Talk:Binary prefix/Archive 11

Archive 5

←

Archive 9

Archive 10

Archive 11

Archive 12

Archive 13

→

Archive 15

"Commonly, historically, and correctly used"

Some here are saying that Adobe, AMD, Apple, Computerworld, Dell, EE Times, HP, IEEE Computer Magazine, Intel, Kingston Technology, Microsoft, Oracle, Samsung, Symantec, Toshiba, and a host of other publications, computer manufacturers and software companies have got it wrong and a bunch of esoteric standards nerds have found the one true method of measuring computer storage capacity.

There is no legal obligation to use this IEC and ANSI/IEEE standard. Every IEEE standard has these disclaimers. "Use of an IEEE Standard is wholly voluntary." "The existence of an IEEE Standard does not imply that there are no other ways to produce, test, measure, purchase, market, or provide other goods and services related to the scope of the IEEE Standard." After the ASME vs Hydrolevel antitrust case [1] was upheld by the Supreme Court[2] standards organizations are abundantly cautious about pushing their standards.

The use of MB and megabyte for binary values is more than common usage, there is 50 years industry practice codified in ANSI/IEEE and other standards. The previous standards formally defined what the industry was already using. Coining new terms like mebibyte is an attempt to change industry practice. In 1984 the ANSI/IEEE Std 91-1984 and IEC 60617-12 standards recommended that everyone start drawing schematic symbols of AND gates as a square box with an ampersand in it. Changing something that the industry thinks is working is very difficult.

The only significant usage is in elite standards groups. I would have the say the adoption of the IEC binary prefixes is minuscule and static. One of the major points on the consumer confusion argument was the difference between RAM, floppy disk and hard disk measurements. Floppy disks are gone, and all hard disk now come with a disclaimer stating that a GB is a billion bytes. The rest of the computer industry is staying with previous ANSI/IEEE standards that define KB, MB and GB as binary units. -- SWTPC6800 (talk) 23:50, 18 March 2008 (UTC)

There is a much longer and deeper history of usage of kB (KB) and MB to mean 1,000 bytes and 1,000,000 bytes respectively. It was the failure of Apple and Microsoft beginning 1984, to explain their then unusual usages that created the ambiguity. Given the binary usage came second it seems unreasonable to now require the rest of us to conform to the rather narrow usage solely with regard to primary memory. IEC and others recognizing the problem came out with an unambiguous prefix system, Ki, Mi, etc. which is slowly being adopted. I seem to recall the same situation when cps was replaced with Hertz; people always resist change. Tom94022 (talk) 00:41, 19 March 2008 (UTC)

If you want to talk about which interpretation is legal, consider that the government hardly ever enforces the use of units of measure except when used to measure something in sold in commerce. The binary prefixes are bigger than their decimal counterparts, so someone who sells something that is described with the binary prefix is not shortchanging the customer, and so is not likely to be prosecuted. --Gerry Ashton (talk) 04:43, 22 March 2008 (UTC)

Those are the only examples you can come up with for the binary usage? I can come up with a lot more that use the prefixes according to the standards. The overwhelmingly vast majority of instances of k- or M- mean 1,000 or 1,000,000, and the overwhelmingly vast majority of people associate k- and M- with 1,000 and 1,000,000, especially in countries that are not the US.

As for the standard binary prefixes, I see several instances of software being updated to use units correctly, and no instances of software being changed back to the ambiguous definition. Maybe we'll see this abandoned in the future, but adoption is clearly increasing at present. — Omegatron 03:58, 26 March 2008 (UTC)

Binary Prefix Confusion from 1968

When Bill Gates and Steve Jobs were 13 years old this letter appeared in Communications of the ACM

Morrison, Donald (March 1968). "Letters to the editor: Abbreviations for computer and memory sizes". Communications of the ACM. 11 (3). ACM Press: pg 150. {{cite journal}}: |pages= has extra text (help)

Editor:

The fact that 2¹⁰ and 10³ are almost but not quite equal creates a lot of trivial confusion in the computing world and around its periphery. One hears, for example, of doubling the size of a 32K memory and getting 65K (not 64K) memories. Doubling again yields a 131K (not 130K) memory. People who use powers of two all the time know that these are approximations to a number they could compute exactly if they wanted to, but they seldom take the trouble. In conversions with outsiders, much time is wasted explaining that we really can do simple arithmetic and we didn't mean exactly what we said.

The confusion arises because we use K, which traditionally means 1000, as an approximation for 1024. If we had a handy name for 1024, we wouldn’t have to approximate. I suggest that κ (kappa) be used for this purpose. Thus a 32κ memory means one of exactly 32,768 words. Doubling it produces a 64κ memory which is exactly 65,536 words. As memories get larger and go into the millions of words, one can speak of a 32κ²(33,554,432-word) memory and doubling it will yield a 64κ² (67,108,864-word) memory. Users of the language will need to have at there fingertips only the first nine powers of 2 and will not need to explain the discrepancies between what they said and what they meant.

Donald R. Morrison

Computer Science, Division 5256

Sandia Corporation, Sandia Base

Albuquerque, N, Mex.

-- SWTPC6800 (talk) 01:42, 19 March 2008 (UTC)

I agree with Morrison, this started out using K as an approximation of 1024, not K=1024! I contend that it was the ubiquitous use of K and M by Apple and Windows to characterize drive size that caused the consumer confusion. We, the practitioners, well understand the distinction but the lay person doesn't. Isn't it a pity that Jobs et al hadn't read Morrison's letter by 1984 when Mac began using K to characterize Floppy Disk formatted capacity. Same thing for Gates with Windows. If, as Morrison suggests, they had taken the trouble to explain what they meant or adopted another symbol, Apple and Microsoft would not have created this mess. Tom94022 (talk) 16:48, 19 March 2008 (UTC)

Exactly. Using "1k" to mean "1024" is perfectly fine. Using "64k" to mean "65536" is what caused the problem. — Omegatron 03:07, 22 March 2008 (UTC)

Not SI Prefixes Prefices

Hmmm... some people keep calling kilo... mega... giga... &c. SI prefixes when used to specify quantities of binary storage units.

When used in this manner, these are not SI prefixes, therefore it is extremely confusing to refer to them as such in the section heading, and I would suggest that to refer to them as such is to imply that the SI use of these prefixes is the only correct use of them ever and that any other use of them is implicitly wrong and bad so don't do it right? Sounds like a non-neutral point of view to me...

84.9.125.170 (talk) 21:07, 25 March 2008 (UTC)

Hello, I think there is a (somewhat philosophical) confusion here involving the use–mention distinction. The right name for the prefixes mega-, giga-, etc is "SI prefixes", and their adoption for the binary sense clearly happened long after they were introduced as part of the SI. Now the "philosophical" question: "Is an SI prefix still an SI prefix if it is not used in the SI sense?" I would say Yes. "SI prefix" refers to the words (the "mention"), not the meaning ("the use"). I agree with you that it needs disambiguation specifying the context in which it is used (indeed, that is what I think is the whole problem :P), but the name for these words remains "SI prefixes", and to make things clear, "SI prefixes used in the binary sense", is the best I could come up with. Making up new names like "SI-like prefixes" or "pseudo-SI prefixes" will only add to the confusion, because the prefixes are *identical* (as you yourself wrote) to the SI prefixes. They are not different prefixes deserving a different name, just the same prefixes used differently. --shreevatsa (talk) 21:29, 25 March 2008 (UTC)

No. There's no confusion and no ambiguity and no philosophical problem. The Système Internationale is what it is, and it defines the prefixes clearly, and it does not define or endorse any "binary sense."

If you buy an "Troy ounce" of gold from me, I can't deliver an avoirdupois ounce and say casually, "Oh, I thought you meant a troy ounce in the avoirdupois sense." When you said "troy ounce," you said exactly what meaning of the word ounce you were referring to, and no misunderstanding is possible.

If you want to use kilo- to mean 1024, fine, when you do so you are not using the SI prefix kilo-, you're using some other kind of prefix. Dpbsmith (talk) 22:49, 25 March 2008 (UTC)

I agree with you, and with 84.9.125.170, that using "kilo-" to mean 1024 is not SI usage, and hence it seems wrong to call it a SI prefix in this context. I felt that the only appropriate name for the {"kilo-", "mega-", "giga-", etc.} set of prefixes is "SI prefixes" (note: name for the prefixes themselves, not usage), so I had called the section "SI prefixes used in the binary sense" to make it both clear what the prefixes were, and that the usage was not SI. You have decided that the appropriate name is "SI prefix names". I think this is a bit cumbersome, but it's okay. For the record, I think "Traditional binary prefixes" is an acceptable name for that section too, and looks better than "SI prefix names, when used in their non-SI binary sense." (The section mentions that these prefixes are identical to the SI prefixes anyway.) --shreevatsa (talk) 23:06, 25 March 2008 (UTC)

Hi

(formatting messed up a bit when I used the :: but this is in reply to the end of the thread above)

If your point is that it may be more appropriate to say 'SI' prefices to highlight the fact that the prefices are derived from SI, then I understand this, although disagree.

When you say that the right name or description for the mega, giga prefices is or are SI prefices I would also disagree. They are merely words or parts of words and their meanings or descriptions derive from their use and application, not their original definition.

If they are re-used as convenient prefices for a similar system of multiples, then this new use or description is not automatically wrong, it then as now has its own separate meaning and existence, and it is completely irrelevant to even mention SI except perhaps by way of explaining the source and reasons behind the adoption.

The meaning and application of words change over time, the application of these prefices has expanded to include their use as binary prefices, and when they are used as such, this use does not match the definition of SI prefices and therefore they cannot be described as such.

I for one see absolutely no ambiguity in the dual use, with the result that if you are talking about

   quantities of bytes / words in any application
   bits on memory chips

then mega means 2^20 and is not an SI prefix, and if you are talking about

   bits on a wire

then mega means 10^6 and is an SI prefix

the main test would be that if you are counting storage units defined in terms of powers of 2 you use prefices that are also defined in terms of powers of 2.

84.9.125.170 (talk) 02:03, 26 March 2008 (UTC)

Prefixes vs prefices

Could we please stop changing prefixes to prefices? First, both are correct and generally Wikipedia policy is to leave things as is in this case. Secondly, most computer science literature uses prefixes rather than prefices. Andareed (talk) 00:04, 4 April 2008 (UTC)

Greetings fellow editors. As this article is about prefices and not a single prefix, I think we should rename it to Binary prefices, and link Binary prefix to it. I also think the answer to the question above is quite obviously no. Oh yeah and I corrected the spelling in the heading. 84.9.125.170 (talk) 05:19, 4 April 2008 (UTC)

To the anonymous editor. I did try to look up prefices without much luck. I did find prefixes more commonly used. Wikipedia's spell checker doesn't recognize prefices but does recognize prefixes. Since prefixes was first, and there is no justification for prefices you really shouldn't be making such global changes. You are also likely in violation of one of the several Wiki policies (watch out for WP:3RR). So please stop it. Tom94022 (talk) 06:34, 4 April 2008 (UTC)

My dialect is British English, with some experience in New Zealand English (I'm a Kiwi) and Singapore English. I suspect "prefices" *may* be a valid variant on prefixes, but I can't say for sure. Dictionary.com returns *nothing* for "prefices" and this for "prefixes". Given that the article original has "prefixes", and that "prefixes" is perfectly valid, WP:ENGVAR ("Retaining the existing variety") suggests that "prefixes" should be retained.  This flag once was red  06:44, 4 April 2008 (UTC)

I'm getting just over 4,000 Google hits for prefices (compared to 3 million prefixes) and I'm inclined to think it's just a hypercorrection error. The -ices ending is the plural of the Latin -ex/-ix ending, but of course prefix is not pref-ix but pre-fix, fix being an English word (albeit with Latin roots) with no reasons whatsoever for Latin inflection. -- Jao (talk) 07:40, 4 April 2008 (UTC)

In reply to the anonymous Tom, Flag, and Jao, and not forgetting the anonymous Andareed, I think prefices both looks and sounds better than prefixes. 84.9.125.170 (talk) 08:23, 4 April 2008 (UTC)

Noted. But what's your point? An editor on Wikipedia thinks A looks and sounds better than B, so we should all ignore past practice, common usage and etymology?  This flag once was red  08:31, 4 April 2008 (UTC)

Isn't that the entire point of the article this talk page is for? Arguing about "looking and sounding better" as opposed to "past practice and common usage"? A war-within-a-war! (I go with "prefixes". I also go with "Mega/Giga" not "Mebi/Gibi", simply because 99% of the time, I'm talking to lay-persons, who think I have a speech impediment when I use "Mebi/Gibi". My job is to fix computers, not spend 10 minutes "educating the masses.") 71.193.198.73 (talk) 19:02, 11 June 2008 (UTC)

If that editor is me then of course yes. 84.9.125.170 (talk) 09:04, 4 April 2008 (UTC)

The editor above was obviously being bold. 0x54097DAA (talk) 18:28, 9 April 2008 (UTC)

On the first occasion, sure. I think we all assumed good faith on the first occasion.  This flag once was red  19:20, 9 April 2008 (UTC)

We need a name for the SI prefixes

This problem with the use-mention distinction crops up again. There is a need for a name for the set of words {"kilo-", "mega-", "giga-" etc.} in the article. Previously, the phrase used was "SI prefixes", because it was felt appropriate at that time. Now there has been some disagreement over this (see above), with the contention that if they are not used in the SI sense, they cannot be called SI prefixes. So if this is valid, then we need an alternative, unambiguous name for the prefixes. Parts of the article use "SI prefixes" for them: "The new standard [..] the SI prefixes will henceforth only have their base-10 meaning [..]" and at least one part uses "SI designation", as in the phrase 'the phrase "decimal unit" will be used to denote "SI designation understood in its standard, decimal, power-of-1000 sense" and "binary unit" will mean "SI designation understood in its binary, power-of-1024 sense."' Both these sentences are talking about the words themselves, but if you take the viewpoint that SI prefixes unambiguously mean "SI prefixes used in the SI sense", then they seem redundant (or wrong). So we need a new name for the words which can be used. I propose "metric prefixes", since these prefixes came into common usage with the metric system in the 18th century, and the phrase does not seem to have the connotations of dictating usage of these prefixes. An alternative name would be fine too. We just need to decide on something, or else this will continue to remain confusing. shreevatsa (talk) 13:57, 4 April 2008 (UTC)

The SI prefix article states "also known as metric prefix" so that could still lead to confusion. I think SI-style or SI-type would work, or use quote marks as in "SI" prefix.

Sorry can't resist saying that the prefix usage in the 18th century would have been mostly if not completely restricted to the French, and therefore not common. 84.9.125.170 (talk) 14:29, 4 April 2008 (UTC)

Yeah, I agree. The problem with "SI-style" or "SI-type" prefixes is that they are misleading -- these prefixes are not like the SI prefixes, they are identical to the SI prefixes. How about "Greek prefixes" or "kilo-style prefixes"?

Or maybe we could put a 'In this article, "metric prefix" will be used to mean the prefixes like kilo-, mega-, giga- etc., independent of the sense in which they are used.' (Or the same for "Greek prefixes". Come to think of it, we could do the same for "SI prefixes" too ;-)) shreevatsa (talk) 04:38, 6 April 2008 (UTC)

I think that the term "metric prefix" still conveys the impression that they are used in the SI sense, and so might confuse the reader. My suggestion is "KMG prefix" as in 'In this article, "KMG prefix" will be used to mean the prefixes like kilo-, mega-, giga- etc., independent of the sense in which they are used.'. Thunderbird2 (talk) 07:56, 6 April 2008 (UTC)

"KMG prefixes" is inventing terminology that some readers may mistake as being "official terminology" so I wouldn't agree with that choice. Just say "traditional prefixes" or "common prefixes" or even "prefixes". Fnagaton 09:59, 6 April 2008 (UTC)

To call them SI-style etc. could be considered appropriate because the most common use of them at the moment is in relation to SI, and more people will recognise them in this context than in any other.

However, these prefixes at least up to mega have been used in science even before SI existed, and the computing industry has used at least kilo to signify binary multiples before SI existed.

If not SI or KMG (which I quite like) then why not MKS-style (or -type) as they were known before the advent of SI. This will clearly indicate their derivation whilst ensuring that they are not incorrectly named. 84.9.125.170 (talk) 11:35, 9 April 2008 (UTC)

So how about "traditional prefixes"? Webster's 1828 dictionary has kilogram as 1,000 grams, they were applied to non-metric explosive power of bombs circa 1945 and, as late as 1968, my Webster's only has decimal meanings for kilo and mega. Tom94022 (talk) 23:56, 22 April 2008 (UTC)

Windows PowerShell

The section about command-line interpreters is really blown out of proportion. I know dozens of shells and command-line interpreters but I don't know any which support binary prefixes. The support in the PowerShell is also pretty obscure "1 kB" becomes 1024? So what happened to the "B", the "bytes"? That's not how you calculate with units. And "mb" is anything but "Megabyte". It might millibit. It's beyond my comprehension what the developers where thinking when they implemented this but I doubt we need a hole section on it. --217.87.114.55 (talk) 16:21, 22 April 2008 (UTC)

Hi, thanks for your feedback. I added this section because I strongly feel that it is impotant to mention the fact that the usage problem described in this article is not only limited to floppy disks, HDDs and optical discs, etc., but also exists in current shell scripting languages that were created even after the introduction of the symbols Ki, Mi, Gi, etc. by the IEC in 1999. The way it has been implemented by MS is as confusing as the floppy disk and CD-ROM/DVD usage of the prefixes.

As stated above the given example, in PS all prefixes are case-insensitive, which is quite distinct from other areas were these symbols are used. Therefore "mb", "MB", "mB" etc. are all the same in this programming language (see example). The reason why the "B" for "bytes" is missing is because this is used for calculations that don't actually require units, it just returns the number of bytes as an integer (the ls command also returns the number of bytes without any prefix). Also there are no millibits etc. only bytes are used in this implementation. Ghettoblaster (talk) 18:26, 22 April 2008 (UTC)

What I mean is that it generalizes from a single example. Is there any other "command-line interpreter" (that is noteworthy) with similar support? Also if we keep listing software that uses prefixes in a way that doesn't conform to IEC 60027-2, even if it's just software since 1999, there would be no end. Microsoft and their Windows are already covered. Even if you want to show how absurd this "feature" is, I think the examples take to much space and are hard to understand if you're not familiar with programming and shells. I believe "PS C:\>" is just the prompt, I find it hard to digest and I don't see how this piece of software deserve a whole section. Let me also add that programmers who are one of the few who might have to work with powers of 1024, virtually never used any prefixes at all in their code. If you want "1000", you have to spell it out. If you want 1024, you have to spell it out. I don't know of any programming language letting you write "1k" instead. That's what makes this whole prefix discussion so absurd. The people who established this convention, aren't using it themselves professionally but use it only colloquially in comments or spoken language. The people who normally wouldn't need to know this at all have only disadvantages because of this convention. This PowerShell example is the first I've ever seen where these prefixes are embedded into a language. It's another example how this corporation solves problems by making them worse. A better place for this might be broken by design. --217.87.102.163 (talk) 20:15, 23 April 2008 (UTC)

Introductory section

Hello, I have reorganised the introductory section mostly back to how it was earlier, because I think it makes more sense. The point is not to group similar usages together etc., but to give a clear idea to a new reader what the issue is. The mention of the JEDEC standard only makes sense after the reader knows that there are two systems of prefixes, and the fact that certain areas of computing always use the Greek prefixes (kilo/mega/giga) in the SI (decimal) sense was put there only to underscore that the issue is ambiguous. Roughly, here is the information that the introduction ought to convey:

• There is a need for binary multipliers
• It was traditional in some areas to use the Greek prefixes for this
• This was inconsistent with SI, with other areas, and even with itself (e.g. "1.44 MB"),
• so it was ambiguous what "megabyte" etc. meant.
• The IEC introduced new prefixes to solve this
• The new prefixes were taken up by many standards organisations, and is now recommended by them.
• The old system is (understandably) still somewhat common, and things like the JEDEC standard reflect this fact.

The article as a whole still needs a LOT of improvement before it is intelligible to a new reader (the point of the article is to be helpful, after all :P), e.g., I think the history section should come after the discussion of the two systems... someone needs to spend time cleaning it up so that it is actually useful and holds a reader's interest (is skimmable, etc.) For now, I have attempted to make the introduction clear. shreevatsa (talk) 20:42, 23 April 2008 (UTC)

IEC

Okay. Exactly WHO gave IEC the authority to tell us what to use? 85.225.114.237 (talk) 12:50, 10 May 2008 (UTC)

No one. Use on WP is determined by consensus. Thunderbird2 (talk) 14:15, 10 May 2008 (UTC)

And the consensus is to use what is commonly used and that is not IEC.DavidPaulHamilton (talk) 22:38, 10 May 2008 (UTC)

And the WP consensus is that there is no consensus. (Although people on both sides claim that THEY have consensus.) 71.193.198.73 (talk) 19:04, 11 June 2008 (UTC)

There is no consensus to use ambiguous units like megabyte, unless accompanied by an explicit disambiguation. The consensus is that use of IEC units is an acceptable way of doing this. For example 128 MB (128 MiB). Thunderbird2 (talk) 22:52, 10 May 2008 (UTC)

MOSNUM mentions IEC is unfamiliar and to use familiar methods.DavidPaulHamilton (talk) 02:08, 11 May 2008 (UTC)

There is no consensus for that statement. That's why it carries a 'disputed' banner. Thunderbird2 (talk) 08:26, 11 May 2008 (UTC)

There is consensus for that statement.DavidPaulHamilton (talk) 04:26, 12 May 2008 (UTC)

What's your evidence that there is consensus, because I sure don't see it. Dpbsmith (talk) 23:37, 28 May 2008 (UTC)

There is no indication the original comment was asking about use on Wikipedia :) shreevatsa (talk) 04:30, 29 May 2008 (UTC)

Its true, but um, then the answer is still consensus. I can go out and say that my method is the way of the future, but if no one listens, then the old method stays. consensus isnt always a majority of ALL people :D 10max01 (talk) 22:38, 3 June 2008 (UTC)

IEC 80000-13

IEC 80000-13 has now replaced subclauses 3.8 and 3.9 of IEC 60027-2:2005, which covers the binary prefixes. Caerwine Caer’s whines 20:50, 26 May 2008 (UTC)

Are there any changes of substance? Thunderbird2 (talk) 21:56, 27 May 2008 (UTC)

These bloody iso/iec-papers are realy hard to come by, if you aren't willing to pay a small fortune. I'll try some other libraries later. In the mean time, this may help: "...This standard cancels and replaces subclauses 3.8 and 3.9 of IEC 60027-2:2005. The only significant change is the addition of explicit definitions for some quantities..." [3]. --213.183.10.41 (talk) 13:10, 1 July 2008 (UTC)

RFC discussion of User:Greg L

A request for comments has been filed concerning the conduct of Greg L (talk · contribs). You are invited to comment on the discussion at Wikipedia:Requests for comment/Greg L. -- — Omegatron (talk) 00:28, 6 June 2008 (UTC) — Omegatron (talk) 00:28, 6 June 2008 (UTC)

Manual of Style

Can we please change the manual of style now? (Wikipedia:MOSNUM#Quantities_of_bytes_and_bits ) 1 KB = 1000 bytes and 1 KiB = 1024 bytes. Enforcing this would clear up a lot of articles like the size limits in the File_allocation_table which are currently wrong. —Preceding unsigned comment added by 98.213.141.241 (talk) 07:06, 3 July 2008 (UTC)

The part of MOSNUM to which you refer was introduced on 7 June 2008, despite a clear consensus against such deprecation of IEC units. Thunderbird2 (talk) 15:03, 5 July 2008 (UTC)

Not true. The text you, TB2, refer to was put there with consensus, the link you provided claiming "despite a clear consensus" misrepresents the truth of where the consensus is. This is because firstly votes (the link you provided is nothing but a vote) don't make consensus, good arguments make consensus. Secondly the real consensus is here. At the time you did not present substantive arguments and you have still have not provided substantive arguments. The consensus is actually for the text in MOSNUM which includes the deprecation of IEC prefixes for the many good reasons given in the link I have provided. Fnagaton 16:34, 5 July 2008 (UTC)

Truth is in the eye of the advocate. I've been watching and participating in this discussion for a number of months; somehow I missed the discussion referred to above and therefore was unable to express my opinion - I would have objected to the rewrite. I notice a number of folks opposed to the deprecation of IEC are also missing from this so-called consensus - a coincidence or manipulation? Looks like the latter to me :-( Tom94022 (talk) 18:30, 5 July 2008 (UTC)

A discussion has been started at WP:MOSNUM concerning the continued deprecation of IEC prefixes. Please comment at the MOSNUM talk page. Thunderbird2 (talk) 18:54, 5 July 2008 (UTC)

32K store

The article included the claim that 32K is used in the binary sense in the following quote:

• "The 8K core stores were getting fairly common in this country in 1954. The 32K store started mass production in 1956; it is the standard now for large machines and at least 200 machines of the size (or its equivalent in the character addressable machines) are in existence today (and at least 100 were in existence in mid-1959)."

There is nothing there that confirms either binary or decimal use, so I weakened the sentence. It would be better to remove it altogether. Thunderbird2 (talk) 19:15, 10 August 2008 (UTC)

WP:OR "Articles may not contain any new analysis or synthesis of published material that serves to advance a position not clearly advanced by the sources." Fnagaton 21:59, 10 August 2008 (UTC)

Exactly. So if the source makes no mention of a binary meaning, the article should not either. It seems to me the claim should go. What do others think? Thunderbird2 (talk) 22:19, 10 August 2008 (UTC)

Cut and paste to this talk page exactly the parts of the source you read that you think are "dubious" and those parts of the source you read which lead you to that conclusion. Fnagaton 22:26, 10 August 2008 (UTC)

(outdent) In 1960 there were two main types of computer architecture, the binary addressed long word machines and the decimal addressed character machines. The IBM 701 could address 4096 words 36 bits wide. (It was updated with the IBM 704 in 1954.) The IBM 1401 was a character addressed machine that used decimal addressing. It could have up to 16,000 8-bit characters. I have seen a 1401 in action and took the photo of the control panel that is in the Wikipedia article.

The letter writer is criticizing a speech by Andrew Booth that was reprinted in ACM. Here is the entire paragraph that has the quote.

Next, size of high speed storage. Dr. Booth regards 4096 words of 40 bits as large (i.e. 163,840 bits): "… only in the last couple of years have machines been produced with storage organs whose total capacity is of this order of magnitude and many machines which are currently manufactured have high-speed storage for only one hundredth of this number of data." The 8K core stores were getting fairly common in this country in 1954. The 32K store started mass production in 1956; it is the standard now for large machines and at least 200 machines of the size (or its equivalent in the character addressable machines) are in existence today (and at least 100 were in existence in mid-1959). It is odd, also, that although Dr. Booth bases many of his points on the state of the art in England, he refers constantly to American developments and milestones.

The letter writer was referring to binary address computers as opposed to "character addressable machines" like the IBM 1401 or the IBM 702.

You should read up on early computers. The early memory devices were columns of mercury. They acted as a delay line. A stream of audio pulses were injected into one end of the tube and traveled to a microphone on the other end. It was then looped back, acting as a shift register. This would hold 1000 digits. This is from a book I have titled "High-Speed Computing Devices" printed in 1950, it has a table of Large-Scale Digital Computing Machines in the United States. This listed all of the computers in the US, all twenty of them. Only ten of them were operational, the other ten were under construction. The computer did not begin with the Apple Mac. -- SWTPC6800 (talk) 03:07, 11 August 2008 (UTC)

None of this proves that the K was intended to take any meaning other than its time honoured one of 1000, only in this case it meant "approximately 1000". To assume otherwise is OR. Thunderbird2 (talk) 07:41, 11 August 2008 (UTC)

Thunderbird2, I note that you have not answered the direct challenge put to you above (22:26, 10 August 2008) and as such you have failed to support your position. Reading the source it is obvious and unambiguous that the 8K and 32K in the cited text refers to binary powers of two quantities. Thunderbird2, it is incorrect to add the "dubious" tag because the source itself (and the evidence from the source posted by SWTPC6800) proves you are wrong. Fnagaton 08:42, 11 August 2008 (UTC)

I think it's somewhat clear that the "32K" there refers to something that is actually 2¹⁵=32768, but the problem is that it's not clear whether the usage was the result of a definition of 'K' as 1024, or the result of approximating '32.768 K' to '32 K' through truncation. For powers of 2 less than 2¹⁶=65536, both truncation and the binary convention give the same result, so it's hard to be sure. (Note that for '32.768', any method of approximation other than rounding down or truncation will actually yield '33', so there is partial support for the K=1024 theory, but this in itself is not conclusive because truncation actually seems a common method: 65536 is closer to 66×1000 but it is often written "65K" as in 65K colours.) shreevatsa (talk) 13:00, 11 August 2008 (UTC)

Precisely. See also IBM’s October 1961 Special Systems features Bulletin, with multiple uses of decimal 32K and 65K as approximations to binary powers, e.g., “This switch has a 65K position and a 32K position.” The claim of binary use is unfounded. Thunderbird2 (talk) 15:23, 11 August 2008 (UTC)

The claim that is unfounded is your claim Thunderbird2 and this is because of exactly the same reasons as posted above by SWTPC6800, you are still wrong and now I see you are willing to violate 3RR to force your edit into this article. I again note that you have still not not answered the direct challenge put to you above (22:26, 10 August 2008). Fnagaton 16:06, 11 August 2008 (UTC)

Shreevatsa makes it clear why 32K is inherently ambiguous and therefore why the dubious cite is valid. Actually any usage 32K or less is inherently ambiguous without asking the author what he/she meant. SWTCP6800's comment is interesting but mainly irrelevant to this discussion. I do suggest the sentence be changed to something like:

Ambiguous usage of K meaning either 1024 or 1000 as in a "32K store" exists as early as 1960.

Also, as a part of this edit war, Fnagaton's reversion wiped out a cite to the Bell 1964 article without stating any reason; anyone know any reason why it shouldn't be reinstated other than TMI?

And if Fnagaton reverts the dubious cite then he clearly is heading towards an WP:3RR violation. Tom94022 (talk) 16:24, 11 August 2008 (UTC)

Tom94022, I restored the article to the last good version edited by Zedlik. From that edit comment you would be correct to assume that I reviewed your edits and those of Thunderbird2 and found they did not improve the article. The same challenge to Thunderbird2 (made at 22:26, 10 August 2008) applies to you Tom94022, would you care to substantiate your position by supplying exact quotes from the source and the reasoning? As Swtpc6800 says, there is no problem with the citation as it is. Fnagaton 16:38, 11 August 2008 (UTC)

The quote is above, provided by SWTCP6800 - would you like me waste space by repeating it? The usage of 32K and 8K therein, in and of themselves, are ambiguous which is now agreed to by Shreevatsan, T2 and me. I actually made it a link in the article because that's what it appears to be to me. Would a better editor please fix the link so that it links to this section. Fnagaton, care to discuss or are you just going to revert after unilaterally deciding the merits? Tom94022 (talk) 17:12, 11 August 2008 (UTC)

That quote does not support your position. I am fully prepared to discuss if you or Thunderbird2 actually provide a substantive argument and actually cite the exact parts of the source. So far neither of you have answered the challenge made above and provided a substantive argument, you attempted to answer it with a vague "The quote is above, provided by SWTCP6800" but as I said that does not support your position. Have you actually read the entire text of the source? You are also wrong to try to claim I've unilaterally decided merits because two other editors have reverted those same edits by Thunderbird2. So are you going to retract your obviously false statement? Fnagaton 17:26, 11 August 2008 (UTC)

Are u serious? Yes I have read the entire letter and unless I missed something the cited quote is the ONLY place in the letter that the symbol K is used in a sense that could be either binary or decimal. It is your POV that the K in 32K is binary but there is no evidence to support your POV. Shreevatsan, T2 and I all agree that it is ambiguous and have provided evidence as to why. In your usual fashion you discount all evidence that does not support your position and provide no evidence in support of your position. Please state why in the cited letter, 32K is unambiguously a binary usage. With regard to the Bell cite, you did revert it, unilaterally - nothing misstated, but in yr usual style you accuse someone of being wrong and demand an apology. Tom94022 (talk) 18:23, 11 August 2008 (UTC)

• The context of the citation, a 1960 letter-to-the-editor in Communications of the ACM is clear. The paragraph reads as follows:

Next, size of high speed storage. Dr. Booth regards 4096 words of 40 bits as larage (i.e., 163,840 bits): “…only in the last couple of years have machines been produced with storage organs whose total capacity is of this order of magnitude and many machines which are currently manufactured have high-speed storage for only one hundredth of this number of data.” The 8K core stores were getting fairly common in this country in 1954. The 32K store started in mass production in 1956;…

This is a discussion of the original “core” memory: magnetic ferrite core memory that was hand-threaded. My computer programming friend has a frame or two of the stuff. And more importantly, such memory was always binary in quantity (it could only be that way).

Arguing against this fact is absurd for two reasons: 1) the citation speaks precisely to the issue of “K” being used in a binary sense and slapping {dispute} tags flies in the face of the obviousness of it, and 2) the nature of computer construction and the convention for describing “high-speed storage” capacity in the binary sense had been carried forward to solid-state silicon memory chips for a long, long time. Please stop disrupting Wikipedia to make a point. If you want to hop into a time machine and change reality, do so. Until then, I’ll have none of this effort of trying to deny reality; no editor has to put up with absurdity. Greg L (talk) 18:54, 11 August 2008 (UTC)

The problem with your argument is that all core memory was NOT binary in quantity, for example, the IBM 705! Shreevatsan and T2 give you examples of such non-binary usage of K which you (and Fnagaton) choose to ignore. Only the author of the letter knows how he meant to use K, and in the absence of an explicit statement 32K and 8K are ambiguous. I am restoring the POV link Tom94022 (talk) 19:23, 11 August 2008 (UTC)

You or Thunderbird2 have still not quoted the exact parts of the source and provided a substantive argument regarding your position. It is not a case of ignoring the examples you claim to have made, the examples you claim to have given do not fit the facts presented in the source. So you are wrong to misrepresent the situation in that way. You did not mention "the Bell cite" in specific relation to "unilaterally", what you did was to make a similar change to Thunderbird2 and with that you implied I was the only one reverting that type of change. You also claimed "without stating any reason" and obviously you are wrong because the edit comment explains why. You are still wrong to imply that for the reason already given above. You are still wrong to misrepresent the situation. Fnagaton 20:56, 11 August 2008 (UTC)

• Tom94022, it doesn’t matter whether all core memory is binary, the vast majority of it is and you should know that. And if you don’t, then, IMO, you lack sufficient understanding of the subject matter to be weighing in here with {dispute} tags; you just degrade articles that most everyone else can understand just fine. Further, as to your statement that “only the author of the letter knows how he meant to use K”, that is such an unfathomly ridiculously thing to write; for the only values the author wrote about are binary! Do you think no one else here can read? The author wrote of 4096 words that were wired for 40 bits each. Or do you think that you can somehow frame how the new burden of proof is such that no letter writer’s intent can ever be known—even if you read it and their words are clear glass? Nice try, but no.

  If you think 163,840 bits aren’t binary, go “correct’ the math articles on Wikipedia too. And after those are all corrected, then come back here and “correct” this article with your new math. The issue at hand is whether “K” was used in a binary sense back in the late 50s and early 60s. Anyone with the common sense God gave a goose knows that’s true. So as regards the suitability of using this citation to support that fact, the second issue is whether this particular letter to the editor was using “K” in the binary sense. Well… Duhhh, what part of “4096” do you not understand? Please stop being disruptive here. Greg L (talk) 21:18, 11 August 2008 (UTC)

Both of you seem to like to shout that only your arguments make sense but that doesn't make it so. Also consistency of argument doesn't seem to be a strong point.

Fnagaton thanks for admitting you unilaterally reverted the Bell portion of my edit. In good faith, perhaps u should restore it? BTW, GregL gets the parts of the quote we think to be ambiguous, why can't you?

GregL, you are the one who said core was always binary and your ad hominem response to being corrected is not particularly persuasive; in fact, given your demonstrated lack of knowledge of the early computer history you might consider recusing yourself. And what is binary about 40?. The demonstrated fact remains that in that time period (late 50s and early 60s) 32K was ambiguous absent some other descriptive material (as in Amdahl and Bell) or some larger number (as in 128K). Tom94022 (talk) 21:41, 11 August 2008 (UTC)

I note you still have not answered the challenge put directly to you. It is obvious that you have not made a substantive argument and I also note that instead of answering the challenge you are now trying to argue about shouting instead. I also note you have continued to misrepresent the situation with your "Fnagaton thanks for admitting..." untrue statement. I note your attempt to change your position as your admission that what you originally wrote is wrong. Are you going to stop misrepresenting the situation and provide a substantive argument? Are you going to stop trying to make personal attacks and actually tackle the real subject? Your question "And what is binary about 40?" is completely irrelevant to this topic and it also highlights how you have failed to answer another question put directly to you, namely the question regarding 4096 in Greg's post above. So, from your posts we have 1) Failing to answer questions 2) Failing to provide a subtantive argument 3) Attempted personal attacks and attempted misrepresentation. Fnagaton 21:50, 11 August 2008 (UTC)

• I can’t help it if you, Tom-whatever don’t understand that “4096” is binary. Your assertion that the letter writer didn’t mean binary math when it’s obvious on the face of it is absurd. Greg L (talk) 21:53, 11 August 2008 (UTC)

• P.S. You and your cohorts have junked this whole article up. It needs a massive cleanup to get it so it can pass for something remotely neutral and factual. But unlike the tactic I’ve seen others take a fancy too, I don’t tend to junk up articles with {neutrality} tags just to pout until I get my way. I do prefer to work in good faith in a constructive fashion if other editors are willing to do the same. Do tell, would you like me to go through this article with a fine-tooth comb and fix everything that isn’t objective? That’s not a threat, by the way; it’s a promise, that I am perfectly willing to make Wikipedia a better place and fix all the bias you and others have introduced into this article. But I’d rather not devote all that time at this moment. Greg L (talk) 22:01, 11 August 2008 (UTC)

Let me try to keep it simple; there are three ways to use prefixes, all with relevant examples well established in the then contemporaneous art, namely, rounding, truncating and binary, as follows:

Exact Number	Integer Value Rounded by K=1,000	Integer Value Truncated by K=1,000	Integer Value Divided by K=1,024
8,192	8K	8K	8K
16,384	16K	16K	16K
32,768	33K	32K	32K
65,536	66K	65K	64K
131,072	131K	131K	128K

The only two prefixed numbers in the 1960 Gruenberger letter are highlighted in yellow above so there is insufficient information in the letter for a neutral observer to conclude what the author meant at that time. The fact that elsewhere in the letter the author uses the decimal number 4,096 is no more relevant than his usage of 40 or 140,000 or any other decimal number, whether divisible by 8 or not. There is no dispute that the author could multiply by 8 or that he was aware that some memories came in binary increments; the question is what did he mean by K in this article at that time. There is evidence that other contemporaneous authors used K with truncation. If either of you have any arguments about why the letter is not ambiguous I would be interested in hearing them, but so far all your shouting adds up to is that it is obvious to you - it is not obvious to at least three other editors. The above table is mathematics so there should be no dispute about what the symbols might stand for. Any neutral observer should conclude that the the use of the prefix K is only unambiguous when applied to known values of 65,536 or higher and that the usage of only 32K and 8K in the 1960 Gruenberger letter makes his usage ambiguous. Tom94022 (talk) 01:26, 12 August 2008 (UTC)

The fact the author uses K is significant considering the context of the letter in the first place. Since the context of the letter makes it obvious that 4096 being binary is used with K means that the simplest explanation is that K is being used in a binary sense. Are you aware of Ockham's razor? You are not applying it to this subject. You want us to believe the author was using some other rounding form then you have to show evidence of that (not just your personal opinion), you have not ergo you are wrong to add it to the article. Fnagaton 08:24, 12 August 2008 (UTC)

I added a reference from 1959 that uses "32k core" on an IBM 704 computer. The IBM 704 core memory units had 4096 36-bit words. Up to 32,768 words could be installed.

• Real, P. (September 1959). "A generalized analysis of variance program utilizing binary logic". ACM '59: Preprints of papers presented at the 14th national meeting of the Association for Computing Machinery. ACM Press: pg 78-1 - 78-5. doi:10.1145/612201.612294. On a 32k core size 704 computer, approximately 28,000 datum may be analyzed, … without resorting to auxiliary tape storage. {{cite journal}}: |pages= has extra text (help)

-- SWTPC6800 (talk) 01:43, 12 August 2008 (UTC)