On Mon, Jun 11, 2007 at 10:15:25PM +0200, Josselin Mouette wrote:
> Le lundi 11 juin 2007 à 15:25 -0400, Joey Hess a écrit :
> > > You seem to fancy the K-is-1024--k-is-1000 convention

> > No, I hate that convention. K and k should only ever refer to 1024.

> /me waits for the day measuring jugs are graduated in powers of two,
> just to please a group of hackers who don't like SI units.

Shall I bring you a half gallon of milk to Scotland, then?

"Pff, base 10, the metric system is so archaic"

--
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
vorlon@debian.org http://www.debian.org/


--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

* Eduard Bloch [070611 22:27]:
> > Can you tell me in which cases you would make a different decision if this was
> > either 2134*1000 or 2134*1024 bytes?
> >
> > In either case, ~ 2 million bytes suits your requirement, or it doesn't.
> > This sounds to me like solving a non-problem, unless you can of course tell me
> > in which situations adding the "B" or "iB" in the field above would solve a
> > real question.
>
> Excuse me? Pretty simple example: you have only 2.03 GB (real GB)
> remaining free space (seen in some disk info tool) on your harddisk and
> you are fetching a 2GB file (2 fake GB, 2GiB in fact). So what, it
> breaks about 99% and displays some very unexpected message.

Isn't uncompressed size filesystem dependent? (At least the space the
package will need when being installed will be).

Hochachtungsvoll,
Bernhard R. Link


--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

On Monday 11 June 2007 15:10:24 Hendrik Sattler wrote:
> Am Montag 11 Juni 2007 22:15 schrieb Josselin Mouette:
> > Le lundi 11 juin 2007 Ã* 15:25 -0400, Joey Hess a écrit :
> > > > You seem to fancy the K-is-1024--k-is-1000 convention
> > >
> > > No, I hate that convention. K and k should only ever refer to 1024.
> >
> > /me waits for the day measuring jugs are graduated in powers of two,
> > just to please a group of hackers who don't like SI units.
>
> And you have to change their world in an useless atempt?
> Abbreviations are ambiguous by design. Who actually says that KB means
> kilobyte?

Well, in SI units, KB never means kilobyte, and is not ambiguous at all;
it's a kelvin·bel.

--
Wesley J. Landaker
OpenPGP FP: 4135 2A3B 4726 ACC5 9094 0097 F0A9 8A4C 4CD6 E3D2

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQBGbd+D8KmKTEzW49IRApuKAJ9O5ckn1d5kjDx9Uc0k/1N0Lxux5gCeIxB4
SHVBFS7j0u7yBz6SmIrf92k=
=7pGP
-----END PGP SIGNATURE-----

Bastian Venthur writes:
> I agree with the "sounds stupid" part, although I don't belive this is a
> valid argument. What I don't believe is your 80 colums argument. Could
> you please name a few of the *many* programs which would have to drop
> information, precision, or significantly change their display to use the
> "KiB" unit?

E.g. all of the "top"-type programs, which display stuff in colums like
memory sizes etc, and are _extremely_ starved for space. It would be
ludicrous to change the memory size displays in such programs from
"224K" to "224KiB" just to give some obsessive-compulsive types a warm
fuzzy.

> On the other hand, we have the chance to avoid user confusion

No one is actually confused.

This "standard" doesn't actually solve a real problem.

-Miles

--
"Whatever you do will be insignificant, but it is very important that
you do it." Mahatma Gandhi


--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

Fine. Stick with Kilobytes, but strictly define it as 10^3 bytes. Just
choose one over the other and be consistent.

On Tue, 2007-06-12 at 01:53 +0900, Miles Bader wrote:
> shirish writes:
> > It isn't just ubuntu or debian but this needs to be done
> > everywhere.
>
> No it doesn't.
>
> The "SI binary prefixes" are an abomination.
>
> "Kibibytes"? Christ... [Did they try pronouncing these horrid things
> when "standarizing" them?!?]
>
> -Miles
>
> --
> We are all lying in the gutter, but some of us are looking at the stars.
> -Oscar Wilde
>
>
--
Alex Jones
http://alex.weej.com/


--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

On 6/11/07, Alex Jones wrote:
> Fine. Stick with Kilobytes, but strictly define it as 10^3 bytes. Just
> choose one over the other and be consistent.

That's not "consistent". Kilobyte has always meant 2^10 bytes. "kilo"
in "kilobyte" is not an SI prefix. SI prefixes only apply to SI
measurements, of which "byte" is not a member. There is no confusion;
the only place where a kilobyte != 2^10 bytes is in hard drive
manufacturer's advertising materials. This is the way it has been for
decades, and it is a perfectly acceptable and desirable standard.

>
> On Tue, 2007-06-12 at 01:53 +0900, Miles Bader wrote:
> > shirish writes:
> > > It isn't just ubuntu or debian but this needs to be done
> > > everywhere.
> >
> > No it doesn't.
> >
> > The "SI binary prefixes" are an abomination.
> >
> > "Kibibytes"? Christ... [Did they try pronouncing these horrid things
> > when "standarizing" them?!?]
> >
> > -Miles
> >
> > --
> > We are all lying in the gutter, but some of us are looking at the stars.
> > -Oscar Wilde
> >
> >
> --
> Alex Jones
> http://alex.weej.com/
>
>
> --
> Ubuntu-devel-discuss mailing list
> Ubuntu-devel-discuss@lists.ubuntu.com
> Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/lis...-devel-discuss
>


--
Mark Reitblatt


--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

On Mon, 2007-06-11 at 19:56 -0500, Mark Reitblatt wrote:
> On 6/11/07, Alex Jones wrote:
> > Fine. Stick with Kilobytes, but strictly define it as 10^3 bytes. Just
> > choose one over the other and be consistent.
>
> That's not "consistent". Kilobyte has always meant 2^10 bytes. "kilo"
> in "kilobyte" is not an SI prefix. SI prefixes only apply to SI
> measurements, of which "byte" is not a member.

Then why bastardise an SI prefix? This surely serves only to confuse
people. Why don't we invent a new word? Should we call it the
"thousandbyte"?

It is simply a convenient accident that 2^10 ~= 10^3. As I'm sure you're
well aware, this approximation starts to become way off as you approach
tera-. In fact, that's about 10% error, which is simply unacceptable.
It's time to move on and accept that the approximation fails with big
numbers.

> There is no confusion;
> the only place where a kilobyte != 2^10 bytes is in hard drive
> manufacturer's advertising materials. This is the way it has been for
> decades, and it is a perfectly acceptable and desirable standard.

And I suppose you think that differences such as that between the
American and the English ton are acceptable and desirable. Let it be
known that I strongly disagree with you here.
--
Alex Jones
http://alex.weej.com/


--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

Miles Bader wrote:
> Bastian Venthur writes:
>
> > On the other hand, we have the chance to avoid user confusion
>
> No one is actually confused.

can you say, in all the thousands of users, not a single one is ever
confused? Not a single one ever wonders if it means 1000 or 1024?
1048576 or 1000000? 1073741824 or 1000000000?

> This "standard" doesn't actually solve a real problem.
http://physics.nist.gov/cuu/Units/binary.html


It does solve a real problem. It solves an ambiguity. Does k mean 1000
or 1024? Does M mean 1000000 or 1048576?

Answer: k mean 1 000
ki means 1 024
m means 1 000 000
mi means 1 048 576

No more ambiguity.

Because you see no problem does not mean there is none.

If you want to use the ``classical'' SI units as a basis, then look no
further than deka: abbreviated da.
http://physics.nist.gov/cuu/Units/prefixes.html

--
John H. Robinson, IV jaqque@sbih.org
http ((((
WARNING: I cannot be held responsible for the above, sbih.org ( )(:[
as apparently my cats have learned how to type. spiders.html ((((


--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

On Mon, Jun 11, 2007 at 02:32:35PM -0700, Steve Langasek wrote:
> On Mon, Jun 11, 2007 at 10:15:25PM +0200, Josselin Mouette wrote:
> > Le lundi 11 juin 2007 à 15:25 -0400, Joey Hess a écrit :
> > > > You seem to fancy the K-is-1024--k-is-1000 convention
>
> > > No, I hate that convention. K and k should only ever refer to 1024.
>
> > /me waits for the day measuring jugs are graduated in powers of two,
> > just to please a group of hackers who don't like SI units.
>
> Shall I bring you a half gallon of milk to Scotland, then?
>
> "Pff, base 10, the metric system is so archaic"
>
The US gallon being the Queen Anne period wine gallon or 0.83
gallons? I'll stick to Imperial thanks - four pints will be fine in
anywhere in Scotland/England/Ireland/Wales and I'll get slightly more

Fuel consumption in UK is defined as relative to the (Imperial)
gallon: the Spanish Empire,of course, did far better - in a good year
they got several thousand miles to the galleon

Andy

K or k in a computer context, is, OF COURSE, 1024

On Mon, Jun 11, 2007 at 09:55:35PM +0200, Magnus Holmgren wrote:
> On Monday 11 June 2007 21:41, Joey Hess wrote:
> > Alex Queiroz wrote:
> > > On 6/11/07, Joey Hess wrote:
> > > >No, I hate that convention. K and k should only ever refer to 1024.
> > >
> > > Like in kg or km?
> >
> > This thread is about units of data.
>
> kbit? kbit/s? kB/s?

So, a Gigabit ethernet card has a rate of 1073741824000 bits a second ?
cool

Mike


--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

On Tuesday 12 June 2007 02:56, Mark Reitblatt wrote:
> On 6/11/07, Alex Jones wrote:
> > Fine. Stick with Kilobytes, but strictly define it as 10^3 bytes. Just
> > choose one over the other and be consistent.
>
> That's not "consistent". Kilobyte has always meant 2^10 bytes. "kilo"
> in "kilobyte" is not an SI prefix. SI prefixes only apply to SI
> measurements, of which "byte" is not a member. There is no confusion;
> the only place where a kilobyte != 2^10 bytes is in hard drive
> manufacturer's advertising materials. This is the way it has been for
> decades, and it is a perfectly acceptable and desirable standard.

That's an argument that's been heard before but it's *wrong*. SI prefixes
*are* used with non-SI units without losing their normal meaning and there is
no reason why bytes should be an exception. Since kilo has always meant 1000,
kilobyte must initially have meant 1000 bytes, before people started to use
it as if to mean 1024. There is confusion; hard drive manufacturers'
advertising material is not the only place where kilobyte != 2^10 bytes.

--
Magnus Holmgren holmgren@lysator.liu.se
(No Cc of list mail needed, thanks)

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQBGbj74k7mRNn1h4+YRAguQAKCst/NypLs1AhFplMILQbORZQWipwCgi0/O
57RfWBrS6WPcSrm4pZDJ6yI=
=B0uR
-----END PGP SIGNATURE-----

On Monday 11 June 2007 23:10, Hendrik Sattler wrote:
> Abbreviations are ambiguous by design. Who actually says that KB means
> kilobyte?

You're arguing that although IEC prefixes eliminate all ambiguity in the area
of amounts and rates of data, there is still some ambiguity left, i.e. IEC
prefixes have to be rejected for not being a perfect solution.

http://en.wikipedia.org/wiki/Perfect_solution_fallacy

> I don't like those Special Interest units in all situations

SI units aren't special. They are universal.

--
Magnus Holmgren holmgren@lysator.liu.se
(No Cc of list mail needed, thanks)

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQBGbj35k7mRNn1h4+YRAvdQAJ42nkbmFUSYWgPfJV1qOA OjMzt2yACg45HE
5rE1SneeX1vaWk/aX401kYk=
=xBbs
-----END PGP SIGNATURE-----

On Tue, Jun 12, 2007 at 08:36:39AM +0200, Magnus Holmgren wrote:
>
> That's an argument that's been heard before but it's *wrong*. SI prefixes
> *are* used with non-SI units without losing their normal meaning and there is
> no reason why bytes should be an exception. Since kilo has always meant 1000,
> kilobyte must initially have meant 1000 bytes, before people started to use
> it as if to mean 1024. There is confusion; hard drive manufacturers'
> advertising material is not the only place where kilobyte != 2^10 bytes.
>
If I remember my history of computing correctly, kilo was not chosen to
mean exactly 1000 when it came to computers. Things were initially done
in powers of 2 (oversimplification). Since 2^10 = 1024 ≈ 1000, kilo was
chosen as the prefix to use, since it already existed. The idea of
going back and redifining the kilo to mean exactly 1000 in the context
of computing was a marketing gimmick.

Besides, there are other units of measure which carry the same name and
have different numerical values based on context (think statute miles
and nautical miles), though I don't think any such examples can be found
in the SI.

Regards,

-Roberto

--
Roberto C. Sánchez
http://people.connexer.com/~roberto
http://www.connexer.com

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQFGbkC/1snWssAFC08RAur5AJ9+QoRTCttaqC/vd2wKMRqlJbt5KACfS1If
+iJFRQRB5fMjMJqoCLWLhxM=
=YxmL
-----END PGP SIGNATURE-----

Hi Thijs!

You wrote:

> We are talking about tools like aptitude here, or at least, the OP does.
> Did you ever have 2 GB free and decided to install a package that would
> exactly fill that space in?

Afaik, we are talking about making the use of the prefixes consistent
over all of Debian, so that everywhere the program says MB, you know
exactly what that means. Doing everywhere except in apt would kind of
defeat the purpose, because then you still can't be sure...

Bas.

--
+--------------------------------------------------------------+
| Bas Zoetekouw | Sweet day, so cool, so calm, so bright, |
|--------------------| The bridall of the earth and skie: |
| bas@zoetekouw.net | The dew shall weep thy fall tonight; |
+--------------------| For thou must die. |
+-----------------------------------------+


--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

Magnus Holmgren writes:
>> No it doesn't.
>>
>> The "SI binary prefixes" are an abomination.
>
> Why - besides pronunciation?

Well among other things, the end result of this whole mess will likely
be to _increase_ confusion, rather than lessen it:

Until now, in a typical computer app, "900K" had an unambiguous meaning:
900*1024.

Now that a bunch of people are all in a misguided frenzy to "correct"
things (which weren't broken), there will almost certainly be cases
where some silly fool will change the _calculation_ but not the label
(e.g., in a case where space is at a premium) -- e.g., they'll keep "K",
but change the calculation to "/ 1000", because that's "correct".

However it's _guaranteed_ that many apps will stick with "/ 1024".

Thus a suffix (e.g. K) which was previously unambiguous in practice
(and no, disk-drive advertisements don't count) will have become more
ambiguous in practice.

[At which point of course the kib-fans will start crying out that
because things are now very confused, we muuuuuuust switch to "gib"
suffixes -- all because they basically screwed things up.]

Great.

-Miles

--
`To alcohol! The cause of, and solution to,
all of life's problems' --Homer J. Simpson


--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

On Tuesday 12 June 2007 08:44, Roberto C. Sánchez wrote:
> On Tue, Jun 12, 2007 at 08:36:39AM +0200, Magnus Holmgren wrote:
> > That's an argument that's been heard before but it's *wrong*. SI prefixes
> > *are* used with non-SI units without losing their normal meaning and
> > there is no reason why bytes should be an exception. Since kilo has
> > always meant 1000, kilobyte must initially have meant 1000 bytes, before
> > people started to use it as if to mean 1024. There is confusion; hard
> > drive manufacturers' advertising material is not the only place where
> > kilobyte != 2^10 bytes.
>
> If I remember my history of computing correctly, kilo was not chosen to
> mean exactly 1000 when it came to computers. Things were initially done
> in powers of 2 (oversimplification). Since 2^10 = 1024 ≈ 1000,kilo was
> chosen as the prefix to use, since it already existed. The idea of
> going back and redifining the kilo to mean exactly 1000 in the context
> of computing was a marketing gimmick.

It is possible that nobody used the word "kilobyte" before it became
conventional to use it to mean "1024 bytes", but if they did, it must have
meant "1000 bytes", by default.

> Besides, there are other units of measure which carry the same name and
> have different numerical values based on context (think statute miles
> and nautical miles), though I don't think any such examples can be found
> in the SI.

Of course not. The utter mess of miles, gallons, tons, pounds
troy/avoirdupois, and so on, is completely irrelevant.

--
Magnus Holmgren holmgren@lysator.liu.se
(No Cc of list mail needed, thanks)

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQBGbkYhk7mRNn1h4+YRAvrLAJ97Q72AufLqo+CuFPKsKb m1NRf4YgCfX0H8
C8F2tD1RW/peoJ4R/82hEYY=
=OCDF
-----END PGP SIGNATURE-----

On Monday 11 June 2007 21:21, Joey Hess wrote:
> Bastian Venthur wrote:
> > I agree with the "sounds stupid" part, although I don't belive this is a
> > valid argument.
>
> It's a perfectly valid argument for me to use to ignore a bad standard.
> If the standard makes me talk funny, I will ignore it or make fun of it.

Most of the time you won't have to say it. Spoken language tends to be less
formal than written language, and 2^10 bytes still is approximately a
kilobyte (and so on up to giga, where the approximation starts to fail). So
you only have to say kibibyte when you need to be precise.

--
Magnus Holmgren holmgren@lysator.liu.se
(No Cc of list mail needed, thanks)

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQBGbkeZk7mRNn1h4+YRAuWRAKDabxKRhabp4mumibFX8n ZbOJf+1ACdHTUh
iqw/iuNtipjeqn6KtDwAVQk=
=r1mP
-----END PGP SIGNATURE-----

On Tue, Jun 12, 2007 at 03:54:25PM +0900, Miles Bader wrote:
> Until now, in a typical computer app, "900K" had an unambiguous meaning:
> 900*1024.

No, its 900 Kelvin aka 626.85°C

Should I say that kb and Mb are kilo bases and mega bases, as in DNA?

Bastian

--
The sight of death frightens them [Earthers].
-- Kras the Klingon, "Friday's Child", stardate 3497.2


--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

On Tuesday 12 June 2007 08:54, Miles Bader wrote:
> Magnus Holmgren writes:
> >> No it doesn't.
> >>
> >> The "SI binary prefixes" are an abomination.
> >
> > Why - besides pronunciation?
>
> Well among other things, the end result of this whole mess will likely
> be to _increase_ confusion, rather than lessen it:

In that case it's not an end result.

--
Magnus Holmgren holmgren@lysator.liu.se
(No Cc of list mail needed, thanks)

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQBGbktVk7mRNn1h4+YRArTSAJ9TVKOii3FwGOBadm1mJc YkDrOFPACgrjzw
b/zEX2Ozk06dv6GRDKHWz5c=
=ZutC
-----END PGP SIGNATURE-----