On 1/31/23 14:40, Richard Kettlewell wrote:
> Pancho <Pancho.Jones@proton.me> writes:
>> On 30/01/2023 18:10, Ken Blake wrote:
>>> So as far as I'm concerned, despite the existing standards, KB mean
>>> 1024, MB means 1024 x 1024, GB means 1024 x 1024 x 1024, etc. and KiB,
>>> MiB, Gib, etc. are almost never used and shouldn't be. A disk drive
>>> that's called 2GB should have 2,147,483,648 bytes, not 2,000,000,000.
>>
>> But...
>>
>> We use decimal for most other stuff, why would we want to use binary
>> for this special case? K means 10^3 not 2^10, M means 10^6, G means
>> 10^9. Why introduce complexity, unnecessary special cases?
>
> Well, it’s hardly ‘introduce’ any more, the convention is decades old.
>
>> What advantage do you think 2^10, 2^20 offers?
>
> Being able to talking about 16GB RAM (or 16GiB if you really must)
> instead of 17.179869184GB RAM.
>

RAM is never 16 GiB either, areas will be reserved by the OS. For
instance, 4 GiB on Windows XP 32 only had about 3.1 GiB available for my
use.

Floating point numbers cannot represent rational numbers like 1/3
precisely/concisely. We live with them for simplicity.

My point was that when I do a calculation for how much RAM I need, I use
decimal, ~17.2 * 10^9 B is the number I want to know, not the prettier
16 GiB.

If you support both systems, you encourage confusion.

Other than that, I agree with what Carlos is saying.

On 1/31/23 19:36, Paul wrote:
> On 1/31/2023 5:58 AM, Pancho wrote:
>
>>
>> But...
>>
>> We use decimal for most other stuff, why would we want to use binary
>> for this special case? K means 10^3 not 2^10, M means 10^6, G means 10^9.
>> Why introduce complexity, unnecessary special cases?
>>
>> What advantage do you think 2^10, 2^20 offers?
>
> Decoding logic is simpler when you use powers-of-two.
>
> This was important... a long time ago.
>
> In this example, someone uses a '139 to decode an address and select a
> device with it.
>
> https://blog.idorobots.org/media/upnod3/ram.png
>
>    Paul

Using Hex for memory addressing is sensible, but that is different.
Forty years ago I did that stuff, but not recently. I use high level
languages, and the kids I worked with understood far less than I did,
about that type of stuff, and didn't care.

Just as I don't know how to double declutch a car.

Pancho <Pancho.Jones@proton.me> writes:

> On 1/31/23 14:40, Richard Kettlewell wrote:
>> Pancho <Pancho.Jones@proton.me> writes:
>>> On 30/01/2023 18:10, Ken Blake wrote:
>>>> So as far as I'm concerned, despite the existing standards, KB mean
>>>> 1024, MB means 1024 x 1024, GB means 1024 x 1024 x 1024, etc. and KiB,
>>>> MiB, Gib, etc. are almost never used and shouldn't be. A disk drive
>>>> that's called 2GB should have 2,147,483,648 bytes, not 2,000,000,000.
>>>
>>> But...
>>>
>>> We use decimal for most other stuff, why would we want to use binary
>>> for this special case? K means 10^3 not 2^10, M means 10^6, G means
>>> 10^9. Why introduce complexity, unnecessary special cases?
>> Well, it’s hardly ‘introduce’ any more, the convention is decades
>> old.
>>
>>> What advantage do you think 2^10, 2^20 offers?
>> Being able to talking about 16GB RAM (or 16GiB if you really must)
>> instead of 17.179869184GB RAM.
>
> RAM is never 16 GiB either, areas will be reserved by the OS. For
> instance, 4 GiB on Windows XP 32 only had about 3.1 GiB available for
> my use.

Why did you bother to ask the question when you’re not going to accept
the answer?

--
https://www.greenend.org.uk/rjk/

On 01/02/2023 08:35, Richard Kettlewell wrote:
> Pancho <Pancho.Jones@proton.me> writes:
>
>> On 1/31/23 14:40, Richard Kettlewell wrote:
>>> Pancho <Pancho.Jones@proton.me> writes:
>>>> On 30/01/2023 18:10, Ken Blake wrote:
>>>>> So as far as I'm concerned, despite the existing standards, KB mean
>>>>> 1024, MB means 1024 x 1024, GB means 1024 x 1024 x 1024, etc. and KiB,
>>>>> MiB, Gib, etc. are almost never used and shouldn't be. A disk drive
>>>>> that's called 2GB should have 2,147,483,648 bytes, not 2,000,000,000.
>>>>
>>>> But...
>>>>
>>>> We use decimal for most other stuff, why would we want to use binary
>>>> for this special case? K means 10^3 not 2^10, M means 10^6, G means
>>>> 10^9. Why introduce complexity, unnecessary special cases?
>>> Well, it’s hardly ‘introduce’ any more, the convention is decades
>>> old.
>>>
>>>> What advantage do you think 2^10, 2^20 offers?
>>> Being able to talking about 16GB RAM (or 16GiB if you really must)
>>> instead of 17.179869184GB RAM.
>>
>> RAM is never 16 GiB either, areas will be reserved by the OS. For
>> instance, 4 GiB on Windows XP 32 only had about 3.1 GiB available for
>> my use.
>
> Why did you bother to ask the question when you’re not going to accept
> the answer?
>

Seriously? "The answer".

When I see something I consider bad practice, I always ask why? What
benefit does it offer?

Quite often I have missed something important, other times I haven't.

Don't you do that? How else do we understand stuff?

Anyway, I wasn't even rejecting your answer. I just didn't feel the
concise notation merited the complexity of a special case.

Pancho <Pancho.Jones@proton.me> writes:
> On 01/02/2023 08:35, Richard Kettlewell wrote:
>> Pancho <Pancho.Jones@proton.me> writes:
>>> On 1/31/23 14:40, Richard Kettlewell wrote:
>>>> Pancho <Pancho.Jones@proton.me> writes:
>>>>> What advantage do you think 2^10, 2^20 offers?
>>>> Being able to talking about 16GB RAM (or 16GiB if you really must)
>>>> instead of 17.179869184GB RAM.
>>>
>>> RAM is never 16 GiB either, areas will be reserved by the OS. For
>>> instance, 4 GiB on Windows XP 32 only had about 3.1 GiB available for
>>> my use.
>> Why did you bother to ask the question when you’re not going to
>> accept
>> the answer?
>
> Seriously? "The answer".
>
> When I see something I consider bad practice, I always ask why? What
> benefit does it offer?
>
> Quite often I have missed something important, other times I haven't.
>
> Don't you do that? How else do we understand stuff?
>
> Anyway, I wasn't even rejecting your answer. I just didn't feel the
> concise notation merited the complexity of a special case.

You asked why, I told you why, you argued with it, and are still doing
so.

--
https://www.greenend.org.uk/rjk/

On Tue, 31 Jan 2023 22:12:13 +0000, Pancho <Pancho.Jones@proton.me>
wrote:

>On 1/31/23 14:40, Richard Kettlewell wrote:
>> Pancho <Pancho.Jones@proton.me> writes:
>>> On 30/01/2023 18:10, Ken Blake wrote:
>>>> So as far as I'm concerned, despite the existing standards, KB mean
>>>> 1024, MB means 1024 x 1024, GB means 1024 x 1024 x 1024, etc. and KiB,
>>>> MiB, Gib, etc. are almost never used and shouldn't be. A disk drive
>>>> that's called 2GB should have 2,147,483,648 bytes, not 2,000,000,000.
>>>
>>> But...
>>>
>>> We use decimal for most other stuff, why would we want to use binary
>>> for this special case? K means 10^3 not 2^10, M means 10^6, G means
>>> 10^9. Why introduce complexity, unnecessary special cases?
>>
>> Well, it’s hardly ‘introduce’ any more, the convention is decades old.
>>
>>> What advantage do you think 2^10, 2^20 offers?
>>
>> Being able to talking about 16GB RAM (or 16GiB if you really must)
>> instead of 17.179869184GB RAM.
>>
>
>RAM is never 16 GiB either,

Yes it is.

> areas will be reserved by the OS. For
>instance, 4 GiB on Windows XP 32 only had about 3.1 GiB available for my
>use.

Yes, but how much RAM there is and how much is available for your use
are two different things. The amount of RAM installed on your computer
is 4GB, not 3.1GB.

For example If I go to System>About, here under Windows 11, it says
Installed RAM 32.0 GB (31.8 usable). How much is installed and how
much is usable are two different things.

On Tue, 31 Jan 2023 21:44:02 +0100, "Carlos E.R."
<robin_listas@es.invalid> wrote:

>On 2023-01-31 19:21, Ken Blake wrote:
>> On Tue, 31 Jan 2023 16:54:57 +0100, "Carlos E. R."
>> <robin_listas@es.invalid> wrote:
>>
>>> On 2023-01-31 16:32, Ken Blake wrote:
>>>> On Tue, 31 Jan 2023 10:58:15 +0000, Pancho <Pancho.Jones@proton.me>
>>>> wrote:
>>>
>>>
>>>> My point, once again, is that when drive manufacturers use the
>>>> established standard for disk drives when almost the rest of the
>>>> computer world does it differently, it confuses people and is a bad
>>>> thing to do. In my view, this is a case where consistency is more
>>>> important than standards.
>>>
>>> Mine is that those computer people are doing it wrong, and the rest of
>>> the world is right.
>>
>> Going by the standards, you are of course correct.
>>
>> But it doesn't matter. What matters is what considered correct by most
>> people.
>>
>>> Computer people have to adapt and say 1 GiB instead og 1 GB.
>>
>> "Have to"? Not a chance. It will never happen. There's only one way to
>> get consistency and that's for the drive manufacturers to use the
>> common understanding of KB, MB, GB etc. that the rest of the computer
>> world uses.
>
>Not going to happen :-)

We agree on that.

>> You will have course say that the drive manufacturers shouldn't change
>> and the rest of the computer world should. That might be a good choice
>> if it were possible, but it's not. There are only a handful of drive
>> manufacturers, but *millions* of computer users. You're not going to
>> change those millions.
>
>Give it time, and teach units in schools.

Not going to happen :-)

>>> Hard disk
>>> people are doing it right since decades.
>>
>>
>> Technically, yes. Practically, no.
>>>
>>> Microsoft, typically, hates standards and goes against.
>>
>> We agree on that. Is Microsoft responsible for the common meanings of
>> MB, GB, TB, etc. being different from the standards and being used the
>> way they are? Probably.
>>
>> But it doesn't matter who is responsible. Whether you or I like it or
>> not (I also don't like it, but I have no real choice other than to
>> accept it), that's the way it is, and we are not going to change it.

On 2023-02-01 16:28, Ken Blake wrote:
> On Tue, 31 Jan 2023 21:44:02 +0100, "Carlos E.R."
> <robin_listas@es.invalid> wrote:
>
>> On 2023-01-31 19:21, Ken Blake wrote:
>>> On Tue, 31 Jan 2023 16:54:57 +0100, "Carlos E. R."
>>> <robin_listas@es.invalid> wrote:
>>>
>>>> On 2023-01-31 16:32, Ken Blake wrote:
>>>>> On Tue, 31 Jan 2023 10:58:15 +0000, Pancho <Pancho.Jones@proton.me>
>>>>> wrote:
>>>>
>>>>
>>>>> My point, once again, is that when drive manufacturers use the
>>>>> established standard for disk drives when almost the rest of the
>>>>> computer world does it differently, it confuses people and is a bad
>>>>> thing to do. In my view, this is a case where consistency is more
>>>>> important than standards.
>>>>
>>>> Mine is that those computer people are doing it wrong, and the rest of
>>>> the world is right.
>>>
>>> Going by the standards, you are of course correct.
>>>
>>> But it doesn't matter. What matters is what considered correct by most
>>> people.
>>>
>>>> Computer people have to adapt and say 1 GiB instead og 1 GB.
>>>
>>> "Have to"? Not a chance. It will never happen. There's only one way to
>>> get consistency and that's for the drive manufacturers to use the
>>> common understanding of KB, MB, GB etc. that the rest of the computer
>>> world uses.
>>
>> Not going to happen :-)
>
> We agree on that.
>
>
>>> You will have course say that the drive manufacturers shouldn't change
>>> and the rest of the computer world should. That might be a good choice
>>> if it were possible, but it's not. There are only a handful of drive
>>> manufacturers, but *millions* of computer users. You're not going to
>>> change those millions.
>>
>> Give it time, and teach units in schools.
>
> Not going to happen :-)

:-)

--
Cheers, Carlos.

On 2/1/23 15:25, Ken Blake wrote:
> On Tue, 31 Jan 2023 22:12:13 +0000, Pancho <Pancho.Jones@proton.me>
> wrote:
>
>> On 1/31/23 14:40, Richard Kettlewell wrote:
>>> Pancho <Pancho.Jones@proton.me> writes:
>>>> On 30/01/2023 18:10, Ken Blake wrote:
>>>>> So as far as I'm concerned, despite the existing standards, KB mean
>>>>> 1024, MB means 1024 x 1024, GB means 1024 x 1024 x 1024, etc. and KiB,
>>>>> MiB, Gib, etc. are almost never used and shouldn't be. A disk drive
>>>>> that's called 2GB should have 2,147,483,648 bytes, not 2,000,000,000.
>>>>
>>>> But...
>>>>
>>>> We use decimal for most other stuff, why would we want to use binary
>>>> for this special case? K means 10^3 not 2^10, M means 10^6, G means
>>>> 10^9. Why introduce complexity, unnecessary special cases?
>>>
>>> Well, it’s hardly ‘introduce’ any more, the convention is decades old.
>>>
>>>> What advantage do you think 2^10, 2^20 offers?
>>>
>>> Being able to talking about 16GB RAM (or 16GiB if you really must)
>>> instead of 17.179869184GB RAM.
>>>
>>
>> RAM is never 16 GiB either,
>
> Yes it is.
>
>> areas will be reserved by the OS. For
>> instance, 4 GiB on Windows XP 32 only had about 3.1 GiB available for my
>> use.
>
>
> Yes, but how much RAM there is and how much is available for your use
> are two different things. The amount of RAM installed on your computer
> is 4GB, not 3.1GB.
>

I care about what is usable to me, just like with disk storage.

> For example If I go to System>About, here under Windows 11, it says
> Installed RAM 32.0 GB (31.8 usable). How much is installed and how
> much is usable are two different things.
>

For me, it says 16 GB and 15.8 GB. I don't know if that 15.8 GB is
exactly 15.8 GiB or only accurate to 1 dp. Either way, the prettiness of
exactly 16 has gone.

The prettiness is gone, but I'm still left with the problem that if I
want to calculate the amount of usable RAM my software data structures
require, in GiB, I have to convert my natural decimal calculations to a
binary format, to avoid the 7.4% difference between the GiB, and the
more orthodox decimal GB. Maybe other programmer don't estimate memory
requirements, don't use algorithms that require a lot of memory? It
wouldn't surprise me, innumeracy is surprisingly high in IT.

Clinging to unnecessary complexity reminds me of the metric martyrs and
their insistence on using imperial weights and measures. When automating
some business process, you often see veterans of the industry try to
cling to unnecessary complexity. I guess if you remove the complexity,
the competitive advantage they have in understanding it disappears, they
are diminished.

As far as I can see, the US government and the standards organizations
have agreed on the decimal GB.

On Wed, 1 Feb 2023 22:21:22 +0000, Pancho <Pancho.Jones@proton.me>
wrote:

>On 2/1/23 15:25, Ken Blake wrote:
>> On Tue, 31 Jan 2023 22:12:13 +0000, Pancho <Pancho.Jones@proton.me>
>> wrote:
>>
>>> On 1/31/23 14:40, Richard Kettlewell wrote:
>>>> Pancho <Pancho.Jones@proton.me> writes:
>>>>> On 30/01/2023 18:10, Ken Blake wrote:
>>>>>> So as far as I'm concerned, despite the existing standards, KB mean
>>>>>> 1024, MB means 1024 x 1024, GB means 1024 x 1024 x 1024, etc. and KiB,
>>>>>> MiB, Gib, etc. are almost never used and shouldn't be. A disk drive
>>>>>> that's called 2GB should have 2,147,483,648 bytes, not 2,000,000,000.
>>>>>
>>>>> But...
>>>>>
>>>>> We use decimal for most other stuff, why would we want to use binary
>>>>> for this special case? K means 10^3 not 2^10, M means 10^6, G means
>>>>> 10^9. Why introduce complexity, unnecessary special cases?
>>>>
>>>> Well, it’s hardly ‘introduce’ any more, the convention is decades old.
>>>>
>>>>> What advantage do you think 2^10, 2^20 offers?
>>>>
>>>> Being able to talking about 16GB RAM (or 16GiB if you really must)
>>>> instead of 17.179869184GB RAM.
>>>>
>>>
>>> RAM is never 16 GiB either,
>>
>> Yes it is.
>>
>>> areas will be reserved by the OS. For
>>> instance, 4 GiB on Windows XP 32 only had about 3.1 GiB available for my
>>> use.
>>
>>
>> Yes, but how much RAM there is and how much is available for your use
>> are two different things. The amount of RAM installed on your computer
>> is 4GB, not 3.1GB.
>>
>
>I care about what is usable to me, just like with disk storage.

As do I. But what either of us cares about has nothing to do with how
much RAM 4GB is.

Pancho wrote on 2/2/23 9:21 am:
> On 2/1/23 15:25, Ken Blake wrote:
>> On Tue, 31 Jan 2023 22:12:13 +0000, Pancho
>> <Pancho.Jones@proton.me> wrote:
>>> On 1/31/23 14:40, Richard Kettlewell wrote:
>>>> Pancho <Pancho.Jones@proton.me> writes:
>>>>> On 30/01/2023 18:10, Ken Blake wrote:
>>>>>> So as far as I'm concerned, despite the existing standards,
>>>>>> KB mean 1024, MB means 1024 x 1024, GB means 1024 x 1024 x
>>>>>> 1024, etc. and KiB, MiB, Gib, etc. are almost never used
>>>>>> and shouldn't be. A disk drive that's called 2GB should
>>>>>> have 2,147,483,648 bytes, not 2,000,000,000.
>>>>>
>>>>> But...
>>>>>
>>>>> We use decimal for most other stuff, why would we want to use
>>>>> binary for this special case? K means 10^3 not 2^10, M means
>>>>> 10^6, G means 10^9. Why introduce complexity, unnecessary
>>>>> special cases?
>>>>
>>>> Well, it’s hardly ‘introduce’ any more, the convention is
>>>> decades old.
>>>>
>>>>> What advantage do you think 2^10, 2^20 offers?
>>>>
>>>> Being able to talking about 16GB RAM (or 16GiB if you really
>>>> must) instead of 17.179869184GB RAM.
>>>
>>> RAM is never 16 GiB either,
>>
>> Yes it is.
>>
>>> areas will be reserved by the OS. For instance, 4 GiB on Windows
>>> XP 32 only had about 3.1 GiB available for my use.
>>
>> Yes, but how much RAM there is and how much is available for your
>> use are two different things. The amount of RAM installed on your
>> computer is 4GB, not 3.1GB.
>
> I care about what is usable to me, just like with disk storage.

How much of that 3.1GB would be useful for you if the other 0.9GB were
not being used?? ;-P I'm guessing NONE!!
--
Daniel

On 2023-02-01 23:21, Pancho wrote:
> On 2/1/23 15:25, Ken Blake wrote:
>> On Tue, 31 Jan 2023 22:12:13 +0000, Pancho <Pancho.Jones@proton.me>
>> wrote:
>>
>>> On 1/31/23 14:40, Richard Kettlewell wrote:
>>>> Pancho <Pancho.Jones@proton.me> writes:
>>>>> On 30/01/2023 18:10, Ken Blake wrote:
>>>>>> So as far as I'm concerned, despite the existing standards, KB mean
>>>>>> 1024, MB means 1024 x 1024, GB means 1024 x 1024 x 1024, etc. and
>>>>>> KiB,
>>>>>> MiB, Gib, etc. are almost never used and shouldn't be.  A disk drive
>>>>>> that's called 2GB should have 2,147,483,648 bytes, not 2,000,000,000.
>>>>>
>>>>> But...
>>>>>
>>>>> We use decimal for most other stuff, why would we want to use binary
>>>>> for this special case? K means 10^3 not 2^10, M means 10^6, G means
>>>>> 10^9. Why introduce complexity, unnecessary special cases?
>>>>
>>>> Well, it’s hardly ‘introduce’ any more, the convention is decades old.
>>>>
>>>>> What advantage do you think 2^10, 2^20 offers?
>>>>
>>>> Being able to talking about 16GB RAM (or 16GiB if you really must)
>>>> instead of 17.179869184GB RAM.
>>>>
>>>
>>> RAM is never 16 GiB either,
>>
>> Yes it is.
>>
>>> areas will be reserved by the OS.  For
>>> instance, 4 GiB on Windows XP 32 only had about 3.1 GiB available for my
>>> use.
>>
>>
>> Yes, but how much RAM there is and how much is available for your use
>> are two different things. The amount of RAM installed on your computer
>> is 4GB, not 3.1GB.
>>
>
> I care about what is usable to me, just like with disk storage.
>
>> For example If I go to System>About, here under Windows 11, it says
>> Installed RAM 32.0 GB (31.8 usable). How much is installed and how
>> much is usable are two different things.
>>
>
> For me, it says 16 GB and 15.8 GB. I don't know if that 15.8 GB is
> exactly 15.8 GiB or only accurate to 1 dp. Either way, the prettiness of
> exactly 16 has gone.
>
> The prettiness is gone, but I'm still left with the problem that if I
> want to calculate the amount of usable RAM my software data structures
> require, in GiB, I have to convert my natural decimal calculations to a
> binary format, to avoid the 7.4% difference between the GiB, and the
> more orthodox decimal GB. Maybe other programmer don't estimate memory
> requirements, don't use algorithms that require a lot of memory? It
> wouldn't surprise me, innumeracy is surprisingly high in IT.
>
> Clinging to unnecessary complexity reminds me of the metric martyrs and
> their insistence on using imperial weights and measures. When automating
> some business process, you often see veterans of the industry try to
> cling to unnecessary complexity. I guess if you remove the complexity,
> the competitive advantage they have in understanding it disappears, they
> are diminished.
>
> As far as I can see, the US government and the standards organizations
> have agreed on the decimal GB.

If all software and docs stick to the units as described by the
standards organizations, there would be no doubts about what that "16 GB
and 15.8 GB" of yours actually means. If some one writes GB it is
decimal, or else he writes GiB. No need to second guess.

--
Cheers, Carlos.

On Thu, 2 Feb 2023 10:37:29 +0100, "Carlos E.R."
<robin_listas@es.invalid> wrote:

>On 2023-02-01 23:21, Pancho wrote:
>> On 2/1/23 15:25, Ken Blake wrote:
>>> On Tue, 31 Jan 2023 22:12:13 +0000, Pancho <Pancho.Jones@proton.me>
>>> wrote:
>>>
>>>> On 1/31/23 14:40, Richard Kettlewell wrote:
>>>>> Pancho <Pancho.Jones@proton.me> writes:
>>>>>> On 30/01/2023 18:10, Ken Blake wrote:
>>>>>>> So as far as I'm concerned, despite the existing standards, KB mean
>>>>>>> 1024, MB means 1024 x 1024, GB means 1024 x 1024 x 1024, etc. and
>>>>>>> KiB,
>>>>>>> MiB, Gib, etc. are almost never used and shouldn't be.  A disk drive
>>>>>>> that's called 2GB should have 2,147,483,648 bytes, not 2,000,000,000.
>>>>>>
>>>>>> But...
>>>>>>
>>>>>> We use decimal for most other stuff, why would we want to use binary
>>>>>> for this special case? K means 10^3 not 2^10, M means 10^6, G means
>>>>>> 10^9. Why introduce complexity, unnecessary special cases?
>>>>>
>>>>> Well, it’s hardly ‘introduce’ any more, the convention is decades old.
>>>>>
>>>>>> What advantage do you think 2^10, 2^20 offers?
>>>>>
>>>>> Being able to talking about 16GB RAM (or 16GiB if you really must)
>>>>> instead of 17.179869184GB RAM.
>>>>>
>>>>
>>>> RAM is never 16 GiB either,
>>>
>>> Yes it is.
>>>
>>>> areas will be reserved by the OS.  For
>>>> instance, 4 GiB on Windows XP 32 only had about 3.1 GiB available for my
>>>> use.
>>>
>>>
>>> Yes, but how much RAM there is and how much is available for your use
>>> are two different things. The amount of RAM installed on your computer
>>> is 4GB, not 3.1GB.
>>>
>>
>> I care about what is usable to me, just like with disk storage.
>>
>>> For example If I go to System>About, here under Windows 11, it says
>>> Installed RAM 32.0 GB (31.8 usable). How much is installed and how
>>> much is usable are two different things.
>>>
>>
>> For me, it says 16 GB and 15.8 GB. I don't know if that 15.8 GB is
>> exactly 15.8 GiB or only accurate to 1 dp. Either way, the prettiness of
>> exactly 16 has gone.
>>
>> The prettiness is gone, but I'm still left with the problem that if I
>> want to calculate the amount of usable RAM my software data structures
>> require, in GiB, I have to convert my natural decimal calculations to a
>> binary format, to avoid the 7.4% difference between the GiB, and the
>> more orthodox decimal GB. Maybe other programmer don't estimate memory
>> requirements, don't use algorithms that require a lot of memory? It
>> wouldn't surprise me, innumeracy is surprisingly high in IT.
>>
>> Clinging to unnecessary complexity reminds me of the metric martyrs and
>> their insistence on using imperial weights and measures. When automating
>> some business process, you often see veterans of the industry try to
>> cling to unnecessary complexity. I guess if you remove the complexity,
>> the competitive advantage they have in understanding it disappears, they
>> are diminished.
>>
>> As far as I can see, the US government and the standards organizations
>> have agreed on the decimal GB.
>
>If all software and docs stick to the units as described by the
>standards organizations, there would be no doubts about what that "16 GB
>and 15.8 GB" of yours actually means. If some one writes GB it is
>decimal, or else he writes GiB. No need to second guess.

I completely agree. But alternatively if all software and docs would
stick to the powers of two definitions of GB, etc. there would be no
doubt about what KB, MB, GB, TB, etc. meant. What's most important is
consistency in the computer world, not what definitions are used.

I don't think we'll ever get consistency.

On 2023-02-02 16:41, Ken Blake wrote:
> On Thu, 2 Feb 2023 10:37:29 +0100, "Carlos E.R."
> <robin_listas@es.invalid> wrote:
>
>> On 2023-02-01 23:21, Pancho wrote:
>>> On 2/1/23 15:25, Ken Blake wrote:
>>>> On Tue, 31 Jan 2023 22:12:13 +0000, Pancho <Pancho.Jones@proton.me>
>>>> wrote:
>>>>
>>>>> On 1/31/23 14:40, Richard Kettlewell wrote:
>>>>>> Pancho <Pancho.Jones@proton.me> writes:
>>>>>>> On 30/01/2023 18:10, Ken Blake wrote:
>>>>>>>> So as far as I'm concerned, despite the existing standards, KB mean
>>>>>>>> 1024, MB means 1024 x 1024, GB means 1024 x 1024 x 1024, etc. and
>>>>>>>> KiB,
>>>>>>>> MiB, Gib, etc. are almost never used and shouldn't be.  A disk drive
>>>>>>>> that's called 2GB should have 2,147,483,648 bytes, not 2,000,000,000.
>>>>>>>
>>>>>>> But...
>>>>>>>
>>>>>>> We use decimal for most other stuff, why would we want to use binary
>>>>>>> for this special case? K means 10^3 not 2^10, M means 10^6, G means
>>>>>>> 10^9. Why introduce complexity, unnecessary special cases?
>>>>>>
>>>>>> Well, it’s hardly ‘introduce’ any more, the convention is decades old.
>>>>>>
>>>>>>> What advantage do you think 2^10, 2^20 offers?
>>>>>>
>>>>>> Being able to talking about 16GB RAM (or 16GiB if you really must)
>>>>>> instead of 17.179869184GB RAM.
>>>>>>
>>>>>
>>>>> RAM is never 16 GiB either,
>>>>
>>>> Yes it is.
>>>>
>>>>> areas will be reserved by the OS.  For
>>>>> instance, 4 GiB on Windows XP 32 only had about 3.1 GiB available for my
>>>>> use.
>>>>
>>>>
>>>> Yes, but how much RAM there is and how much is available for your use
>>>> are two different things. The amount of RAM installed on your computer
>>>> is 4GB, not 3.1GB.
>>>>
>>>
>>> I care about what is usable to me, just like with disk storage.
>>>
>>>> For example If I go to System>About, here under Windows 11, it says
>>>> Installed RAM 32.0 GB (31.8 usable). How much is installed and how
>>>> much is usable are two different things.
>>>>
>>>
>>> For me, it says 16 GB and 15.8 GB. I don't know if that 15.8 GB is
>>> exactly 15.8 GiB or only accurate to 1 dp. Either way, the prettiness of
>>> exactly 16 has gone.
>>>
>>> The prettiness is gone, but I'm still left with the problem that if I
>>> want to calculate the amount of usable RAM my software data structures
>>> require, in GiB, I have to convert my natural decimal calculations to a
>>> binary format, to avoid the 7.4% difference between the GiB, and the
>>> more orthodox decimal GB. Maybe other programmer don't estimate memory
>>> requirements, don't use algorithms that require a lot of memory? It
>>> wouldn't surprise me, innumeracy is surprisingly high in IT.
>>>
>>> Clinging to unnecessary complexity reminds me of the metric martyrs and
>>> their insistence on using imperial weights and measures. When automating
>>> some business process, you often see veterans of the industry try to
>>> cling to unnecessary complexity. I guess if you remove the complexity,
>>> the competitive advantage they have in understanding it disappears, they
>>> are diminished.
>>>
>>> As far as I can see, the US government and the standards organizations
>>> have agreed on the decimal GB.
>>
>> If all software and docs stick to the units as described by the
>> standards organizations, there would be no doubts about what that "16 GB
>> and 15.8 GB" of yours actually means. If some one writes GB it is
>> decimal, or else he writes GiB. No need to second guess.
>
>
> I completely agree. But alternatively if all software and docs would
> stick to the powers of two definitions of GB, etc. there would be no
> doubt about what KB, MB, GB, TB, etc. meant. What's most important is
> consistency in the computer world, not what definitions are used.

But there would be confusion to the many people for which K is 1000, and
have to deduce from context if this is 1000 or 1024.

Like disc manufacturers, always using 10^

>
> I don't think we'll ever get consistency.

I don't like bibytes units, I have been all my life doing 2^
calculations. But now that there is a standardization by the
organizations that do standards, I decided to accept it in full.like it
or not.

--
Cheers,
Carlos E.R.

Carlos E. R. wrote on 3/2/23 3:12 am:
> On 2023-02-02 16:41, Ken Blake wrote:

<Snip>

>> I completely agree. But alternatively if all software and docs would
>> stick to the powers of two definitions of GB, etc. there would be no
>> doubt about what KB, MB, GB, TB, etc. meant. What's most important is
>> consistency in the computer world, not what definitions are used.
>
> But there would be confusion to the many people for which K is 1000, and
> have to deduce from context if this is 1000 or 1024.
Should we all go on strike until Society accepts that the smaller
letter, 'k', represents the smaller number, 1000, and the larger letter
.. 'K', represents the larger number, 1024?? Etc., etc.!

Who's with me?? ;-P
--
Daniel

On Fri, 3 Feb 2023 22:24:01 +1100, Daniel65
<daniel47@nomail.afraid.org> wrote:

>Carlos E. R. wrote on 3/2/23 3:12 am:
>> On 2023-02-02 16:41, Ken Blake wrote:
>
><Snip>
>
>>> I completely agree. But alternatively if all software and docs would
>>> stick to the powers of two definitions of GB, etc. there would be no
>>> doubt about what KB, MB, GB, TB, etc. meant. What's most important is
>>> consistency in the computer world, not what definitions are used.
>>
>> But there would be confusion to the many people for which K is 1000, and
>> have to deduce from context if this is 1000 or 1024.
>Should we all go on strike until Society accepts that the smaller
>letter, 'k', represents the smaller number, 1000, and the larger letter
>. 'K', represents the larger number, 1024?? Etc., etc.!
>
>Who's with me?? ;-P

Not me. I don't think that's a great idea. It would be too hard to
recognize the difference and too hard to remember what's what.

K and k is essentially no different from KB and KiB, just spelled
differently.

On 2023-02-03 15:12, Ken Blake wrote:
> On Fri, 3 Feb 2023 22:24:01 +1100, Daniel65
> <daniel47@nomail.afraid.org> wrote:
>
>> Carlos E. R. wrote on 3/2/23 3:12 am:
>>> On 2023-02-02 16:41, Ken Blake wrote:
>>
>> <Snip>
>>
>>>> I completely agree. But alternatively if all software and docs would
>>>> stick to the powers of two definitions of GB, etc. there would be no
>>>> doubt about what KB, MB, GB, TB, etc. meant. What's most important is
>>>> consistency in the computer world, not what definitions are used.
>>>
>>> But there would be confusion to the many people for which K is 1000, and
>>> have to deduce from context if this is 1000 or 1024.
>> Should we all go on strike until Society accepts that the smaller
>> letter, 'k', represents the smaller number, 1000, and the larger letter
>> . 'K', represents the larger number, 1024?? Etc., etc.!
>>
>> Who's with me?? ;-P
>
>
> Not me. I don't think that's a great idea. It would be too hard to
> recognize the difference and too hard to remember what's what.

Right.

And the SI already says it is "k", lower case, which means "kilo". There
is no "K". So that would be a new change on the standards.

B and b are used for byte and bits, respectively. And we forget.

--
Cheers,
Carlos E.R.

On 03-02-2023 15:48, Carlos E. R. wrote:
> On 2023-02-03 15:12, Ken Blake wrote:
>> On Fri, 3 Feb 2023 22:24:01 +1100, Daniel65
>> <daniel47@nomail.afraid.org> wrote:
>>
>>> Carlos E. R. wrote on 3/2/23 3:12 am:
>>>> On 2023-02-02 16:41, Ken Blake wrote:
>>>
>>> <Snip>
>>>
>>>>> I completely agree. But alternatively if all software and docs would
>>>>> stick to the powers of two definitions of GB, etc. there would be no
>>>>> doubt about what KB, MB, GB, TB, etc. meant. What's most important is
>>>>> consistency in the computer world, not what definitions are used.
>>>>
>>>> But there would be confusion to the many people for which K is 1000,
>>>> and
>>>> have to deduce from context if this is 1000 or 1024.
>>> Should we all go on strike until Society accepts that the smaller
>>> letter, 'k', represents the smaller number, 1000, and the larger letter
>>> . 'K', represents the larger number, 1024?? Etc., etc.!
>>>
>>> Who's with me?? ;-P
>>
>>
>> Not me. I don't think that's a great idea. It would be too hard to
>> recognize the difference and too hard to remember what's what.
>
> Right.
>
> And the SI already says it is "k", lower case, which means "kilo". There
> is no "K". So that would be a new change on the standards.
>
> B and b are used for byte and bits, respectively. And we forget.
>
Furthermore, a lowercase m stands for milli (which, for bytes, seems
rather silly).

with <ZE9DL.424202$Tcw8.32549@fx10.iad> jjb wrote:
> On 03-02-2023 15:48, Carlos E. R. wrote:
>> On 2023-02-03 15:12, Ken Blake wrote:
>>> On Fri, 3 Feb 2023 22:24:01 +1100, Daniel65
>>> <daniel47@nomail.afraid.org> wrote:
>>>> Carlos E. R. wrote on 3/2/23 3:12 am:
>>>>> On 2023-02-02 16:41, Ken Blake wrote:

*SKIP*
>> And the SI already says it is "k", lower case, which means "kilo".
>> There is no "K". So that would be a new change on the standards.
>> B and b are used for byte and bits, respectively. And we forget.
> Furthermore, a lowercase m stands for milli (which, for bytes, seems
> rather silly).

Unless speeds or densities.

Also, </usr/share/misc/units.dat> is worth checking. Turns out 'K' is
already taken (it's Kelvin).

Also, I was musing about nice tangent: is byte primitive or derived?
If derived then does it come from mass and/or temperature? Turns out --
dead end, 'bit' is kinda primitive, 'byte' is derived. Such a loss :(

--
Torvalds' goal for Linux is very simple: World Domination
Stallman's goal for GNU is even simpler: Freedom