Jump to content

Talk:Binary prefix/Archive 7

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 1Archive 5Archive 6Archive 7Archive 8Archive 9Archive 10

JEDEC

I believe JEDEC is a minor organization compared to the ten or so IEC prefix supporters mentioned in this talk page (try the google test, for example), and is thus given undue weight in the article. I am fairly sure, given talk going on WP:MOSNUM talk, that any attempt to remedy the situation would be met with an edit war, so I'm just going to add a POV tag to the section describing what might be wrong with it. If I was pessimist with this, you have my permission to remove the tag together with the section. --SLi 20:12, 12 April 2007 (UTC)

Also see Wikipedia talk:Manual of Style (dates and numbers)#Binary prefixes straw poll and Wikipedia talk:Manual of Style (dates and numbers)#Binary prefixes vs JEDEC Standard for discussion of this issue. --SLi 20:23, 12 April 2007 (UTC)

I agree that the JEDEC section as presented is POV and therefore that the tag should not be removed. As a remedy I suggest that the section be shortened and an equally weighted IEEE section be added.Tom94022 21:23, 12 April 2007 (UTC)
Isn't the IEEE already covered by the IEC section? The only difference is the abbreviations for bit and byte, which can be covered in the IEC section. If this is decided to be undue weight, we should just remove the section, but I'm not totally convinced it is undue weight. — Omegatron 21:45, 12 April 2007 (UTC)
The difference between IEEE and JEDEC is glaring. Comparing them is really like comparing a Porsche and a Lada. IEEE is big and known, JEDEC is small and obscure. Presenting the two as somehow equals is biased. For example, from the Google test (which I agree is not entirely without problems): IEEE, 90.7M hits; JEDEC, 1.46M hits. See the difference between the two? --SLi 22:05, 12 April 2007 (UTC)
JEDEC is important in its own rite, but it certainly doesn't have the clout of IEEE. Heck, even IEEE isn't able to enforce preferred style standards (pick up a copy of Spectrum or an IEEE computer society publication and see if you can find 'mebi' :). My issue with all this JEDEC nonsense is simply that they never claim to define how terms should be used by others; they just follow good typical technical writing practices and define how they themselves use terms. How this can be taken as an endorsement of common usage or an active denial of the IEC prefixes is beyond my comprehension (especially in light of their acknowledgement in a recent standard). The IEEE has actually explicitly endorsed the IEC prefixes and explicitly recommended their usage by electrical and computer engineering professionals. -- mattb @ 2007-04-12T22:40Z


This is the second sentence of the Binary prefixes article. "When discussing things that naturally come in powers of two (such as computer memory sizes)." The organization that develops the standards for semiconductor memory should be relevant to the binary prefix article. You cannot purchase a computer that does not use JEDEC standard memory. -- SWTPC6800 01:55, 13 April 2007 (UTC)

This still doesn't address the issue that JEDEC doesn't, in fact or in theory, set unit conventions. -- mattb @ 2007-04-13T02:21Z
The companies that make binary things that Wikipedia readers buy use JEDEC standards. Someone might want to know why every memory module is labeled with 512MB or 1GB. -- SWTPC6800 15:20, 13 April 2007 (UTC)
Isn't that why we have section 5 "Usage notes"? Since JEDEC is a semiconductor trade organization and Standard 100B.01 applies the binary definition "as a prefix to units of semiconductor storage capacity" it seems that a short paragraph is appropriate for section 5.2. Perhaps we should rename 5.2 to Semiconductor storage instead of Computer memory :-)Tom94022 17:24, 13 April 2007 (UTC)

Some Background

"The JEDEC Solid State Technology Association (Once known as the Joint Electron Device Engineering Council), is the semiconductor engineering standardization body of the Electronic Industries Alliance (EIA), a trade association that represents all areas of the electronics industry." [1]

"The Electronic Industries Alliance (EIA) is a national trade organization that includes the full spectrum of U.S. manufacturers. The Alliance is a partnership of electronic and high-tech associations and companies whose mission is promoting the market development and competitiveness of the U.S. high-tech industry through domestic and international policy efforts. EIA, headquartered in Arlington, Va., comprises nearly 1,300 member companies whose products and services range from the smallest electronic components to the most complex systems used by defense, space and industry, including the full range of consumer electronic products." [2]

"Accredited by the American National Standards Institute (ANSI), EIA provides a forum for industry to develop standards and publications in our major technical areas: electronic components, consumer electronics, electronic information, telecommunications, and Internet security. The department also administrates the EIA Technology Council; a body of member companies that explores how emerging technologies potentially will affect segments of the electronics industry. The Council is also involved in creating a technology roadmap for the Alliance." [3]

The Electronic Industries Alliance (EIA) / JEDEC has agreement with the IEC [4] to produce Publicly Available Specifications [5]

Microprocessor companies like AMD [6] and Intel [7] are members of JEDEC and to ensure their processors can utilize the latest memory systems.

All of the semiconductor memory manufactures are members of JEDEC. Here is was Micron says about the upcoming DDR3 memory standard.

"Evaluation samples of Micron's leading-edge 1Gb DDR3 components are available to select customers with production expected to begin early next year. Micron's 1Gb DDR3 components will be available in various output configurations (x4, x8 and x16), and will be fully compliant to the most recent JEDEC DDR3 specifications (JEDEC is the leading developer of standards for the semiconductor industry). These components will support module densities from 512 megabytes (MB) through 4 gigabytes (GB) and a variety of module types including FBDIMMs, UDIMMs, SODIMMs, and RDIMMs." [8]

Why semiconductor companies use JEDEC

JEDEC is the standards organization of choice for semiconductor companies because for the flexibility and speed of the process. JEDEC Standard 21 for Configurations for Solid State Memories is published in a 3-ring binder that can be updated monthly. Each semiconductor company (and other members) gets a vote. In IEC each country gets a vote and the process takes years.

Solid State Times OCTOBER, 1999 Volume 1, Issue 4, Page 2 [9]

JEDEC has entered into an important new Memorandum of Understanding with the International Electrotechnical Commission (IEC) to expedite the process of moving semiconductor standards into the international marketplace. The agreement, which became effective September 1, 1999 is part of the IEC-PAS (Publicly Available Specifications) program. The program’s objective is to speed up the international standards process in areas of rapidly developing technology. It is intended to deliver information to the marketplace quickly since PAS are de facto standards. If PAS standards are approved and accepted by the international community, they ultimately will become IEC international standards.

Under the recently signed agreement, JEDEC will identify the specific standards it wishes to submit for PAS consideration. If the IEC accepts the JEDEC standard as a PAS to be entered into their Program of Work (PW), the standard will be published under the joint logos of JEDEC and the IEC, and royalties on the sale of the PAS will be divided equally between the two organizations. If the IEC later adopts the PAS as an IEC international standard, the standard will be published under the IEC’s sole logo and the division of royalties will end.

A significant feature of the new agreement is the fact that JEDEC will continue to have the right to continue developing the relevant technology, to revise and modify the underlying JEDEC standard, and to continue publishing the underlying JEDEC standard on the World Wide Web.

Solid State Times JANUARY, 2000 Volume 2, Issue 1, Page 10 [10]

The IEC works differently than the JEDEC committees where it is one country one vote instead of the JEDEC one company one vote. This voting practice does not favor the larger Semiconductor Countries over the smaller ones. Many smaller countries like the IEC because they can voice their opinion without being overpowered by the larger countries. The IEC normally meets twice per year to discuss new proposals and to review the present balloting results.

The IEC has 4 basic stages of balloting which include New Work Item Proposal (NP), Committee Draft (CD), Committee Draft for Vote (CDV), and Final Draft International Standard (FDIS). Since the IEC meetings are International and their meeting schedule is every 6-months, the balloting cycle can easily be 2-3 years in length. The IEC management group is working on ways to reduce this cycle time in 1/2 in order to meet the needs for the Semiconductor Market.

SWTPC6800 01:18, 13 April 2007 (UTC)

Proposed non- POV JEDEC units of semiconductor storage capacity

JEDEC Solid State Technology Association, the semiconductor engineering standardization body of the Electronic Industries Alliance (EIA) in Standard 100B.01[1] defines in the binary sense K, M and G as prefixes to units of semiconductor memory, noting that these definitions are “only included to reflect common usage” and that in accordance with IEEE/ASTM SI 10-1997, “This practice frequently leads to confusion and is deprecated.”

How about the above as a replacement?Tom94022 04:31, 13 April 2007 (UTC)
Much better, though I still don't particularly think the JEDEC needs its own section in this article. Still, your text is a marked improvement. -- mattb @ 2007-04-13T04:58Z
Good point, perhaps if Omegatron agrees we can put it into section 5.2, Computer memory. Alternatively, we could add a short IEEE section, more or less analogous to JEDEC. I'm open. Tom94022 05:20, 13 April 2007 (UTC)
If I agree? Wha? What do I have to do with this? — Omegatron 17:57, 13 April 2007 (UTC)
It appears that Omegatron added the JEDEC section on 7 April. On 12 April Sli flagged it as POV; to which I agreed. Sli has already given his permission to remove the POV. I'm not sure exactly how Wikipedia would want us to adjudicate this issue, but I think if Omegatron agrees to my proposed changes (or something similar) then we can move on.Tom94022 16:35, 14 April 2007 (UTC)

Yes, if the facts challenge the kibibyte belief system they must be removed. :) -- SWTPC6800 13:58, 13 April 2007 (UTC)

If you're going to make such an implication against others, why haven't you yet answered the charge that the JEDEC doesn't actually standardize unit convention? The JEDEC doesn't claim to standardize more than semiconductor memory, why do you persist in the argument that their usage of common terms is prescriptive? I've pointed this out many times to you and have yet to see you respond. This has been a weak argument since the moment you brought it up and most of your claims are based in total misunderstanding of what the JEDEC does and what standard documentation practices are. -- mattb @ 2007-04-13T14:14Z
The docuements produced from the JEDEC do say things like "JEDEC Stadards document", that is prescriptive. i.e. "Sanctioned or authorized by long-standing custom or usage.". Fnagaton 14:33, 13 April 2007 (UTC)
The standard document itself does not include unit prefixes, though. It merely defines them as they are used by the document as is standard technical writing practice. -- mattb @ 2007-04-13T15:03Z
The standards document defines them. It doesn't say "thou shalt only use this for verily it is thus" for each definition, but that is the danger when reading something out of context and in isolation of the whole document. By being defined for use in the standards document that appears to be good enough for all of the other documents that referene it use use those terms. Look at the language used in document JESD100B.01:

Page 1 : "JEDEC STANDARD" Page 2 : "No claims to be in conformance with this standard may be made unless all requirements stated in the standard are met." Page 9 : "TERMS, DEFINITIONS, AND LETTER SYMBOLS FOR MICROCOMPUTERS, MICROPROCESSORS, AND MEMORY INTEGRATED CIRCUITS" (The caps is how the text is displayed in the document.) That's actually prescriptive language, especially page 2. At least I've not seen any actual use of the binary prefix terms in any documents from the JEDEC despite extensive searches looking for those terms, meanwhile I have seen plenty of kilo-, mega-, KB, MB used. Fnagaton 15:30, 13 April 2007 (UTC)

Uh, that's like saying that a standard which defines the words should and must (a common practice) for its internal usage standardizes the usage of these words. That's just not how it works. --SLi 16:19, 13 April 2007 (UTC)
If that's what the argument hinges on, we can always email the JEDEC and ask whether a semiconductor device would be considered out of spec with their standards if they were marketed using IEC prefixes. -- mattb @ 2007-04-13T16:43Z
I've been told sending an email and then trying to use the reply would be WP:OR and is not allowed. For the result of that email to be accept as a reliable source it would need to be published by a few reliable entities with editorial control. Fnagaton 11:26, 14 April 2007 (UTC)
That's not a reason not to ask. The answer can be reported here and can inform our judgement even if it cannot go directly in the article. More information is always good. Better information is always good. And the inquiry might even spark publication; whoever answers might decide that it would be a good idea to clarify this in some press release or article or in a website FAQ or something like that. Dpbsmith (talk) 14:44, 14 April 2007 (UTC)
Also, the standard says in the footnote that they are deprecated (or at least quotes a major standard saying so, without any sign of disagreement with that). Don't you think this is of any importance at all? But I do understand you, it's certainly tempting to have selective vision when that's the only way you can find support for your POV. --SLi 16:51, 13 April 2007 (UTC)
Well it's not the JEDEC explicitly saying they are deprecated it's the quote in the footnote from the IEEE. So what it actually means is that the terms kilo, mega are defined in the standard and there is a footnote. The question would be to ask the JEDEC if it is standards correct to use the terms kilo,mega. I've highlighted the wider picture here. You are the one who is being selective by only fixating on tiny portions of standards documents or tiny fragments on of the MoS when you thnk it suports your point of view. Fnagaton 10:10, 14 April 2007 (UTC)
As I read the JEDEC document, the footnote pretty clearly states that the binary IEC definition is "an alternative definition" to the "common usage" definitions therein. Under the definition of mega it tabulates the alternative (IEC binary) definitions. So it seems like we don't have to ask JEDEC anything, Ki, Mi, etc are permitted under this specification, at least as I read it. Tom94022 16:35, 14 April 2007 (UTC)
As I read the standard using kilobyte and megabyte are acceptable terms to use in a document that wants to be notd as following the standard. Fnagaton 19:24, 15 April 2007 (UTC)

Federal Trade Commission recognizes JEDEC "In the Matter of Rambus, Inc." Docket No. 9302[11] is the largest standardization case the Federal Trade Commission has brought in a decade. The opinion cites JEDEC as being "an industrywide standard-setting organization" in the opening paragraph. The FTC has spent the last several years supporting JEDEC efforts to promulgate memory standards for RAM. RAM is that thinging consumers put in their computers and it is measured in binary units. The Binary Units that all producers of DRAM use is MB and GB (Mb and Gb for bits). They use the JEDEC standards not the esoteric IEC standards.

By that way, if you read all 120 pages of the FTC ruling they never mention that consumers are confused by memory size. -- SWTPC6800 17:30, 14 April 2007 (UTC)

You are a genius for digging up that reference. Is there some kind of "research barnstar" award I can give? ;) I'm glad to see that reference has also been included in this related talk Wikipedia_talk:Manual of Style (dates and numbers). Fnagaton 19:24, 15 April 2007 (UTC)

Windows XP Picture Example

Could you find a more complicated image to show the problem? Anyone who uses the Windows Disk Management tab should know what 1 MB of storage is. Grandmothers typically don't use more than 3 partitions. Windows XP Disk management for 160 GB disk SWTPC6800 14:09, 13 April 2007 (UTC)

Was your first question sarcastic? If not, I'm confused. If so, I said right in the image description that it's cluttered because of all the partitions and should be replaced by something else eventually. But I don't want to rearrange my backup drive just to take a screenshot. Maybe it could be photoshopped? — Omegatron 18:00, 13 April 2007 (UTC)

Images

  • We should have images from other OSes besides Linux and XP. Those are just the ones I have.
  • We shouldn't have too many images, so replacing the XP total drive size image with an equivalent from OS X is fine with me
  • I kind of wanted to show the same physical 160 GB drive in several different contexts, though, to illustrate how it differs, so I would prefer if the 160×109 examples aren't replaced.
    • Of course, if someone else wants to do the same thing I did, and show the outside of an external hard drive box, format it with two or three partitions, and then show how it appears in XP, Vista, Win98, Ubuntu, OS X, etc etc. that would be even better. — Omegatron 17:59, 14 April 2007 (UTC)


I agree that we shouldn't have too many images and I think we have already reached that point! I'm not sure we need any other OS's beyond those necessary to illustrate the various usages. That is, the relevant minimum set of images is using one drive and showing how it is displayed in "SI GB", "GiB" and "binary GB" by various OS's. Using one drive and showing the various ways it can be displayed really illustrates the source of consumer confusion. From an historical perspective, the early usage, i.e., currently 128K MAC System is also relevant. That makes 5 images:

any GB drive,
OS display of same drive in GB
OS display of same drive in GiB
OS display of same drive in binary GB
earliest example of volume OS with binary SI units

Omegatron has done a great job with his 160 GB drive in XP, Linux and Gnome (although arguably Gnome is not an OS). IMHO, all the rest are redundant with little informative value and should be dropped or at most placed in another article, e.g. Examples Of Binary Prefix Usage (including the 128K MAC if someone comes up with an earlier and more significant example).Tom94022 19:04, 14 April 2007 (UTC)

I thought it was important to show memory, hard drive, and data rate measurements, too, though, as examples of common usage. — Omegatron 21:50, 14 April 2007 (UTC)
From the big picture perspective, there appears to be two issues, 1) usage by computer memory vs. the Rest Of The World, and 2) confusion between disk drive specs & OS display. In order to keep the article to a reasonable length, I suggest we just highlight with the minimum number of illustrations to cover just these two issues. The 5 images above cover the disk issue, so adding one more image for memory is a good idea, but shouldn't we stop at that point - after all, just about every other unit system is SI (the last kilo I bought had 1,000 grams :-)? Tom94022 22:03, 14 April 2007 (UTC)
Fine with me. — Omegatron 22:34, 14 April 2007 (UTC)

Strange text

Hi,

About the text A programmer could easily mentally calculate that "8 kay times 4 kay is 32 meg" and get it exactly right, within this powers-of-two only context. This convenience is likely one source of originally adapting "Kilo" and "Mega" from SI as shorthand for 1024 and 1,048,576, as specialized jargon within a segment of the industry..

I'm not sure I understand this. What does it have to do with what terms are used? AFAICS it works just as well with kibi and mebi. --SLi 00:40, 15 April 2007 (UTC)

It does now, but kibi and mebi didn't exist then. The text is saying that calling 1024 a 'K' was convenient. Quirkie 00:09, 30 April 2007 (UTC)
I agree - It's a strange example that works both with decimal and binary units. It would have been more relevant for instance to state that a programmer easily can calculate that "128 buffers of 8 K each would take up exactly 1 meg". —The preceding unsigned comment was added by 85.112.166.110 (talk) 12:27, 14 May 2007 (UTC).

Inconsitent texts

The article Megabit states 'The megabit is a unit of information, abbreviated Mbit or sometimes Mb.' The template with the table of quantities of bits clearly shows that the abbreviation is Mb and not Mbit. Shouldn't the statement be reversed so it would say: 'The megabit is a unit of information, abbreviated Mb or sometimes Mbit.' ? —Preceding unsigned comment added by 145.7.182.13 (talk) 09:15, 29 October 2007 (UTC)

Fnagaton's edits

Fnagaton, you removed "such as computer memory sizes", and added "rather than using the system described below", claiming that this makes the intro more neutral.

Computer memory is the most obvious example of a use of binary prefixes, and I can't fathom how mentioning this would be considered biased. What does this bias the intro in favor of? And there are more than one "systems described below". Which one should the reader assume you are talking about? How is removal of this phrase biased?

Can you explain why you made this edit, besides a continuing pattern of antagonism? — Omegatron 17:00, 1 June 2007 (UTC)

Section 'consumer confusion'

it says "Apparently, some computer programmers were unaware that disk drive manufacturers used the SI notation when specifying and/or advertising capacity of their hard disk drives" This is competly bogus. There is no chance that any developers, then or now, was not aware of the practice of the drive manufacturere. by 1984 they had all switch to decimal prefixed Bytes... Practically, we (then and now) represent memory size and file size with memory unit, and calculating 'decimal' unit implied a binary to decimal conversion that is not trivial (converting a number of byte in KB is a matter of shifting right by 10 bits, which is a single and fast assembler instruction. dividing by a 1000 was a much more expensive proposition CPU wise ( [http://home.comcast.net/~fbui/intel/d.html#div 13 times more cycles for a 386). Developer back then knew just as much as hard drive manufacturer knew that KB/MB... was used by said developers and memory manufacturers to mean 1024... Furthermore, these were not ans still are not 'SI' notation. MB and GB (but not KB) can possibly be described as using a 'SI prefix', but that is not a 'SI notation' by any stretch of the imagination. -- Shmget 07:57, 8 June 2007 (UTC)

then it says "This mixed presentation..." What mixed presentation ? The presentation is not mixed, as the text explain later "would report the space available on a 41,959,424 byte hard disk drive as 40 MB". that is not 'mixed representation', that is a consistent presentation with all the other size in the OS. a 'mixed' representation would be to say that a a disk as 40MB of space of 40,000,000 and then send a message saying that there is 150 KB left after copying 38 1MB-file on it.... or to say that the hibernate file to swap 256MB of RAM to disk is... 268.43 MB !!!! -- Shmget 07:57, 8 June 2007 (UTC)

If you contend the programmers willfully ignored the HDD manufacturers specifications and advertisements, would it be less POV to say, "For unknown reasons, some computer programmers chose to ignore the disk drive manufacturers practice of using SI prefixes when specifying and/or advertising capacity of their hard disk drives"
I content that, for the purpose this enciclopedic article, it is irrelevant who ignored what and why. One could just as well turn the statement around and say that hard drive manufacturer chose to ignore the usage of computer memory, file size and floppy disk size... But either way you turn that phrase would be POV. I content that that phrase - or the same phrase turned around, is useless and bring nothing to the article but some POV... -- Shmget 06:35, 9 June 2007 (UTC)
I hope you are not seriously contending that it was to save a few machine cycles.
I am dead serious - Shmget 06:35, 9 June 2007 (UTC)
First of all, in the relevant time period, at the machine level, capacity was available as a string of Hex characters.
No kidding, you mean that at a 'relevant period', had disk didn't have Head, Cylinder and Sector but just a 'capacity as a string of Hex characters' ?
No programmer in his right mind would internally represent any measure in a mixed notation of decimal characters with binary prefixes.
You are right, and they don't, there is no mixed representation, internally there are Bytes, or power of two of bytes... taht's it.
I don't know of an assembler or compiler that supports prefixes of any sort.
In so called 'binary unt' that is call 'shifting'. For instance an inode contain a block-size and a number of block. the block-size is typically 5112 bytes. So when you want the size in KB you just 'shift right' the number of block by 1 bit. (See Posix description of '-k' option of du or ls for some discussions on that )
 When I use them in my programs I define them as constants and use them consistently, i.e. using pseudo-code
decimal decMyNumber = 16777216m;
deximal decMega = 1000000m;
hex hexMyNumber = F00000x;
hex hexMega = 100000x;
decMyMegaNumber = decMyNumber / decMega;
hexMyMegaNumber = hexMyNumber / hexMega;
you would never see a
whatTheHeckIs = decMyNumber / hexMega;
It would be all Hex internally!
You confuse the bottle and the wine. you are confusing a string that represent a number and the binary representation of that number. that being said the example above use 'division', but really hexMyMegaNumber = hexMyNumber / hexMega; is number_of_megabyte = number_of_byte >> 20; (BTW, what programming language use suffix for numeric constant ??? I mean I have seen some variety with the prefix used to represent hexa constant, but I don't recall ever seen a programming language that used a suffix for that - and for good reason, that would make parsing unnecessarily painful) -- Shmget 06:35, 9 June 2007 (UTC)
The conversion would be for display purposes and for that purpose the difference between a binary shift followed by a hex to decimal conversion is not much more than a hex to decimal conversion followed by a decimal shift. After conversion with data in a BCD format the shift is a byte shift (4 bit) shift. There is no difference in time to speak of (3 vs 4 ?) and if u just move a pointer, no time at all.
"if u just move a pointer, no time at all." ??? euh ? how exactly to you 'move a pointer' to get rid of least significant digits ? BTW shifting 'before' converting to ascii reduce the number of digit to convert. so there IS a difference. You alos miss the point that these size are not just there to be display, that they are often already expressed in power of tow, like the number of 512 bytes blocks... for that specific example: If I have the number of block of a file, all I have to do is: nb_of_KB = (nb_blocks >> 1) and then bin_to_ascii(nb_of_block). To display that using 'decimal' unit, I would need to do nb_byte = nb_block << 9 (which mean that I need a 64 bits register or I am limited to (2^32 - 1) for my file size, then calculate nb_kB = nb_bytes / 1000, and finally bin_to_ascii(nb_kB). Now imagine that you actually have a 8086 a 6502 or a 8086 processor and picture doing the later with these 8/16 bits register processor.... (the former is quite easy). -- Shmget 06:35, 9 June 2007 (UTC)
The mixed presentation is the mixture of a decimal digits with a Binary Prefix Abbreviation!
What are you talking about ? I truly don;t undertand what you pint is. For one thing I never use 'binary prefix' nor does any of the operating system mentionned here. they used KB,MB,GB, these are Units, that have a value of respectvely 2^10, 2^20 and 2^30 bytes. For example If I say I am 80 years old or I am 960 months old, I am using 'decimal' number (as in 80 and 960) with some unit... how is that a mixture of anything ? is it because some else may chose to say that I am 2.524 Gs old ? -- Shmget 06:35, 9 June 2007 (UTC)
The fact that the unit abbreviation B is not SI recognized is irrelevant.
When one side insist that they have the 'true'(tm) meaning on the ground of the holy BIPM, then one cannot just turn around and in the very same unit commit BIPM blasphemy. So yes it is very relevant to this whole 'fuss' by the hard drive industry.
And your point about k v K is really meaningless with regards to HDDs, I don't think there has ever been an HDD measured in k or K by the manufacturers. The first one was 5 million characters.Tom94022 21:25, 8 June 2007 (UTC)
I'm sure you noticed it wasn't 5 MB. But the point is that contrary to what the current formulation says, system programmer where in no what obligated nor even remotely expected to adopt the hard-drive manufacturer definitions, quite the contrary, having done so would have created internal inconsistency. you would have had displayed a 300 'MB' (hard-drive manufacturer definition) drive containing a 256MB file size, containing the dump of you 256MB memory, but still you would have only 30 MB left of that drive... Now THAT would have been mixed and confusing. -- Shmget 06:35, 9 June 2007 (UTC)

RAM storage capacity is a result of the number of address lines. Disk storage capacity is a result of tracks, sectors, and platters. Most Flash memory emulates disk storage. [12] (There are exceptions but this covers most of the cases.)

It is a very rare solid state memory whose size is not a binary multiple. Intel's first dynamic RAM had 1024 bits. Formatted disk storage has always had capacities unrelated to binary multiples. It would be a dim witted computer programmer that did not know the difference.

There would be a case for consumer confusion among English majors, but not for computer programmers. The current statement in the Binary Prefix article is absurd. -- SWTPC6800 05:08, 12 June 2007 (UTC)

  1. ^ JEDEC Solid State Technology Association (December 2002), "Terms, Definitions, and Letter Symbols for Microcomputers, Microprocessors, and Memory Integrated Circuits", JESD 100B.01