width fields like the serial number.  i'd be interested
> where this tradition popped up.  0 would make more sense.
>
>
> I risk being wrong--as always--and say it must have popped up in a normal
> ASCII environment. 0x20 = 32, the ASCII code point for a simple
> whitespace. BIOS routines know how to display a whitespace, or any ASCII
> character, in text mode. I remember somewhere back in time I could load
> AL with an ASCII character, call interrupt 0x0A service 0x0E, and have
> the character printed on the screen and the cursor moved one character to
> the right. This was (is?) fairly standard and time-proven. And it worked
> (works?) everywhere, at least in the PC world.
>
>
>
>
> DOS string routines used $ character termination. (AX = 09, DX=(address
> of $ terminated string) INT 21h, if IIRC).
>
>
> The print routines in the BIOS I knew of took a length parameter in the
> CX register (also IIRC)
>
>
> Why do I sometimes still yearn for the simplicity of DOS?  Maybe it's
> Vista that makes me feel so.  Maybe it was the amount of stuff we could
> do with so very little RAM back then.  Running protected mode servers
> really wasn't all THAT bad. :-)
>
>
> Perhaps I'm just getting old.
>
>
>
>
>
>
> --On Monday, November 03, 2008 7:06 AM -0500 erik quanstrom
> <quanstro@quanstro.net> wrote:
>
>
>
> This courtesy of the ACPI spec: ""RSD PTR " (Notice that this
> signature must contain a trailing
> blank character.)"
>
> So where do we get the guys who design this stuff? Can we send them
> back? Or put them in an infinite loop in a time machine (oh wait see
> the subject).
>
>
> i think it's a tradition at this point to use 0x20 and not 0x00 to
> fill a fixed-with signature.  ata identify device uses 0x20 to fill
> out fixed-width fields like the serial number.  i'd be interested
> where this tradition popped up.  0 would make more sense.
>
> - erik
>
>
>
>
>
>
>
>
>
>

Re: [9fans] punched cards live

[Eris Discordia]

> i think it's a tradition at this point to use 0x20 and not 0x00 to
> fill a fixed-with signature.  ata identify device uses 0x20 to fill
> out fixed-width fields like the serial number.  i'd be interested
> where this tradition popped up.  0 would make more sense.

I risk being wrong--as always--and say it must have popped up in a normal
ASCII environment. 0x20 = 32, the ASCII code point for a simple whitespace.
BIOS routines know how to display a whitespace, or any ASCII character, in
text mode. I remember somewhere back in time I could load AL with an ASCII
character, call interrupt 0x0A service 0x0E, and have the character printed
on the screen and the cursor moved one character to the right. This was
(is?) fairly standard and time-proven. And it worked (works?) everywhere,
at least in the PC world.

--On Monday, November 03, 2008 7:06 AM -0500 erik quanstrom
<quanstro@quanstro.net> wrote:

>> This courtesy of the ACPI spec: ""RSD PTR " (Notice that this
>> signature must contain a trailing
>> blank character.)"
>>
>> So where do we get the guys who design this stuff? Can we send them
>> back? Or put them in an infinite loop in a time machine (oh wait see
>> the subject).
>
> i think it's a tradition at this point to use 0x20 and not 0x00 to
> fill a fixed-with signature.  ata identify device uses 0x20 to fill
> out fixed-width fields like the serial number.  i'd be interested
> where this tradition popped up.  0 would make more sense.
>
> - erik
>
>

Re: [9fans] punched cards live

[erik quanstrom]

> It's the ACPI Secure Computing Initiative:  fixed input format == no
> buffer overflow vulnerabilities.  Long live Herman Hollerith!
>

if bios wants to own the os, using buffer overflow in acpi
seems like more effort than necessary.

- erik

[9fans] punched cards live

[ron minnich]

This courtesy of the ACPI spec: ""RSD PTR " (Notice that this
signature must contain a trailing
blank character.)"

So where do we get the guys who design this stuff? Can we send them
back? Or put them in an infinite loop in a time machine (oh wait see
the subject).

ron

Re: [9fans] punched cards live

[erik quanstrom]

> This courtesy of the ACPI spec: ""RSD PTR " (Notice that this
> signature must contain a trailing
> blank character.)"
>
> So where do we get the guys who design this stuff? Can we send them
> back? Or put them in an infinite loop in a time machine (oh wait see
> the subject).

i think it's a tradition at this point to use 0x20 and not 0x00 to
fill a fixed-with signature.  ata identify device uses 0x20 to fill
out fixed-width fields like the serial number.  i'd be interested
where this tradition popped up.  0 would make more sense.

- erik

Re: [9fans] punched cards live

[Brantley Coile]

As the subject says, punched cards.  Except it was a '40'X on the 360.

On Nov 3, 2008, at 7:06 AM, erik quanstrom wrote:

>> This courtesy of the ACPI spec: ""RSD PTR " (Notice that this
>> signature must contain a trailing
>> blank character.)"
>>
>> So where do we get the guys who design this stuff? Can we send them
>> back? Or put them in an infinite loop in a time machine (oh wait see
>> the subject).
>
> i think it's a tradition at this point to use 0x20 and not 0x00 to
> fill a fixed-with signature.  ata identify device uses 0x20 to fill
> out fixed-width fields like the serial number.  i'd be interested
> where this tradition popped up.  0 would make more sense.
>
> - erik
>
>
>

Re: [9fans] punched cards live

[John DeGood]

It's the ACPI Secure Computing Initiative:  fixed input format == no
buffer overflow vulnerabilities.  Long live Herman Hollerith!

ron minnich wrote:
> This courtesy of the ACPI spec: ""RSD PTR " (Notice that this
> signature must contain a trailing
> blank character.)"
>
> So where do we get the guys who design this stuff? Can we send them
> back? Or put them in an infinite loop in a time machine (oh wait see
> the subject).
>
> ron
>