Greetings all,
I see that in the very latest Configure.help version, 2.76, available at http://www.tuxedo.org/~esr/cml2/
Eric has decided to follow the following standard:
IEC 60027-2, Second edition, 2000-11, Letter symbols to be used in electrical technology - Part 2: Telecommunications and electronics.
and has changed all the abbreviations for Kilobyte (KB) to KiB, Megabyte (MB) to MiB, etc, etc.
Now, granted that this is the "standard", should there be some discussion related to this
change, or is everyone comfortable with this? It certainly made me do a double take.
Here is a snippet from the diff between versions 2.75 and 2.76 of Configure.help:
@@ -344,8 +344,8 @@
If you are compiling a kernel which will never run on a machine with
more than 960 megabytes of total physical RAM, answer "off" here
(default choice and suitable for most users). This will result in a
- "3GB/1GB" split: 3GB are mapped so that each process sees a 3GB
- virtual memory space and the remaining part of the 4GB virtual memory
+ "3GiB/1GiB" split: 3GiB are mapped so that each process sees a 3GiB
+ virtual memory space and the remaining part of the 4GiB virtual memory
space is used by the kernel to permanently map as much physical memory
as possible.
Steven
I believe that the main purpose of documentation, help etc is to get the
information across in a way that is most easily understood, ie that
minimises the number of support questions.. ..and everyone surely knows
what GB, MB and KB stand for. So let's leave it at that. Where's the "i"
in "megabyte" ? Or is 1MiB 1000000 bytes, rather than 1048576?
It's confusing enough with the 10 "Mb" networking / 1.44 "MB" floppy
distinction already..
At 11:02 -0700 Steven Cole wrote:
>Now, granted that this is the "standard", should there be some discussion related to this
>change, or is everyone comfortable with this? It certainly made me do a double take.
>
>Here is a snippet from the diff between versions 2.75 and 2.76 of Configure.help:
>
>@@ -344,8 +344,8 @@
> If you are compiling a kernel which will never run on a machine with
> more than 960 megabytes of total physical RAM, answer "off" here
> (default choice and suitable for most users). This will result in a
>- "3GB/1GB" split: 3GB are mapped so that each process sees a 3GB
>- virtual memory space and the remaining part of the 4GB virtual memory
>+ "3GiB/1GiB" split: 3GiB are mapped so that each process sees a 3GiB
>+ virtual memory space and the remaining part of the 4GiB virtual memory
> space is used by the kernel to permanently map as much physical memory
> as possible.
On Thu, 2001-12-20 at 13:02, Steven Cole wrote:
> I see that in the very latest Configure.help version, 2.76, available at
> http://www.tuxedo.org/~esr/cml2/ Eric has decided to follow the
> following standard: IEC 60027-2, Second edition, 2000-11, Letter symbols
> to be used in electrical technology - Part 2: Telecommunications and
> electronics. and has changed all the abbreviations for Kilobyte (KB) to
> KiB, Megabyte (MB) to MiB, etc, etc.
Personally? This is hideous, confusing, and unneeded.
GiB in my mind is a gigabit, anyhow.
Robert Love
On Thu, Dec 20, 2001 at 06:16:24PM +0000, Matt Bernstein wrote:
> I believe that the main purpose of documentation, help etc is to get the
> information across in a way that is most easily understood, ie that
> minimises the number of support questions.. ..and everyone surely knows
> what GB, MB and KB stand for. So let's leave it at that. Where's the "i"
> in "megabyte" ? Or is 1MiB 1000000 bytes, rather than 1048576?
>
> It's confusing enough with the 10 "Mb" networking / 1.44 "MB" floppy
> distinction already..
I'm afraid I have to disagree with you. Not enough people
know that KB stands for 1000B. I know through experience that
KB is often used for 1024B. The introduction of KiB makes it
clear that 1024B is intended.
--
Randolph Bentson
[email protected]
Hi!
You are close - he uses "MiB" as short for "mebi" - Mega-binary.
Don't laugh - this is official! It's exactly for what you said:
What is 1 MB?
1.000.000 Byte
or
1.048.576 Byte
For a short reading I recommend this:
http://physics.nist.gov/cuu/Units/binary.html
ciao
Michael
Matt Bernstein wrote:
> I believe that the main purpose of documentation, help etc is to get the
> information across in a way that is most easily understood, ie that
> minimises the number of support questions.. ..and everyone surely knows
> what GB, MB and KB stand for. So let's leave it at that. Where's the "i"
> in "megabyte" ? Or is 1MiB 1000000 bytes, rather than 1048576?
>
> It's confusing enough with the 10 "Mb" networking / 1.44 "MB" floppy
> distinction already..
>
> At 11:02 -0700 Steven Cole wrote:
>
>
>>Now, granted that this is the "standard", should there be some
discussion related to this
>>change, or is everyone comfortable with this? It certainly made me
do a double take.
>>
>>Here is a snippet from the diff between versions 2.75 and 2.76 of
Configure.help:
>>
>>@@ -344,8 +344,8 @@
>> If you are compiling a kernel which will never run on a machine with
>> more than 960 megabytes of total physical RAM, answer "off" here
>> (default choice and suitable for most users). This will result in a
>>- "3GB/1GB" split: 3GB are mapped so that each process sees a 3GB
>>- virtual memory space and the remaining part of the 4GB virtual memory
>>+ "3GiB/1GiB" split: 3GiB are mapped so that each process sees a 3GiB
>>+ virtual memory space and the remaining part of the 4GiB virtual
memory
>> space is used by the kernel to permanently map as much physical memory
>> as possible.
>>
>
> -
> To unsubscribe from this list: send the line "unsubscribe
linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
>
On Thu, 20 Dec 2001, Randolph Bentson wrote:
> I'm afraid I have to disagree with you. Not enough people
> know that KB stands for 1000B. I know through experience that
> KB is often used for 1024B. The introduction of KiB makes it
> clear that 1024B is intended.
Given the confusion its raised in this thread already, this strikes
me as a particularly bad change. Some of those definitions in
Configure.help have been there for the better part of 10 years
now, and I don't recall anyone in the past few years needing to
ask for clarification.
Is this really that important ?
Dave.
--
| Dave Jones. http://www.codemonkey.org.uk
| SuSE Labs
On Thursday 20 December 2001 12:47, Michael Dunsky wrote:
> Hi!
>
> You are close - he uses "MiB" as short for "mebi" - Mega-binary.
> Don't laugh - this is official! It's exactly for what you said:
>
> What is 1 MB?
> 1.000.000 Byte
> or
> 1.048.576 Byte
>
>
> For a short reading I recommend this:
>
> http://physics.nist.gov/cuu/Units/binary.html
>
>
> ciao
Tebi or not Tebi? That is the question. Mebi this is a good idea,
but I doubt it. How many understand people (Americans at least)
understand SI units before this change?? I say, gibe a break!
--
[email protected].
Steven Cole <[email protected]>:
> I see that in the very latest Configure.help version, 2.76,
> available at http://www.tuxedo.org/~esr/cml2/ Eric has decided to
> follow the following standard: IEC 60027-2, Second edition, 2000-11,
> Letter symbols to be used in electrical technology - Part 2:
> Telecommunications and electronics. and has changed all the
> abbreviations for Kilobyte (KB) to KiB, Megabyte (MB) to MiB, etc,
> etc.
This change came as a patch from David Woodhouse. I think the new
abbreviations are awful ugly, myself, but they do have the virtue of
not being ambiguous. So I swallowed hard and took the patch.
--
<a href="http://www.tuxedo.org/~esr/">Eric S. Raymond</a>
The right to buy weapons is the right to be free.
-- A.E. Van Vogt, "The Weapon Shops Of Isher", ASF December 1942
Hi all :))
>You are close - he uses "MiB" as short for "mebi" - Mega-binary.
Personally I don't like very much the abbreviations, but I must
recognize that they remove all possible ambiguity for the
Configure.help. With MiB, GiB, etc... you're completely sure that you
are talking about 2^20, 2^30 and not 10^6, 10^9, etc...
So I think that is a good idea in general to use that
abbreviations for the binary units. Moreover, it's official and
correct use. Eric made a sensible decision here.
Ra?l
On Thursday 20 December 2001 11:00 am, Dave Jones wrote:
> On Thu, 20 Dec 2001, Randolph Bentson wrote:
> > I'm afraid I have to disagree with you. Not enough people
> > know that KB stands for 1000B. I know through experience that
> > KB is often used for 1024B. The introduction of KiB makes it
> > clear that 1024B is intended.
>
> Given the confusion its raised in this thread already, this strikes
> me as a particularly bad change. Some of those definitions in
> Configure.help have been there for the better part of 10 years
My *mind* has had those definitions hard-coded for the better part of 8
years. This is only going to serve to confuse newcomers (and for that
matter, anyone who doesn't read this thread.)
to: linux-kernel
from: Joe Newbie[1-areallyfrigginbignumber]
date: mid-2.5 continuing well beyond any foreseeable date.
message: "What's a GiB? And why are the Men in Black taking over the
kernel?"
I leave you with that thought.
On Thursday 20 December 2001 11:52 am, Eric S. Raymond wrote:
> Steven Cole <[email protected]>:
> > I see that in the very latest Configure.help version, 2.76,
> > available at http://www.tuxedo.org/~esr/cml2/ Eric has decided to
> > follow the following standard: IEC 60027-2, Second edition, 2000-11,
> > Letter symbols to be used in electrical technology - Part 2:
> > Telecommunications and electronics. and has changed all the
> > abbreviations for Kilobyte (KB) to KiB, Megabyte (MB) to MiB, etc,
> > etc.
>
> This change came as a patch from David Woodhouse. I think the new
> abbreviations are awful ugly, myself, but they do have the virtue of
> not being ambiguous. So I swallowed hard and took the patch.
I personally agree with Matt Bernstein who wrote:
>I believe that the main purpose of documentation, help etc is to get the
>information across in a way that is most easily understood, ie that
>minimises the number of support questions.
I've obviously been wrong all along on this, but I had always assumed
(incorrectly) that mega when applied to binary things really meant
2 to the 20th, and that the confusion came from disk manufacturers who
chose whatever definition made their products seem bigger.
So, if we really want to be pedantically correct, let's all swallow
_really_ hard, and go all the way on this. Or, are the UGLY changes
below actually technically wrong?
Cheers, and happy holidays,
Steven
--- Configure.help.2.76 Thu Dec 20 12:00:46 2001
+++ Configure.help Thu Dec 20 12:08:13 2001
@@ -2,7 +2,8 @@
# Eric S. Raymond <mailto:[email protected]>
# Steven Cole <mailto:[email protected]>
#
-# Merged version 2.76: current with 2.4.17-rc2/2.5.1.
+# Merged version 2.77: current with 2.4.17-rc2/2.5.1.
+# Warning: A barf bag is a recommended accessory this version.
#
# This version of the Linux kernel configuration help texts
# corresponds to kernel versions 2.4.x and 2.5.x.
@@ -334,25 +335,25 @@
# Choice: himem
High Memory support
CONFIG_NOHIGHMEM
- Linux can use up to 64 Gigabytes of physical memory on x86 systems.
+ Linux can use up to 64 Gibibytes of physical memory on x86 systems.
However, the address space of 32-bit x86 processors is only 4
- Gigabytes large. That means that, if you have a large amount of
+ Gibibytes large. That means that, if you have a large amount of
physical memory, not all of it can be "permanently mapped" by the
kernel. The physical memory that's not permanently mapped is called
"high memory".
If you are compiling a kernel which will never run on a machine with
- more than 960 megabytes of total physical RAM, answer "off" here
+ more than 960 mebibytes of total physical RAM, answer "off" here
(default choice and suitable for most users). This will result in a
"3GiB/1GiB" split: 3GiB are mapped so that each process sees a 3GiB
virtual memory space and the remaining part of the 4GiB virtual memory
space is used by the kernel to permanently map as much physical memory
as possible.
- If the machine has between 1 and 4 Gigabytes physical RAM, then
+ If the machine has between 1 and 4 Gibibytes physical RAM, then
answer "4GB" here.
- If more than 4 Gigabytes is used then answer "64GB" here. This
+ If more than 4 Gibibytes is used then answer "64GB" here. This
selection turns Intel PAE (Physical Address Extension) mode on.
PAE implements 3-level paging on IA32 processors. PAE is fully
supported by Linux, PAE mode is implemented on all recent Intel
@@ -370,12 +371,12 @@
4GB
CONFIG_HIGHMEM4G
Select this if you have a 32-bit processor and between 1 and 4
- gigabytes of physical RAM.
+ gibibytes of physical RAM.
64GB
CONFIG_HIGHMEM64G
Select this if you have a 32-bit processor and more than 4
- gigabytes of physical RAM.
+ gibibytes of physical RAM.
Normal floppy disk support
CONFIG_BLK_DEV_FD
@@ -13509,7 +13510,7 @@
Enable kernel debugging symbols
CONFIG_DEBUGSYM
When this is enabled, the User-Mode Linux binary will include
- debugging symbols. This enlarges the binary by a few megabytes,
+ debugging symbols. This enlarges the binary by a few mebibytes,
but aids in tracking down kernel problems in UML. It is required
if you intend to do any kernel development.
@@ -18296,7 +18297,7 @@
MSND buffer size (KiB)
CONFIG_MSND_FIFOSIZE
- Configures the size of each audio buffer, in kilobytes, for
+ Configures the size of each audio buffer, in kibibytes, for
recording and playing in the MultiSound drivers (both the Classic
and Pinnacle). Larger values reduce the chance of data overruns at
the expense of overall latency. If unsure, use the default.
@@ -18791,7 +18792,7 @@
CONFIG_REMOTE_DEBUG
If you say Y here, it will be possible to remotely debug the MIPS
kernel using gdb. This enlarges your kernel image disk size by
- several megabytes and requires a machine with more than 16 MiB,
+ several mebibytes and requires a machine with more than 16 MiB,
better 32 MiB RAM to avoid excessive linking time. This is only
useful for kernel hackers. If unsure, say N.
@@ -24244,7 +24245,7 @@
kernel using gdb, if you have the gdb-sh-stub package from
http://www.m17n.org (or any conforming standard LinuxSH BIOS) in FLASH or
EPROM. This enlarges your kernel image disk size by several
- megabytes but allows you to load, run and debug the kernel image
+ mebibytes but allows you to load, run and debug the kernel image
remotely using gdb. This is only useful for kernel hackers. If
unsure, say N.
Steven Cole <[email protected]>:
> I've obviously been wrong all along on this, but I had always assumed
> (incorrectly) that mega when applied to binary things really meant
> 2 to the 20th, and that the confusion came from disk manufacturers who
> chose whatever definition made their products seem bigger.
>
> So, if we really want to be pedantically correct, let's all swallow
> _really_ hard, and go all the way on this. Or, are the UGLY changes
> below actually technically wrong?
Alas, I don't know.
--
<a href="http://www.tuxedo.org/~esr/">Eric S. Raymond</a>
No kingdom can be secured otherwise than by arming the people. The possession
of arms is the distinction between a freeman and a slave.
-- "Political Disquisitions", a British republican tract of 1774-1775
In article <[email protected]> you wrote:
> Eric has decided to follow the following standard:
> IEC 60027-2, Second edition, 2000-11,
> Letter symbols to be used in electrical technology -
> Part 2: Telecommunications and electronics.
> and has changed all the abbreviations for Kilobyte (KB) to KiB,
> Megabyte (MB) to MiB, etc, etc.
I did this for nettools (i.e. ifconfig), too:
RX bytes:2120660294 (1.9 GiB) TX bytes:341183013 (325.3 MiB)
man page:
Since net-tools 1.60-4 ifconfig is printing byte counters
with SI units. So 1 KiB are 2^10 byte. Note, the numbers
are truncated to one decimal (which can by quite a large
error if you consider 0.1 PiB is 112.589.990.684.262 bytes :)
...
SEE ALSO
route(8), netstat(8), arp(8), rarp(8), ipchains(8)
http://physics.nist.gov/cuu/Units/binary.html - Prefixes
for binary multiples
On Thu, Dec 20, 2001 at 01:52:13PM -0500, Eric S. Raymond wrote:
> Steven Cole <[email protected]>:
> > I see that in the very latest Configure.help version, 2.76,
> > available at http://www.tuxedo.org/~esr/cml2/ Eric has decided to
> > follow the following standard: IEC 60027-2, Second edition, 2000-11,
> > Letter symbols to be used in electrical technology - Part 2:
> > Telecommunications and electronics. and has changed all the
> > abbreviations for Kilobyte (KB) to KiB, Megabyte (MB) to MiB, etc,
> > etc.
What? AFAIK 'K' means 1000 in SI. However since computers use binary
numbers, the number (2^n) which was the most closer to 1000 was selected to
be used as 'K' for indicating information amount, where n=10. [for decimal
numbers 10^n (n=3) is used for 'K']. And so on with 'M', 'G' ... Sorry if
I was wrong about this ...
- Gabor
* Steven Cole schrieb am 20.12.01 um 20:32 Uhr:
> - If the machine has between 1 and 4 Gigabytes physical RAM, then
> + If the machine has between 1 and 4 Gibibytes physical RAM, then
> answer "4GB" here.
^^^
this has to be GiB then, does'n it?
>
> - If more than 4 Gigabytes is used then answer "64GB" here. This
> + If more than 4 Gibibytes is used then answer "64GB" here. This
^^^^
and 64 GiB for
that one?
-Marc
--
| |
| http://www.links2linux.de <-- Von Linux-Usern fuer Linux-User |
On Thu, 20 Dec 2001, [iso-8859-2] G?bor L?n?rt wrote:
> On Thu, Dec 20, 2001 at 01:52:13PM -0500, Eric S. Raymond wrote:
> > Steven Cole <[email protected]>:
> > > I see that in the very latest Configure.help version, 2.76,
> > > available at http://www.tuxedo.org/~esr/cml2/ Eric has decided to
> > > follow the following standard: IEC 60027-2, Second edition, 2000-11,
> > > Letter symbols to be used in electrical technology - Part 2:
> > > Telecommunications and electronics. and has changed all the
> > > abbreviations for Kilobyte (KB) to KiB, Megabyte (MB) to MiB, etc,
> > > etc.
>
> What? AFAIK 'K' means 1000 in SI. However since computers use binary
> numbers, the number (2^n) which was the most closer to 1000 was selected to
> be used as 'K' for indicating information amount, where n=10. [for decimal
> numbers 10^n (n=3) is used for 'K']. And so on with 'M', 'G' ... Sorry if
> I was wrong about this ...
>
One of the many bad things about changing this kind of stuff is that
it doesn't even follow the rules, i.e., upper case is used for proper
names an/or where there could be a conflict between a previously-defined
abbreviation such as milliampere and megampere (mA, MA). Instead, most
everybody uses K for kilo and it's as absolutely incorrect as possible.
The existing symbols work by fiat. You can't make them "correct" by
following incorrect rules.
If we change anything......, we should define a new system of units,
PI, instead of SI. The basic unit is measurement is the Penguin. It is
abbreviated as p.
Powers of 2:
2 ^ 0 = p (1)
2 ^ 1 = dp dipenguin
2 ^ 2 = qp hepenguin
2 ^ 3 = op octpenguim
2 ^ 4 = hp hexpenguim
2 ^ 5 = ddp duodipenguin
2 ^ 6 = oop octoctpenguin
2 ^ 7 = ohp octohexpenguin
2 ^ 8 = hhp hexahexpenguin
2 ^ 9 = dhhp duohexahexpenguin
2 ^ 10 = kp kilopenguin
2 ^ 20 = mp megapenguin
2 ^ 30 = gp gigapenguin
...etc.
........ otherwise we should leave it alone!
Cheers,
Dick Johnson
Penguin : Linux version 2.4.1 on an i686 machine (797.90 BogoMips).
Santa Claus is coming to town...
He knows if you've been sleeping,
He knows if you're awake;
He knows if you've been bad or good,
So he must be Attorney General Ashcroft.
On Thu, Dec 20, 2001 at 04:05:26PM -0500, Richard B. Johnson wrote:
> One of the many bad things about changing this kind of stuff is that
> it doesn't even follow the rules, i.e., upper case is used for proper
> names an/or where there could be a conflict between a previously-defined
> abbreviation such as milliampere and megampere (mA, MA). Instead, most
OK, that's true, 'MA' is a nightmare even for the first sight ...
> everybody uses K for kilo and it's as absolutely incorrect as possible.
> The existing symbols work by fiat. You can't make them "correct" by
> following incorrect rules.
Oh well, sorry, so let's say about 'k' and 'm'. However an engineer friend
of mine has just say that 'K' is 1024, and 'k' is 1000 ... I dunno anymore ...
[however I've never seen 'Kg' instead of 'kg', but 'mB' or 'mb' are ugly
when compared with 'Mb' and 'MB', not counting that 'b' is bit and 'B' is
byte ... well ... it's confusing sometimes ...]
> 2 ^ 0 = p (1)
> 2 ^ 1 = dp dipenguin
> 2 ^ 2 = qp hepenguin
Nice ;-)
Sorry for OT, I'm going to convert myself into private mails when somebody
will reply ...
- Gabor
On Dec 20, 2001 22:14 +0100, G?bor L?n?rt wrote:
> > everybody uses K for kilo and it's as absolutely incorrect as possible.
> > The existing symbols work by fiat. You can't make them "correct" by
> > following incorrect rules.
>
> Oh well, sorry, so let's say about 'k' and 'm'. However an engineer friend
> of mine has just say that 'K' is 1024, and 'k' is 1000 ... I dunno anymore ...
Well, they are wrong, because 'K' is Kelvin, and not kilo-.
> [however I've never seen 'Kg' instead of 'kg', but 'mB' or 'mb' are ugly
> when compared with 'Mb' and 'MB', not counting that 'b' is bit and 'B' is
> byte ... well ... it's confusing sometimes ...]
Especially since few people work on 1/1000 of a byte (i.e. 'm' is milli,
like mm=millimeter, and not 'M' which is Mega-). So mB and mb are just
plain wrong.
Cheers, Andreas
--
Andreas Dilger
http://sourceforge.net/projects/ext2resize/
http://www-mddsp.enel.ucalgary.ca/People/adilger/
Ra?lN??ez de Arenas Coronado wrote:
> >You are close - he uses "MiB" as short for "mebi" - Mega-binary.
>
> Personally I don't like very much the abbreviations, but I must
> recognize that they remove all possible ambiguity for the
> Configure.help. With MiB, GiB, etc... you're completely sure that you
> are talking about 2^20, 2^30 and not 10^6, 10^9, etc...
Hah! Everytime I have seen "MiB" before this thread I'd though it meant
"Million Bytes", because everyone knows that MB in computers is a
megabyte (<ahem>).
Actually I think I have only ever seen MiB and GiB in the context of
disk drives before now.
-- Jamie
On Thu, Dec 20, 2001 at 10:14:22PM +0100, G?bor L?n?rt wrote:
> On Thu, Dec 20, 2001 at 04:05:26PM -0500, Richard B. Johnson wrote:
> > One of the many bad things about changing this kind of stuff is that
> > it doesn't even follow the rules, i.e., upper case is used for proper
> > names an/or where there could be a conflict between a previously-defined
> > abbreviation such as milliampere and megampere (mA, MA). Instead, most
>
> OK, that's true, 'MA' is a nightmare even for the first sight ...
>
> > everybody uses K for kilo and it's as absolutely incorrect as possible.
> > The existing symbols work by fiat. You can't make them "correct" by
> > following incorrect rules.
>
> Oh well, sorry, so let's say about 'k' and 'm'. However an engineer friend
> of mine has just say that 'K' is 1024, and 'k' is 1000 ... I dunno anymore ...
i have seen kB instead of KB in many places. and the only place i've
ever seen kilo abbreviated as K has been with respect to binary.
> [however I've never seen 'Kg' instead of 'kg', but 'mB' or 'mb' are ugly
> when compared with 'Mb' and 'MB', not counting that 'b' is bit and 'B' is
> byte ... well ... it's confusing sometimes ...]
i was going to comment about simply using lowercase equivalents, but
then milli already has 'm', although the concept of a millibyte (or even
millibit) is absurd.
-mike
--------------------------------------------------------------------------
/~\ The ASCII all that is gold does not glitter
\ / Ribbon Campaign not all those who wander are lost
X Against HTML -- jrr tolkien
/ \ Email!
radiusd+mysql: http://www.cafes.net/~diz/kiss-radiusd
--------------------------------------------------------------------------
At 07:47 PM 12/20/01 +0100, Michael Dunsky wrote:
>You are close - he uses "MiB" as short for "mebi" - Mega-binary.
>Don't laugh - this is official! It's exactly for what you said:
>[snip]
>For a short reading I recommend this:
>
>http://physics.nist.gov/cuu/Units/binary.html
Yet on this page you point to, the first paragraph under the table reads as
follows:
"It is important to recognize that the new prefixes for binary multiples
are not part of the International System of Units (SI), the modern metric
system."
That disclaimer makes it highly UNofficial.
It also goes against legacy use, as well as use within the kernel and GNU
utilities for informative messages.
We have enough problem introducing non-technical people to Linux as it is
without inflicting new and obscure abbreviations. Notice that the
discussion is about changes to Configure.help, the WORST place to start
introducing new and not-widely-used notation.
Hey, it's not a bad idea, but I want it adopted first in things like
stories in the mass media before we start introducing it in Configure.help.
Stephen Satchell
Mike Eldridge wrote:
> i have seen kB instead of KB in many places. and the only place i've
> ever seen kilo abbreviated as K has been with respect to binary.
>
>>[however I've never seen 'Kg' instead of 'kg', but 'mB' or 'mb' are ugly
>>when compared with 'Mb' and 'MB', not counting that 'b' is bit and 'B' is
>>byte ... well ... it's confusing sometimes ...]
>>
>
> i was going to comment about simply using lowercase equivalents, but
> then milli already has 'm', although the concept of a millibyte (or even
> millibit) is absurd.
Why? For instance a millibyte/s might be a hearbeat across a LAN every
hour or so or it might be a control traffic requirement for a deep space
probe. You might not have an immediate use for the term but it has a
specific meaning - and certainly isn't "absurd" (see definition on
http://www.dict.org).
Engineers not (yet) being familiar with the relatively new SI (and IEEE)
binary prefixes is just about acceptable. "Engineers" that misuse k/K
and (worse!) m/M should be in a different field entirely. The SI system
is generally taught as basic science to pre-teenagers. There is no
excuse!
Mike
P.S. Merry Christmas / Mid winter festival / whatever you choose
to celebrate :-)
On 12/21/2001 10:58:30 AM GMT Mike Jagdis wrote:
>
>Mike Eldridge wrote:
>
>> i was going to comment about simply using lowercase equivalents, but
>> then milli already has 'm', although the concept of a millibyte (or even
>> millibit) is absurd.
>
>Why? For instance a millibyte/s might be a hearbeat across a LAN every
>hour or so or it might be a control traffic requirement for a deep space
>probe. You might not have an immediate use for the term but it has a
>specific meaning - and certainly isn't "absurd" (see definition on
>http://www.dict.org).
So, is it 1/1024 or 1/1000 bytes ? :-)
regards,
chris
Christian Groessler wrote:
> On 12/21/2001 10:58:30 AM GMT Mike Jagdis wrote:
>>
>> Mike Eldridge wrote:
>>
>>> i was going to comment about simply using lowercase equivalents, but
>>> then milli already has 'm', although the concept of a millibyte (or even
>>> millibit) is absurd.
>>
>> Why? For instance a millibyte/s might be a hearbeat across a LAN every
>> hour or so or it might be a control traffic requirement for a deep space
>> probe. You might not have an immediate use for the term but it has a
>> specific meaning - and certainly isn't "absurd" (see definition on
>> http://www.dict.org).
>
> So, is it 1/1024 or 1/1000 bytes ? :-)
1/1024. Because we are talking about byte.
Rene
On Fri, 2001-12-21 at 06:50, Rene Engelhard wrote:
> >> Why? For instance a millibyte/s might be a hearbeat across a LAN every
> >> hour or so or it might be a control traffic requirement for a deep space
> >> probe. You might not have an immediate use for the term but it has a
> >> specific meaning - and certainly isn't "absurd" (see definition on
> >> http://www.dict.org).
> >
> > So, is it 1/1024 or 1/1000 bytes ? :-)
>
> 1/1024. Because we are talking about byte.
What does bytes have to do with anything? Is it
1/(2^3 * 10^7) or 1/(2^3 * 2^7)? We're talking about expressing a number
of "bytes"; terms of the base number system don't have any bearing --
and that's the problem. RAM and addressing are restricted to expressions
in terms of binary numbers, as in 2^10, 2^20, etc. Hard drive
manufacturers feel it's neccessary to express storage in terms of base
10 numbers of bytes, even though a sector is 2^9 bytes. In networking,
absolute numbers of bits on the wire are whats important. Though for
some reason telecom engineers have pinned megabit as 1,024,000 bits.
Experienced CS people can glean the proper definition from context, but
the terms should really lend themselves to accurate definition all the
time. If I just say off the cuff that I'm going to send you a megabyte
of data, do I mean 1,000,000 bytes, 1,048,576 bytes, or 1,024,000 bytes?
With the new measures those would be a megabyte, a mebibyte, and 1,024
kilobytes respectively.
Personally I feel that "kibibyte(KiB)" and "mebibyte(MiB)" are silly,
but they are technically unambiguous.
Regards,
Reid
--
"Public sentiment is everything. With public sentiment, nothing can
fail; without it nothing can succeed." -- Abraham Lincoln
>
> Rene
Reid Hekman wrote:
> On Fri, 2001-12-21 at 06:50, Rene Engelhard wrote:
>>>> Why? For instance a millibyte/s might be a hearbeat across a LAN every
>>>> hour or so or it might be a control traffic requirement for a deep space
>>>> probe. You might not have an immediate use for the term but it has a
>>>> specific meaning - and certainly isn't "absurd" (see definition on
>>>> http://www.dict.org).
>>>
>>> So, is it 1/1024 or 1/1000 bytes ? :-)
>>
>> 1/1024. Because we are talking about byte.
>
> What does bytes have to do with anything? Is it
> 1/(2^3 * 10^7) or 1/(2^3 * 2^7)? We're talking about expressing a number
> of "bytes"; terms of the base number system don't have any bearing --
> and that's the problem. RAM and addressing are restricted to expressions
Right.
8 Bit = 1 Byte
1024 Byte = 1 KB
1024 KB = 1 MB
1024 MB ...
So we are talking about that, beacuse the X-Byte is defined as 1024 and not
as 1000 of the previous step.
> in terms of binary numbers, as in 2^10, 2^20, etc. Hard drive
> manufacturers feel it's neccessary to express storage in terms of base
> 10 numbers of bytes, even though a sector is 2^9 bytes. In networking,
> absolute numbers of bits on the wire are whats important. Though for
> some reason telecom engineers have pinned megabit as 1,024,000 bits.
> Experienced CS people can glean the proper definition from context, but
> the terms should really lend themselves to accurate definition all the
> time. If I just say off the cuff that I'm going to send you a megabyte
> of data, do I mean 1,000,000 bytes, 1,048,576 bytes, or 1,024,000 bytes?
What _you_ mean can not be determied from me.
But *I* would mean 1.048.574, otherwise I would say the 9xxx number or say
nearly 1 MB.
> With the new measures those would be a megabyte, a mebibyte, and 1,024
> kilobytes respectively.
That's the sense of them.
Rene
Rene Engelhard wrote:
> Christian Groessler wrote:
>>So, is it 1/1024 or 1/1000 bytes ? :-)
>>
>
> 1/1024. Because we are talking about byte.
Bollocks. How can I put this politely? Don't anyone ever send me
a CV unless you know the technical basics!
Go look up "SI binary prefix" and "SI prefix" on Google. You might
not _like_ the binary prefixes (I don't either) but they're what's
been standardized and they're unambiguous. It does no good to claim
that it's enough that *you* know what you mean. This isn't Alice in
Wonderland (you can look that reference up in your spare time :-) ).
Mike
P.S. After you've understood the SI system you should be able to
tell us what the binary prefix equivalent for m is and why K as
a prefix is a mark of stupidity...
Mike Jagdis wrote:
> Rene Engelhard wrote:
>
>> Christian Groessler wrote:
>>> So, is it 1/1024 or 1/1000 bytes ? :-)
>>>
>>
>> 1/1024. Because we are talking about byte.
>
> Bollocks. How can I put this politely? Don't anyone ever send me
> a CV unless you know the technical basics!
OK, I answered to fast, not reading the mails correct and without
concentration
Surely, millibytes does not make any sense...
> Go look up "SI binary prefix" and "SI prefix" on Google. You might
> not _like_ the binary prefixes (I don't either) but they're what's
> been standardized and they're unambiguous. It does no good to claim
> that it's enough that *you* know what you mean. This isn't Alice in
> Wonderland (you can look that reference up in your spare time :-) ).
A long time ago the MB was under computer scientiests defined as 1024 KB ans
we had used it it one or two decades. Then the hardware manufactures started
to call 1 MB = 1000 MB and that's shit.
I know this problem, I also know the problem, that is gonna be confusing.
I do not insist to change it but I think it is better to make clear what is
defined as 1000 or 1024 insted using MB for both.
> P.S. After you've understood the SI system you should be able to
> tell us what the binary prefix equivalent for m is and why K as
> a prefix is a mark of stupidity...
Have I said that?
Then I apopolgize. Otherwise I read the mail I replied not as careful as I
should and I talked at this moment whitch someone so I was not as
concentraded as I had to...
Rene
On Fri, Dec 21, 2001 at 03:48:22PM +0000, Mike Jagdis wrote:
> Rene Engelhard wrote:
>
> > Christian Groessler wrote:
> >>So, is it 1/1024 or 1/1000 bytes ? :-)
> >>
> >
> > 1/1024. Because we are talking about byte.
>
> Bollocks. How can I put this politely? Don't anyone ever send me
> a CV unless you know the technical basics!
The IEC has adopted these units, not SI. And you could argue that
the IEC has gotten it completely wrong.
Basic unit of length: 1 m
Basic unit of weight: 1 g
...
Basic unit of information: 1 *bit*
Not a byte. A byte is 8 basic units of information.
To be consistent (as with SI), all prefixes should be applied to
the basic unit of measurement - in this case, a bit. And one
kilobit == 1000 bits; one megabit == 1 million bits.
Saying "kilobyte" is like saying "micromicrofarad" or "megakilometer".
However, you *could* also argue that the IEC has it correct. IIRC,
physicists researching electrostatics/magnetostatics at least used to use
(and maybe still do) the centimeter/gram/second units - e.g. an abcoulomb,
which ends up being 10 coulombs. For whatever reason, convention among
this group of people means using cgs units. For whatever reason,
convention among computer people is to use units of 8 bits - 1 byte -
as a measurement standard.
> Go look up "SI binary prefix" and "SI prefix" on Google. You might
> not _like_ the binary prefixes (I don't either) but they're what's
> been standardized and they're unambiguous. It does no good to claim
> that it's enough that *you* know what you mean. This isn't Alice in
> Wonderland (you can look that reference up in your spare time :-) ).
SI standards have been around for years. Yet many mechanical
engineers in the US still use English units. Convention and
economics dictate that they do so; any change in this field is quite
slow.
Somehow I expect that the same convention and economics factors will
also dominate the argument over prefixes for bits of information
in this field for years to come as well.
-Bob
On Fri, Dec 21, 2001 at 11:43:40AM -0600, Bob Glamm wrote:
> Basic unit of weight: 1 g
Actually the base unit of mass is 1 kg, which relates to history.
See http://physics.nist.gov/cuu/Units/units.html
Andreas
--
Andreas Ferber - dev/consulting GmbH - Bielefeld, FRG
---------------------------------------------------------
+49 521 1365800 - [email protected] - http://www.devcon.net
> If you would pay more attention, you can see that on most drives there is
> a small note that says: 1MB = 1000000 bytes. This is why the drive
> capacity is smaller than the manufacturer says.
http://www.seagate.com/products/discsales/discselect/A1a2.html#cap
Capacity:
Capacity is the amount of data that the drive can store, after
formatting. Most disc drive companies, including Seagate, calculate disc
capacity based on the assumption that 1 megabyte = 1000 kilobytes and 1
gigabyte=1000 megabytes.
Disks have a natural measurement of capacity based on an integral number
of 512byte blocks. So kilobytes (1024) makes sense for them.
The only marketing wizardry is to use the smaller of:
1 megabyte = 1000 kilobytes
instead of:
1 megabyte = 1024 kilobytes
There are valid arguements for both interpretations.
--
__O
Lineo - For Embedded Linux Solutions _-\<,_
PGP Fingerprint: 28 E2 A0 15 99 62 9A 00 (_)/ (_) 88 EC A3 EE 2D 1C 15 68
Stuart Lynne <[email protected]> http://www.lineo.com 604-461-7532
On Thursday 20 December 2001 04:41 pm, Jamie Lokier wrote:
> Ra?lN??ez de Arenas Coronado wrote:
> > >You are close - he uses "MiB" as short for "mebi" - Mega-binary.
> >
> > Personally I don't like very much the abbreviations, but I must
> > recognize that they remove all possible ambiguity for the
> > Configure.help. With MiB, GiB, etc... you're completely sure that you
> > are talking about 2^20, 2^30 and not 10^6, 10^9, etc...
>
> Hah! Everytime I have seen "MiB" before this thread I'd though it meant
> "Million Bytes", because everyone knows that MB in computers is a
> megabyte (<ahem>).
Me, I've got the "men in black" song running through my head now. And I
don't particularly like it, either.
This reminds me of people going after metric speed limit signs with shotguns
back in the late 70's. I never understood this impulse before now.
Trust the ISO to change the BINARY nomenclature on something that has, until
now, DEFAULTED to binary. I get the feeling nobody at the ISO actually uses
computers much. Is it too late to sober them up?
Rob
Please NO NO NO NO!!
Why on earth is it good to develop misunderstandings and inconsistency
with well- and widely-known historical abbrevs?
I (and I think I'm far not alone) would hate to see those abbrevs. I
really don't want to vomit every time I read configure.help or an
ifconfig output.
This is a 3-year old decision, and haven't seen it in use anywhere before.
If this knew style would be the common use in IT, then this change is OK.
But _not_ now. (and hopefully never).
So may I suggest considering this change a few years later, _if_ it comes
into common use?
On Thu, 20 Dec 2001, Bernd Eckenfels wrote:
> In article <[email protected]> you wrote:
> > Eric has decided to follow the following standard:
> > IEC 60027-2, Second edition, 2000-11,
> > Letter symbols to be used in electrical technology -
> > Part 2: Telecommunications and electronics.
> > and has changed all the abbreviations for Kilobyte (KB) to KiB,
> > Megabyte (MB) to MiB, etc, etc.
>
> I did this for nettools (i.e. ifconfig), too:
>
> RX bytes:2120660294 (1.9 GiB) TX bytes:341183013 (325.3 MiB)
>
> man page:
>
> Since net-tools 1.60-4 ifconfig is printing byte counters
> with SI units. So 1 KiB are 2^10 byte. Note, the numbers
> are truncated to one decimal (which can by quite a large
> error if you consider 0.1 PiB is 112.589.990.684.262 bytes :)
> ...
> SEE ALSO
> route(8), netstat(8), arp(8), rarp(8), ipchains(8)
> http://physics.nist.gov/cuu/Units/binary.html - Prefixes
> for binary multiples
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
--
pozsy
On Sat, Dec 22, 2001 at 11:24:17AM +0100, Pozsar Balazs wrote:
| Please NO NO NO NO!!
|
| Why on earth is it good to develop misunderstandings and inconsistency
| with well- and widely-known historical abbrevs?
|
| I (and I think I'm far not alone) would hate to see those abbrevs. I
| really don't want to vomit every time I read configure.help or an
| ifconfig output.
|
|
| This is a 3-year old decision, and haven't seen it in use anywhere before.
| If this knew style would be the common use in IT, then this change is OK.
| But _not_ now. (and hopefully never).
|
| So may I suggest considering this change a few years later, _if_ it comes
| into common use?
I can understand your point about not jumping into something that will turn
out (possibly) to be a big flop and cause new confusion. However, I'd like
to point out that any new idea will _never_ become adopted if everyone takes
the position of "I'm not going to do it until most everyone else does first".
If this is going to be the common usage, I believe we need to be the lead on
this and do it. The question is, how can we determine if it will be common
_before_ anyone else takes the lead to make it be so.
--
-----------------------------------------------------------------
| Phil Howard - KA9WGN | Dallas | http://linuxhomepage.com/ |
| [email protected] | Texas, USA | http://phil.ipal.org/ |
-----------------------------------------------------------------
On Sat, Dec 22, 2001 at 11:24:17AM +0100, Pozsar Balazs wrote:
> Why on earth is it good to develop misunderstandings and inconsistency
> with well- and widely-known historical abbrevs?
There is no well understood abrev. for power of two units, thats why
somebody had to introduce them.
Actually it was not my Idea, but since I got multiple error reports for
multible writings (and that even while the original count is visible in the
ouput) I decided to go with the standard.
Greetings
Bernd
--
(OO) -- [email protected] --
( .. ) ecki@{inka.de,linux.de,debian.org} http://home.pages.de/~eckes/
o--o *plush* 2048/93600EFD eckes@irc +497257930613 BE5-RIPE
(O____O) When cryptography is outlawed, bayl bhgynjf jvyy unir cevinpl!
At 04:47 AM 12/22/01 -0600, Phil Howard wrote:
>I can understand your point about not jumping into something that will turn
>out (possibly) to be a big flop and cause new confusion. However, I'd like
>to point out that any new idea will _never_ become adopted if everyone takes
>the position of "I'm not going to do it until most everyone else does first".
Let's make clear what we are talking about. We are talking about making
the change to the one place that will be exposed to the non-technical user,
but not anywhere else (internal documentation, comments to the source code,
output from /proc, userland programs). In other words, the absolutely
worst place in the GNU/Linux system to introduce new and confusing usage
that is not widely used.
Get the hint? Let's change usage FIRST in the places that don't have the
exposure of a Help File to the general public, THEN consider making the
change once the users (technical and non-) have had a chance to voice their
opinion.
I don't know about y'all, but my non-technical clients have enough problems
with the historical abbreviations and terminology without being thrown this
curve. Unless you can get the business community to cry out for the change
(so that BSD, what's left of BeOS, and Microsoft make the change) this is a
Bad Idea(tm).
Look, I agree that there is significant merit to KiB et. al., but the
marketplace has not always selected that which is best. That's the nature
of the marketplace.
Stephen Satchell
On Sat, 22 Dec 2001, Bernd Eckenfels wrote:
> On Sat, Dec 22, 2001 at 11:24:17AM +0100, Pozsar Balazs wrote:
> > Why on earth is it good to develop misunderstandings and inconsistency
> > with well- and widely-known historical abbrevs?
>
> There is no well understood abrev. for power of two units, thats why
> somebody had to introduce them.
>
> Actually it was not my Idea, but since I got multiple error reports for
> multible writings (and that even while the original count is visible in the
> ouput) I decided to go with the standard.
What error reports do you receive?
This might be a standard, but _it_is_not_adopted_anywhere! (at least I
haven't seen it anywhere (I might be blind)).
I would rather call standard something which is adopted in real life.
--
Balazs Pozsar
On Sat, 22 Dec 2001 21:18:08 +0100 (MET), Pozsar Balazs wrote:
>
>This might be a standard, but _it_is_not_adopted_anywhere! (at least I
>haven't seen it anywhere (I might be blind)).
>I would rather call standard something which is adopted in real life.
No, that would be a socalled defacto standard. Ie. not necessarily a standard
proposed and recommended by a standards body. For instance, you may think that
the UK using pounds and ounces is standard - after all they are both being
used in real life. But it is contrary to the standard. (as given by the SI
system).
rgds,
Per
regards,
Per Jessen, Zurich
http://www.enidan.com - home of the J1 serial console.
Windows 2001: "I'm sorry Dave ... I'm afraid I can't do that."
In article <Pine.GSO.4.30.0112222114120.9228-100000@balu> you wrote:
> What error reports do you receive?
about kb, Kb, kB beeing wrong (even if the number in bytes is listed ppl
complaint they had to pay more money then expected... i wont comment on this,
but I have now a solution which will get used, I am sure :)
Greetings
Bernd
On Saturday 22 December 2001 03:22 pm, Per Jessen wrote:
> On Sat, 22 Dec 2001 21:18:08 +0100 (MET), Pozsar Balazs wrote:
> >This might be a standard, but _it_is_not_adopted_anywhere! (at least
> > I haven't seen it anywhere (I might be blind)).
> >I would rather call standard something which is adopted in real
> > life.
>
> No, that would be a socalled defacto standard. Ie. not necessarily a
> standard proposed and recommended by a standards body. For instance,
> you may think that the UK using pounds and ounces is standard - after
> all they are both being used in real life. But it is contrary to the
> standard. (as given by the SI system).
Not anymore. Apparently it is now illiegal in the UK to use Imperial
measurements (or for that matter, anything but standard metric) in
transactions. If people are still using the Imperial system in day to
day life, they won't be very soon (and yes, this has infact been
enforced at least once that I've heard of.)
I just eagerly await the day that the U.S. government decides to stick
its ass in and tell U.S. citizens (and whatever other countries they
decide to pressure into agreeing with them) what we can and can't use
to abbreviate 1024bytes, since we're all too busy arguing over wether
or not to abide by the defacto standard that has been in use longer
than I've been ALIVE (which yes, means I've not been alive very long.)
Let me know when we're done fighting about wether or not we should
confuse everyone (further) that tries to configure their first kernel.
[email protected] said:
> Look, I agree that there is significant merit to KiB et. al., but the
> marketplace has not always selected that which is best. That's the
> nature of the marketplace.
The marketplace selected Windows.
If you are more interested in the choices of the marketplace than in
technical correctness, one has to wonder what you're doing on this mailing
list.
--
dwmw2
On Sun, Dec 23, 2001 at 10:43:34AM +0000, David Woodhouse wrote:
>
> [email protected] said:
> > Look, I agree that there is significant merit to KiB et. al., but the
> > marketplace has not always selected that which is best. That's the
> > nature of the marketplace.
>
> The marketplace selected Windows.
>
> If you are more interested in the choices of the marketplace than in
> technical correctness, one has to wonder what you're doing on this mailing
> list.
Well, I always assumed Linux was supposed to "work damn well", not
"be technically correct".
--
Vojtech Pavlik
SuSE Labs
> Not anymore. Apparently it is now illiegal in the UK to use Imperial
> measurements (or for that matter, anything but standard metric) in
> transactions. If people are still using the Imperial system in day to
Half accurate
Over time we are gradually shifting more and more to metric units. Its slow
because the only fair way to let elderly people who don't know metric handle
things has been to wait until they pass on.
Weights and measures have gone from imperial, through imperial and metric
now to metric values. We still have people selling 454g bags in part
because of the time it takes them to change sizes.
Other areas - alcoholic drink sizes, road distances and road speeds have yet
to be changed, but eventually will be.
Alan
On Sat, 22 Dec 2001 23:21:11 -0800, Nicholas Knight wrote:
>On Saturday 22 December 2001 03:22 pm, Per Jessen wrote:
>> On Sat, 22 Dec 2001 21:18:08 +0100 (MET), Pozsar Balazs wrote:
>> >This might be a standard, but _it_is_not_adopted_anywhere! (at least
>> > I haven't seen it anywhere (I might be blind)).
>> >I would rather call standard something which is adopted in real
>> > life.
>>
>> No, that would be a socalled defacto standard. Ie. not necessarily a
>> standard proposed and recommended by a standards body. For instance,
>> you may think that the UK using pounds and ounces is standard - after
>> all they are both being used in real life. But it is contrary to the
>> standard. (as given by the SI system).
>
>Not anymore. Apparently it is now illiegal in the UK to use Imperial
>measurements (or for that matter, anything but standard metric) in
>transactions. If people are still using the Imperial system in day to
>day life, they won't be very soon (and yes, this has infact been
>enforced at least once that I've heard of.)
Yeah, I know of those too. I lived in London when that case with the greengrocer
or butcher or whatever was on. I think the guy was prosecuted.
That case was partially about the metric system, partially about EU legislation.
The UK government had agreed that by a certain date, all supermarkets etc.
would be required by law to mark all goods with a per 100g price. Not per ounce
or whatever.
This is of course caused a few people to whine ....
And soon AFAIR, at least one supermarket turned around, and stopped also marking
goods in price per ounce etc. Downright silly.
>
>Let me know when we're done fighting about wether or not we should
>confuse everyone (further) that tries to configure their first kernel.
Sooner or later, the IEC notation will also make it into the schools, and students
will be graduating knowing nothing but the IEC notation. If we haven't changed by
then, that's when we'll see some true confusion.
/Per Jessen, Zurich.
regards,
Per Jessen, Zurich
http://www.enidan.com - home of the J1 serial console.
Windows 2001: "I'm sorry Dave ... I'm afraid I can't do that."
At 10:43 AM 12/23/01 +0000, David Woodhouse wrote:
>If you are more interested in the choices of the marketplace than in
>technical correctness, one has to wonder what you're doing on this mailing
>list.
Nice ad hominem attack, David. Attack the messenger. Good boy.
If you recall (or perhaps you have placed me in your kill file) I mentioned
that the last place you want to make a sudden change in terminology is in
the most public of places, the configuration help file. You also missed my
call to use the new abbreviations FIRST within the kernel itself, in /proc,
and in those userland utilities that report system resource usage. If the
universe of Linux users accept it (translation: no flames erupt) then you
can consider populating the configuration help files with them.
I also mentioned that we have a very, very large base of "legacy users" who
do not understand what MiB would be (outside of the context of the movie
_Men in Black_) and who would become very, very confused. In short, making
the change would CONFUSE THE NON-TECHNICAL USERS more than they already are.
I'm not against technical correctness. I'm against witless, thoughtless,
blind deployment of an idea without considering the consequences of that
deployment.
Don't underestimate the power of "the market." Have you seen much about
MINIX lately?
Stephen Satchell
At 10:18 AM 12/23/01 +0100, Per Jessen wrote:
>Sooner or later, the IEC notation will also make it into the schools, and
>students
>will be graduating knowing nothing but the IEC notation. If we haven't
>changed by
>then, that's when we'll see some true confusion.
Or the professors and research assistants will say "no" and that will be
that. Don't forget that the IEC doesn't cast things in stone, either -- as
THEY get feedback from their members, they may decide to shit-can the whole
thing.
That's another reason to take it slowly -- let's see if the idea has
serious legs.
Stephen Satchell
[email protected] said:
> >If you are more interested in the choices of the marketplace than in
> >technical correctness, one has to wonder what you're doing on this mailing
> >list.
> Nice ad hominem attack, David. Attack the messenger. Good boy.
That's not what I understood 'ad hominem' to mean. My understanding was that
ad hominem involved an attack on the person making the argument, followed by
an obviously false assertion that such attack renders the person's arguments
invalid, even though the details of the attack made are completely unrelated
to the matter being discussed.
Thus observing that you sent your mail using a Windows MUA, then declaring
that your argument is invalid because you're a Windows user and therefore
obviously mentally deficient, would be an ad hominem attack.
My response, though it could possibly be called an 'attack' if you were
feeling particularly thin-skinned, was definitely based upon the discussion
at hand - I expressed surprise at the criterion of marketplace acceptability
which you used to justify your position.
> I also mentioned that we have a very, very large base of "legacy
> users" who do not understand what MiB would be (outside of the
> context of the movie _Men in Black_) and who would become very, very
> confused. In short, making the change would CONFUSE THE
> NON-TECHNICAL USERS more than they already are.
But this term _is_ used outside that context. And the context it's used in,
in just about all cases, makes it blindingly obvious to all but the densest
reader what the intended meaning is. Maybe _those_ people will remain
slightly confused about where we mean 10^3 and where we mean 2^10, but at
least people with a clue no longer have to be confused about such things.
As an example - what possible meaning could you contrive for 'KiB' in the
following:
This lets you select the page size of the kernel. For best IA-64
performance, a page size of 8KiB or 16KiB is recommended. For best
IA-32 compatibility, a page size of 4KiB should be selected (the vast
majority of IA-32 binaries work perfectly fine with a larger page
size). For Itanium systems, do NOT chose a page size larger than
16KiB.
Surely it's difficult to imagine anyone reading that and coming to any
other conclusion than the correct one?
I accept that is often appropriate to 'dumb down' documentation and
explanations somewhat to cater for the lowest common denominator members of
the audience.
It is much more rarely appropriate to dumb it down so far that it becomes
factually inaccurate. The tuition of physics at high school, in Further
Education and then Higher Education is perhaps an example of when such
oversimplifications are necessary and appropriate. Some people will never
need to know that Newton's Laws break down, and even if that weren't the
case, they wouldn't have a whelk's chance in a supernova of understanding
Relativity anyway. So why trouble them with it?
But in the situation at hand, there is no justification for catering to the
ignorant in our documentation to the extent that it becomes inaccurate. The
difficulty in understanding the correct text is just not sufficient to
justify the inaccuracies.
--
dwmw2
Eric S. Raymond wrote:
>
>
> This change came as a patch from David Woodhouse. I think the new
> abbreviations are awful ugly, myself, but they do have the virtue of
> not being ambiguous. So I swallowed hard and took the patch.
>
This could even have the nice side effect of teaching something to Linux
newbies (mainly the fact that the difference between 2^10 and 10^3
matters in some areas). I see 2 cases :
- already encountered the kiB/MiB/GiB notation and understood the
meaning: no problem if we take out of the equation the aesthetic of the
abreviations.
- this is a new thing for the reader, 3 cases:
. Computer literate person : she uses her intuition and understand its
meaning : no problem apart the time used to put her intuition at work,
. Computer illiterate person which don't care enough : she doesn't
understand the difference with kB/MB/GB and takes the notation as a
different syntax but with the same semantic : the only problem would be
a temporary confusion (from a fraction of a second to several minutes)
until this assumption is made. It certainly would be made by computer
illiterate people who are lost in the first place: we trade a
misunderstanding for another.
. Computer illiterate person which really tries to understand : she
doesn't understand and take the time to document herself : no problem
she might discover something she didn't even thought about.
This is a simplified view but I believe the actual readers' behaviour
would be somehow a combination of several of the above cases.
So what's the tradeoff :
* aesthetic and shor time spent in temporary confusion or reflexion
* for clarity and education of some people in the end.
Hiding complexity in the docs would only keep some users ignorant.
This is my personal opinion but don't we prefer educated users instead
of ignorant ones ?
I find the choice obvious...
We could argue on the choice of these particular abreviations against
others but as I don't see any other around...
LB
Hi!
> > This change came as a patch from David Woodhouse. I think the new
> > abbreviations are awful ugly, myself, but they do have the virtue of
> > not being ambiguous. So I swallowed hard and took the patch.
> >
>
> This could even have the nice side effect of teaching something to Linux
> newbies (mainly the fact that the difference between 2^10 and 10^3
> matters in some areas). I see 2 cases :
>
> - already encountered the kiB/MiB/GiB notation and understood the
> meaning: no problem if we take out of the equation the aesthetic of the
> abreviations.
Heh, is it kiB or KiB? Anyway, I guess yes MiB units should be used.
We already have them in dmesg output. (And btw it confused me because
it reported *disk size* in MiB.... So I assumed MiB must be 10e6).
Pavel
--
Philips Velo 1: 1"x4"x8", 300gram, 60, 12MiB, 40bogomips, linux, mutt,
details at http://atrey.karlin.mff.cuni.cz/~pavel/velo/index.html.
Pavel Machek wrote:
>
> Heh, is it kiB or KiB?
Don't know, didn't read IEC 60027-2. From Steven first post on the
subject and Configure.help content, I must be wrong and it is KiB.
Lionel.
Hi Mike.
>> [however I've never seen 'Kg' instead of 'kg', but 'mB' or 'mb' are
>> ugly when compared with 'Mb' and 'MB', not counting that 'b' is bit
>> and 'B' is byte ... well ... it's confusing sometimes ...]
> i was going to comment about simply using lowercase equivalents, but
> then milli already has 'm', although the concept of a millibyte (or
> even millibit) is absurd.
I remember reading somewhere that the Voyager space probes now
communicate information back to earth at the incredibly fast rate of 10
minutes per byte - which my mental maths says is somewhere in the region
of 1.6 mB per second - using precicely the unit you decry.
> --------------------------------------------------------------------------
> /~\ The ASCII all that is gold does not glitter
> \ / Ribbon Campaign not all those who wander are lost
> X Against HTML -- jrr tolkien
> / \ Email!
> --------------------------------------------------------------------------
Have you purloined the emblem of the Scottish National Party ???
Refer to http://www.snp.org.uk/ for details.
Best wishes from Riley.
Hi Rene.
>> If I just say off the cuff that I'm going to send you a megabyte of
>> data, do I mean 1,000,000 bytes, 1,048,576 bytes, or 1,024,000
>> bytes?
> What _you_ mean can not be determied from me. But *I* would mean
> 1.048.574, otherwise I would say the 9xxx number or say nearly 1 MB.
You obviously give short change as a matter of course then...
(Assuming that you use . as a thousands separator rather than as a
decimal point the way most of the Enlish speaking world does)
Best wishes from Riley.
[email protected] (Pozsar Balazs) wrote on 22.12.01 in <Pine.GSO.4.30.0112221113120.2091-100000@balu>:
> Please NO NO NO NO!!
>
> Why on earth is it good to develop misunderstandings and inconsistency
> with well- and widely-known historical abbrevs?
It's not. The sooner we bury this computer-originated idiocy of using
decimal prefixes to denote binary values, the better.
You *do* remember which version has the longer history, right?
MfG Kai
On a sunny Sat, 22 Dec 2001 04:47:43 -0600 Phil Howard gathered a sheaf of
electrons and etched in their motions the following immortal words:
> On Sat, Dec 22, 2001 at 11:24:17AM +0100, Pozsar Balazs wrote:
> I can understand your point about not jumping into something that will
> turn out (possibly) to be a big flop and cause new confusion. However,
> I'd like to point out that any new idea will _never_ become adopted if
> everyone takes the position of "I'm not going to do it until most
> everyone else does first".
I hope *everyone* takes just that position. MiB GiB. yeah, and I've got a
miB right here. fractional bits. ha!
fucking moronic is a phrase that springs to mind. probably something an
accountant thought of whilst twiddling his/her thumbs staring at excel.
In article <[email protected]> you write:
>Engineers not (yet) being familiar with the relatively new SI (and IEEE)
>binary prefixes is just about acceptable. "Engineers" that misuse k/K
>and (worse!) m/M should be in a different field entirely. The SI system
>is generally taught as basic science to pre-teenagers. There is no
>excuse!
>
How many of them learn it though?
Jonathan (occasionally guilty of s/kB/KB/ himself).
On Wednesday 02 January 2002 11:17, Jonathan Amery wrote:
> In article <[email protected]> you write:
> >Engineers not (yet) being familiar with the relatively new SI (and IEEE)
> >binary prefixes is just about acceptable. "Engineers" that misuse k/K
> >and (worse!) m/M should be in a different field entirely. The SI system
> >is generally taught as basic science to pre-teenagers. There is no
> >excuse!
>
> How many of them learn it though?
>
> Jonathan (occasionally guilty of s/kB/KB/ himself).
>
For the 10th time, the K v. k issue is due to the standards
body ignoring common sense and following tradition instead.
All positive powers of ten should have upper-case letters
(D,H,K,M,T,P)
and negative powers of ten should use lower-case letters.
(d,c,m,n,p)
The KB meaning 2^10 B instead of 10^3 B is just plain dumb,
and that's why the standards body tried to fix it with KiB.
But again, this solution was considered to look and sound
goofy and to be based on stupid mathematical games;
hence this whole long thread. <rant>A thread which has shown
to me that most comp. sci. folks lack common sense and
are pendantic to the max.</rant>
--
[email protected].
On January 2, 2002 09:17 pm, Timothy Covell wrote:
> On Wednesday 02 January 2002 11:17, Jonathan Amery wrote:
> > In article <[email protected]> you write:
> > >Engineers not (yet) being familiar with the relatively new SI (and IEEE)
> > >binary prefixes is just about acceptable. "Engineers" that misuse k/K
> > >and (worse!) m/M should be in a different field entirely. The SI system
> > >is generally taught as basic science to pre-teenagers. There is no
> > >excuse!
> >
> > How many of them learn it though?
> >
> > Jonathan (occasionally guilty of s/kB/KB/ himself).
>
> For the 10th time, the K v. k issue is due to the standards
> body ignoring common sense and following tradition instead.
> All positive powers of ten should have upper-case letters
> (D,H,K,M,T,P)
> and negative powers of ten should use lower-case letters.
> (d,c,m,n,p)
So if the box says '16 mB' flash, that's 16 millibytes, right?
> The KB meaning 2^10 B instead of 10^3 B is just plain dumb,
> and that's why the standards body tried to fix it with KiB.
> But again, this solution was considered to look and sound
> goofy and to be based on stupid mathematical games;
> hence this whole long thread. <rant>A thread which has shown
> to me that most comp. sci. folks lack common sense and
> are pendantic to the max.</rant>
Yes, true, and?
--
Daniel
On Wednesday 02 January 2002 22:23, Daniel Phillips wrote:
> On January 2, 2002 09:17 pm, Timothy Covell wrote:
> > On Wednesday 02 January 2002 11:17, Jonathan Amery wrote:
> > > In article <[email protected]> you write:
> > > >Engineers not (yet) being familiar with the relatively new SI (and
> > > > IEEE) binary prefixes is just about acceptable. "Engineers" that
> > > > misuse k/K and (worse!) m/M should be in a different field entirely.
> > > > The SI system is generally taught as basic science to pre-teenagers.
> > > > There is no excuse!
> > >
> > > How many of them learn it though?
> > >
> > > Jonathan (occasionally guilty of s/kB/KB/ himself).
> >
> > For the 10th time, the K v. k issue is due to the standards
> > body ignoring common sense and following tradition instead.
> > All positive powers of ten should have upper-case letters
> > (D,H,K,M,T,P)
> > and negative powers of ten should use lower-case letters.
> > (d,c,m,n,p)
>
> So if the box says '16 mB' flash, that's 16 millibytes, right?
Of course that's what it should mean, but obviously the folks
who packaged it were not too up on things. And, let's face
the music here, if we as group of supposedly smart people
can't come close to reaching any consensus, can we blame
marketing folks for mislabeling something???
>
> > The KB meaning 2^10 B instead of 10^3 B is just plain dumb,
> > and that's why the standards body tried to fix it with KiB.
> > But again, this solution was considered to look and sound
> > goofy and to be based on stupid mathematical games;
> > hence this whole long thread. <rant>A thread which has shown
> > to me that most comp. sci. folks lack common sense and
> > are pendantic to the max.</rant>
>
> Yes, true, and?
Sorry, I was feeling frustrated and should've kept that internal.
<rant>
But the fact is that I've presented what I think are good arguments
on how to "fix" things in a rational and _consistant_ manner.
That is the point of metrics after all, to have a simple, rational,
and consistant set of units of measurement to replace the
inconsistant units that were previously used. If we can't do
this, then there is no point in converting. Americans are right
to continue using English/America/Imperial units if even the
so called cognescenti can't get metrics right.
</rant>
--
[email protected].