Author Topic: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?  (Read 20450 times)

0 Members and 1 Guest are viewing this topic.

Online David Hess

  • Super Contributor
  • ***
  • Posts: 16545
  • Country: us
  • DavidH
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #25 on: September 15, 2014, 03:45:01 am »
Let me know when I can buy 32.768k x 8 EPROMs with a part number something like 27C256144.
 

Offline rs20

  • Super Contributor
  • ***
  • Posts: 2317
  • Country: au
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #26 on: September 15, 2014, 04:04:34 am »
Let me know when I can buy 32.768k x 8 EPROMs with a part number something like 27C256144.

... or just don't be dumb and buy 32KiB x 8 EPROMs with the same part number it's always been.  Yours is just ridiculous logic.
 

Offline miguelvp

  • Super Contributor
  • ***
  • Posts: 5550
  • Country: us
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #27 on: September 15, 2014, 06:04:00 am »
Let me know when I can buy 32.768k x 8 EPROMs with a part number something like 27C256144.

... or just don't be dumb and buy 32KiB x 8 EPROMs with the same part number it's always been.  Yours is just ridiculous logic.

I don't see any mention of KiB in the datasheet

http://download.siliconexpert.com/pdfs/2011/10/12/23/41/53/625/atm_/manual/at27c256r-15jc.pdf

It does mention "organized as 32K by 8 bits."

Edit: which is good because at work we don't use KB we use just K and we all know it means 1024
« Last Edit: September 15, 2014, 06:06:39 am by miguelvp »
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 16545
  • Country: us
  • DavidH
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #28 on: September 15, 2014, 06:43:23 am »
Let me know when I can buy 32.768k x 8 EPROMs with a part number something like 27C256144.

... or just don't be dumb and buy 32KiB x 8 EPROMs with the same part number it's always been.  Yours is just ridiculous logic.

I picked an old well known old part as a simple example but it applies just as well to new ones.  There is a distinct lack of 67.108864 Mbit SRAM and 4.294967296 Gbit DRAM on the market.
 

Offline rs20

  • Super Contributor
  • ***
  • Posts: 2317
  • Country: au
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #29 on: September 15, 2014, 06:58:49 am »
Let me know when I can buy 32.768k x 8 EPROMs with a part number something like 27C256144.

... or just don't be dumb and buy 32KiB x 8 EPROMs with the same part number it's always been.  Yours is just ridiculous logic.

I picked an old well known old part as a simple example but it applies just as well to new ones.  There is a distinct lack of 67.108864 Mbit SRAM and 4.294967296 Gbit DRAM on the market.

And there's a whole lot of 64 Mib SRAM and 4 Gib DRAM on the market. What's your point? You seem to think that you're being forced to use base-1000 notation when you're actually just being asked to disambiguate whenever you use base-1024. I think taking the latter option, by pressing the 'i' key on your keyboard, is a lot more sensible than stating the capacity in base-1000 with 9 decimal places.  Also, since JEDEC oversees SRAM and DRAM, feel free to adopt JEDEC's standards and call it 64 Mb SRAM as a third option, I'm not going to pedantic and say that that's wrong.  Again, where you go wrong is if you were to think that JEDEC's standards are more relevant and applicable than the IEC's in, let's say, every single context other than RAM and flash chips.
 

Offline lpc32

  • Frequent Contributor
  • **
  • Posts: 454
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #30 on: September 15, 2014, 11:47:06 am »
I don't use it because it sounds stupid. If a new notation, my vote is for MB^2 and MB^10, etc.

I never got the cries about "HDD manufacturers are cheating". Base 10 is the most natural one by default, and it's always been used by HDDs.

 

Offline rs20

  • Super Contributor
  • ***
  • Posts: 2317
  • Country: au
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #31 on: September 15, 2014, 01:23:27 pm »
I don't use it because it sounds stupid. If a new notation, my vote is for MB^2 and MB^10, etc.

I never got the cries about "HDD manufacturers are cheating". Base 10 is the most natural one by default, and it's always been used by HDDs.

Even I (a strong supported of KiB, MiB, etc) don't particularly like saying or writing "kibibyte". But I think that's fine; written language is often more precise and unambiguous than spoken English (where there's a chance for the listener to clarify). So if I'm taking notes at a meeting, and someone says "64 megabyte file", I'll write down "64 MiB file" (if that's what was meant). When asked to read it back out, I'll say "64 megabyte files". No-one has to say weird words, but the record is completely unambiguous.

The question of notation is completely orthogonal to the question of pronounciation/how it sounds. So I don't really understand your first two sentences.

I fully concur that while HDD manufacturers were doing it for marketing reasons, they're also doing exactly the right thing.
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 16545
  • Country: us
  • DavidH
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #32 on: September 15, 2014, 05:55:58 pm »
I don't use it because it sounds stupid. If a new notation, my vote is for MB^2 and MB^10, etc.

I never got the cries about "HDD manufacturers are cheating". Base 10 is the most natural one by default, and it's always been used by HDDs.

The HDD and other mass storage manufacturers did *not* always use base 10.  I distinctly remember when they made the transition one manufacturer at a time.  There were even heated debates about it online (such as it existed) with the more cynical concluding it was for marketing reasons.  Not long after the change, the whole "formatted versus unformatted" capacity meme was used to justify the difference in numbers even though the discrepancy was insignificant before the change to base 10.

This was all before Eternal September and the rise of the web so supporting documentation is largely offline in paper form.
 

Offline Kjelt

  • Super Contributor
  • ***
  • Posts: 6459
  • Country: nl
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #33 on: September 16, 2014, 02:16:38 pm »
Kilobyte will always mean 1024. To ensure that the new definition fails I started a rival standard where kibibyte = 1000 bytes. Confusion, uncertainty and doubt will destroy it.
nonsense the same argument would then go for kilogram, I can assure you that is never going to change anymore. Better get used to it or make some mistakes in the near future because a smarter colleague did use the right abbreviation.
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 16545
  • Country: us
  • DavidH
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #34 on: September 16, 2014, 03:46:06 pm »
Before the IEC got involved, one interesting place where I ran across 1000 versus 1024 is in oscilloscope record lengths but for a different reason it was irrelevant.

All of my DSOs have record lengths that are multiplies of 1024 which would be awkward without a power-of-2 number of points per display division (*) but dividing the standard 1-2-5 sequence into a power-of-2 would also be awkward.  Who wants 31.25, 62.5, or 156.3 nanoseconds per point?

Their solution was to use a power-of-10 number of points per displayed division, 50 or 100 (254 points per inch!) is typical but I have seen modern lower resolution (!) displays with 20 or 25, so 10 divisions of 50 or 100 points yields 500 or 1000 points across the display width and the extra 24 points from the power-of-2 waveform record are displayed through overscan which is rather oddly convenient even if it initially looks like an error.

The only place I have been tripped up by 1000 versus 1024 is when dealing with interfaces where the sample rates were not explicitly specified and they used 1024 or some mixture of 1000 and 1024 instead of 1000.  I am used to dealing with non-power-of-2 data length in power-of-2 length buffers or the reverse to avoid associative cache problems.

(*) Some modern DSOs suffer from aliasing produced when the acquisition record is displayed which is doubly annoying when they advertise huge record lengths as the solution to aliasing.
 

Offline Kjelt

  • Super Contributor
  • ***
  • Posts: 6459
  • Country: nl
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #35 on: September 17, 2014, 12:22:57 pm »
Instead of trying to re-write history and millions of pages of documentation and code, I'm just using KiB to mean multiples of 1000 and keeping the original meaning of KB.
You don't need to rewrite history to be forward compatible you just say from now on it is no longer backwards compatible.
So if you read pre 2009 documents you have to think for yourself if kB is 1000 or 1024 (which was screwed up many times in the past anyway) and for newer documentation just hold on to the standard.
You know if you read roman literature that IV means the numeral 4 and that it means the Initialization vector if you are reading about aes cbc mode encryption.
What you are doing is saying oh darn we were wrong in the past but lets continue to be wrong in the future and really be wrong this time since it has another value.
 

Offline amyk

  • Super Contributor
  • ***
  • Posts: 8240
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #36 on: September 17, 2014, 12:52:12 pm »
I don't use it because it sounds stupid. If a new notation, my vote is for MB^2 and MB^10, etc.

I never got the cries about "HDD manufacturers are cheating". Base 10 is the most natural one by default, and it's always been used by HDDs.

The HDD and other mass storage manufacturers did *not* always use base 10.  I distinctly remember when they made the transition one manufacturer at a time.  There were even heated debates about it online (such as it existed) with the more cynical concluding it was for marketing reasons.  Not long after the change, the whole "formatted versus unformatted" capacity meme was used to justify the difference in numbers even though the discrepancy was insignificant before the change to base 10.

This was all before Eternal September and the rise of the web so supporting documentation is largely offline in paper form.
The 10MB drive in the IBM XT was 306 cylinders, 4 heads, and 17 sectors/track, holding 10653696 bytes. This would be "10.16MB" or even "10.2MB" if they were marketed today.

Manufacturers tended to be even more generous back then - a common size was the 615/4/26 "30MB" drive, that actually contained slightly over 31MB.

If anyone made a 1TB drive that actually contained 2^40 bytes (or more), and marketed it as "1 true TB", at the same price as other "1TB" drives, I'd bet they could gain a lot of marketshare. The difference between a binary/decimal TB is 99,511,627,776 bytes - more than the size of an entire drive several years ago, and good for storing a huge amount more data. They could be clever and say things like "now with bonus 90GB for the same price"!
« Last Edit: September 17, 2014, 01:00:35 pm by amyk »
 

Offline rs20

  • Super Contributor
  • ***
  • Posts: 2317
  • Country: au
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #37 on: September 17, 2014, 01:04:14 pm »
...and marketed it as "1 true TB"...

Also known as 1 TiB...
 

Offline lpc32

  • Frequent Contributor
  • **
  • Posts: 454
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #38 on: September 17, 2014, 01:36:04 pm »
Well, that's interesting. When did HDDs transition from MB2 to MB10?
« Last Edit: September 17, 2014, 01:37:40 pm by lpc32 »
 

Offline mikeselectricstuff

  • Super Contributor
  • ***
  • Posts: 13694
  • Country: gb
    • Mike's Electric Stuff
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #39 on: September 17, 2014, 02:26:20 pm »
I think a major problem is it just sounds silly.
Until the hard disk manufacturers' marketing dickheads messed things up, you could tell the meaning unambiguously from context - kilobyte = 1024, kilo<anything else> = 1000
 
 
Youtube channel:Taking wierd stuff apart. Very apart.
Mike's Electric Stuff: High voltage, vintage electronics etc.
Day Job: Mostly LEDs
 

Offline SirNick

  • Frequent Contributor
  • **
  • Posts: 589
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #40 on: September 17, 2014, 07:44:22 pm »
A "fix" for a non-problem that invites more confusion (because existing material uses the "wrong" term now) is not a fix for anything.  K, M, G, T, etc., are meant to provide "orders of magnitude" values.  If you want exact amounts, don't use a multiplier.  Specify in bits, bytes, pixels, whatever.

ls -l vs. ls -lh.

There.  No ambiguity.  1,000,000 bytes is 1,000,000 bytes in everyone's parlance.  If you say 1MB, I know you mean "approximately 1 million bytes -- give or take".  Will it fit on a 1MB disk?  I dunno.  Specify that in bytes and see which is larger.  Problem solved, can we stop being ridiculously pedantic now?  ;)
 

Offline Kjelt

  • Super Contributor
  • ***
  • Posts: 6459
  • Country: nl
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #41 on: September 17, 2014, 08:05:41 pm »
  K, M, G, T, etc., are meant to provide "orders of magnitude" values.
K? Oh you mean k.  ;)
 

Offline SirNick

  • Frequent Contributor
  • **
  • Posts: 589
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #42 on: September 17, 2014, 09:01:25 pm »
I'll kut you.  ;D
 

Offline rs20

  • Super Contributor
  • ***
  • Posts: 2317
  • Country: au
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #43 on: September 18, 2014, 12:42:14 am »
I think a major problem is it just sounds silly.
Until the hard disk manufacturers' marketing dickheads messed things up, you could tell the meaning unambiguously from context - kilobyte = 1024, kilo<anything else> = 1000

Also network interfaces? One Mbps is unambiguously 1000000 bps, right? Is One MBps 1048576 Bps? It strikes me that the only thing that does use 1024 is binary address RAM and flash, and a few operating systems that have inappropriately extended this idea on to reporting HDD sizes?

A "fix" for a non-problem that invites more confusion (because existing material uses the "wrong" term now) is not a fix for anything.  K, M, G, T, etc., are meant to provide "orders of magnitude" values.  If you want exact amounts, don't use a multiplier.  Specify in bits, bytes, pixels, whatever.

ls -l vs. ls -lh.

There.  No ambiguity.  1,000,000 bytes is 1,000,000 bytes in everyone's parlance.  If you say 1MB, I know you mean "approximately 1 million bytes -- give or take".  Will it fit on a 1MB disk?  I dunno.  Specify that in bytes and see which is larger.  Problem solved, can we stop being ridiculously pedantic now?  ;)

Not sure if trolling, but this is ridiculously inconsistent. You'd be the sort of person that opposed the introduction of the metric system; that caused *real* problems, but I think we can agree it was the right thing to do in the long run. Make up your mind, is precision important or not?  If you think that it's a good idea for G to unavoidably imply a +/- 10% error, why are you getting so upset about existing material using the "wrong" term now? The existing material was always vague, right, so what difference does it make?

This is one thing that confuses me about all this whinging. The current situation is pretty messed up, but we all get by, I concur. All the IEC is saying is, "here, take this extra available set of prefixes: MiB, GiB, etc". They shall always means precisely 1024, 1048576. What MB means now is completely fuzzy, messed up, and context specific, and of course it will remain so. Anyone who requests a 4MB page from linux and expects a 4,000,000 byte page is an idiot. But what's the problem with introducing a new, parallel, unambiguous option? Saying that kB and MB "will get messed up" is just :palm:
 

jucole

  • Guest
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #44 on: September 18, 2014, 01:32:10 pm »
I don't like KibiByte but then I'm not really one for arranging my food tins in the cupboard with 10mm spacing.
 

Offline rs20

  • Super Contributor
  • ***
  • Posts: 2317
  • Country: au
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #45 on: September 18, 2014, 01:46:35 pm »
I don't like KibiByte but then I'm not really one for arranging my food tins in the cupboard with 10mm spacing.

Real nice. Run out of logic, resort to name-calling.
 

Offline Kjelt

  • Super Contributor
  • ***
  • Posts: 6459
  • Country: nl
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #46 on: September 18, 2014, 01:54:20 pm »
Real nice. Run out of logic, resort to name-calling.
+1  but seeing his nick I was afraid to hear something like "do you feel lucky Punk?"  :-DD
 

jucole

  • Guest
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #47 on: September 18, 2014, 04:41:57 pm »
I don't like KibiByte but then I'm not really one for arranging my food tins in the cupboard with 10mm spacing.

Real nice. Run out of logic, resort to name-calling.

yes because it defies logic;  so what are the benefits of adding a new set of SI units as opposed to making the hard disk manufacturers simply print a correct label?  Which in my case for the drive sat on my desk should just read 465.7 GB and not 500 GB.



 

jucole

  • Guest
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #48 on: September 18, 2014, 04:45:13 pm »
Real nice. Run out of logic, resort to name-calling.
+1  but seeing his nick I was afraid to hear something like "do you feel lucky Punk?"  :-DD

Classic film, but not right!  you're not thinking 4th dimensionally ;-)
 

Offline SirNick

  • Frequent Contributor
  • **
  • Posts: 589
Re: Will Kibibytes catch-on or can we turn things back to the JEDEC standard?
« Reply #49 on: September 18, 2014, 06:45:17 pm »
Not sure if trolling, but this is ridiculously inconsistent.
The second one.  By that I mean I'm trolling you.  (Or do I...?  Could go either way.)

You'd be the sort of person that opposed the introduction of the metric system
And you would be the sort of person who is wrong.  I would like to see it as a worldwide standard, which would mean changing some of my own habits.  I'm a'ight with that.  The difference is, an inch would still be an inch, not "supposedly an inch but maybe a cm".

Make up your mind, is precision important or not?
Of course it is.  That's why, when I need a precise number, I leave off the postfix, and state the real number.  The only time the value is still precise with a SI postfix is when it's a perfect multiple (1.0M), or you specify the value out to the required number of decimal places (32.768k).  If you have to specify more than two or three decimals, you probably shouldn't be shifting scales that lose precision.

But, there are ample times when I need to convey the general quantity of something -- like a 10k resistor, for example.  We understand it may not (and probably won't be) 10,000.000 ohms.  It has a tolerance, also specified if it matters, and we all agree to accept any error one way or the other.  And, we also understand that a 5% error on a 100R resistor is closer to the actual value than 5% on a 1M resistor.  Yet, no one is proposing new units for resistors...

If you think that it's a good idea for G to unavoidably imply a +/- 10% error, why are you getting so upset about existing material using the "wrong" term now? The existing material was always vague, right, so what difference does it make?
I'm not terribly upset about it being "wrong" now, I just find it unhelpful to say there's now an unambiguous unit for 2^x, but the 10^x unit is still just as ambiguous as ever -- except now, it's "supposed to be" accurate too.  The only way you'd know for sure is if someone left a footnote saying "actual MB, not MiB" -- and if it mattered either way, why not footnote with "MB = 1,000,000" as we're doing now?  In most scenarios, "MB = ~1 million" is good enough, so... as I said... it's a non-fix for a non-problem.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf