Wikipedia talk:Manual of Style (dates and numbers)/Archive B10

From Wikipedia, the free encyclopedia
Jump to: navigation, search

Proposed change to wording of "Binary prefix" (3/3): continuing discussion

There is no basis for a policy that deprecates Binary SI units

Am I missing something, but since, (MOSNUM:Which system to use), requires that editors should use the units employed in the current scientific literature on that topic and since the IEEE Computer Society is one of the leading such publisher of such literature, deprecation seems to be in violation of stated policy and more the POV of a rabid editors such as Fnagaton. This whole thread started when Tbird tried to clarify the existing policy allowing Binary SI units since Fnagton had turned the policy on its head, misinterpreting the plain language of the existing policy to not allow Binary SI units. I for one will not support any policy that deprecates Binary SI units. Tom94022 (talk) 00:17, 22 March 2008 (UTC)

  • Interpolated note: GUIDELINES ARE NOT POLICIES. Please do not confuse them. To raise every minor guideline issue as a huge "policy" fight is the fallacy of appeal to emotion if you are doing it intentionally. If you are not, then please study WP:POLICY really thoroughly before engaging in any more heated debates on MOS* topics. Most of these things rage on and on pointlessly (and repeatedly, reappearing every few months to rehash the same arguments) because some people are not taking the time to fully understand the purpose of the MOS and its nature. These errors in perception range everywhere from "this is going to be a binding policy like WP:COPYRIGHT so I have to fight it to the death because the topic in question is a pet peeve of mine" all the way to "MOS has to say exactly what [my favorite paper-based style manual] says because it is the same thing with the same purpose - it's a general style guide for the world to use in all contexts", among many other mistaken notions. I don't mean this to sound too harsh - I made a mistake or two like this a year or more ago when I first ventured in here. Learned fast. — SMcCandlish [talk] [cont] ‹(-¿-)› 00:17, 5 April 2008 (UTC)
  • P-u-h-l-e-e-ze! Who are you trying to kid? Do you think the rest of us here are not going to read and parse the logic of what you just wrote? The goal in all technical writing is to clearly communicate to the intended audience with minimal confusion. Do you really expect us to believe that the typical reader of the computer-related articles here on Wikipedia are members of Institute of Electrical and Electronics Engineers or the International Electrotechnical Commission? That’s a metric ton of weapons-grade bullonium and you know it Tom. Other encyclopedias don’t routinely use the IEC 60027-2 binary prefixes in numerical equivalencies to denote the capacity of computer memory and storage for one reason only: because such symbols aren’t used in the computer industry or their packaging, manuals, or advertising, nor do any general-circulation computer magazines use them. Consequently, the terms are unfamiliar to the typical general-interest reader. This much is just too obvious.

    Finally, as to your characterization of Fnagaton as a “rabid” editor pushing a POV, I suggest you go cool off in a corner somewhere and come back when you have something to add that can remotely be considered as constructive. Greg L (my talk) 02:22, 22 March 2008 (UTC)

  1. Software and research papers use them.
  2. Fnagaton should have been blocked with Sarenne. — Omegatron 04:05, 22 March 2008 (UTC)
Your repeated misrepresentation, threats about blocking and personal attacks show that you are continuing use harassment against me instead of tackling the issue. The fact remains you are in the wrong so stop using personal attacks otherwise I will be forced to report you again. Fnagaton 11:05, 23 March 2008 (UTC)
Indeed. I don't agree with Fnagaton on this, but I hardly need to suggest that he's got a crazy-making brain disease to make a case that he's misinterpreting how to go about changing the guideline. The fact that Fnagaton is passionate about this issue, as others have been before (on both sides) has nothing to do with the validity of their arguments either way. Having been accused of WP:DE simply for being passionate and steadfast myself in the past, I sympathize in a Voltaire way - I defend Fnagaton's right to express what he is thinking (civilly), but if I disagree with his logic I'll certainly say so, since that's where the reason in argument is. Debate by flamethrowing is unproductive pen...sword-waving. I.e., everybody please chill. — SMcCandlish [talk] [cont] ‹(-¿-)› 00:36, 5 April 2008 (UTC)
Please be respectful toward other editors. Tom is correct that the IEEE-CS is one of the leading publishers of scientific articles related to communications and computing (the IEEE as a whole may even be the leading such publisher), and that (not society membership) is relevant to the standard set out in the "Which system to use" text that you yourself have previously cited. I do not know what units are the most common in IEEE-CS papers.
And while I am not singling anyone out, it is a factual statement that a number of editors have been pretty rabid about removing IEC units from the encyclopedia, and some of them have in fact misinterpreted or misrepresented the contemporary wording of MOSNUM to be more strongly anti-IEC than was intended. Virtually any time anything related to binary units is brought up on this page, we hear the refrain of "IEC units must be banned", and then an argument usually ensues. — Aluvus t/c 06:49, 22 March 2008 (UTC)
  • Omegatron, Regarding point #1, I’m sure that’s true. But you know as well as I do why that’s not enough. As for point #2, that’s water under the bridge and is no relevance to moving forward with developing a MOSNUM policy that makes better sense than what we currently have. Greg L (my talk) 04:13, 22 March 2008 (UTC)
Greg L, since when do "packaging, manuals, advertising or general-circulation computer magazines" constitute scientific literature? I don't believe the standard is the knowledge of the general purpose reader; my wife, for example, college educated and a competent computer user doesn't know the difference between byte and bit, much less MB vs MiB vs 220B. And given her college experience, she well knows kilo = 1,000. All technical articles run into this problem and need linkages and/or footnotes to help such users. If an editor chooses to use Binary SI prefixes either as first article or for disambiguation, IMO, a link is the best way to clarify the meaning for the "general purpose reader." BTW, I think what i wrote is rather straight forward and doesn't need any parsing, perhaps u should cool down. Also, since at least Omegatron, TBird2 and I don't support deprecating, there is no consensus to change the currently posted MOS, agree? Tom94022 (talk) 05:25, 22 March 2008 (UTC)
The SI system usually comes under consideration when debating policy here, but WP is under no obligation to follow it. SI is made for numerous modes and registers, whereas WP is purely online and for a wide range of readers. Tony (talk) 07:56, 22 March 2008 (UTC)
  • You’re taking the “scientific literature” term far too literally Tom. It means the real-world writings the typical Wikipedia reader will likely encounter elsewhere and that obviously doesn’t mean draft standards and policy papers from the International Electrotechnical Commission; that much is just common sense as it applies to “communicating with the intended audience.” The lead-in to the draft proposal laid out the premiss for the policy clearly enough: “The 1999 IEC proposal on binary prefixes (IEC 60027-2) … has not been widely adopted by the computing industry and trade magazines and is therefore unfamiliar to most readers.” In this context “real world” as it applies to our typical reader obviously means all the brochures, data sheets, advertisements, owners manuals, and magazines one encounters in real life.

    Note footnote #1 in the ‘Seagate’ example; I lifted that text verbatim right off the Seagate data sheet for that hard drive. That’s how the “real world” disambiguates the term “GB.” That’s what readers are exposed to in real life. Few knew here at MOSNUM when we permitted use of the IEC 60027-2 prefixes that Wikipedia would effectively be marching off into the snow proudly blowing our “Follow Me!” kazoos, only to find that no one else in the industry nor any other encyclopedia or magazine was inclined to even look out of their cabin windows at us on this one. Wikipedia is alone on this and it’s unwise to continue to embrace such a failed policy.

    Now, having read your argument above, as well as your “straight forward” [sic] words on topics as varied as describing another editor being “rabid”, I think you’ve made your feelings abundantly clear here. You believe explaining the magnitude of values using terms like “GiB” is a good idea and you won’t be changing your mind any time soon. I’ll assume we can mark you down in the “Don’t like this proposal” column. Fair enough? Greg L (my talk) 09:23, 22 March 2008 (UTC)

You can mark me down as opposing this proposal and any other that deprecates SI Binary Prefixes. By my count there are now four editors so inclined. BTW, you really do need to cool off, "... marching off into the snow proudly blowing our “Follow Me!” kazoos ..." - give me a break. 22:42, 22 March 2008 (UTC) —Preceding unsigned comment added by Tom94022 (talkcontribs)
  • I’m not at all upset up on this. Your unfounded suggestion that I am does not establish you as the wise, cool-headed voice of reason here Tom. I simply think your position on this is unwise and the current policy does readers of Wikipedia a disservice. Note that I haven’t written any computer articles here on Wikipedia so I don’t have a vested interest in this as an editor. I subscribe to several computer magazines and daily read on-line computer magazines. The first time I encountered the IEC prefixes was here on Wikipedia. This is obviously my personal opinion, but my reaction that first time was to to bury my head and say “Oh, geeze” to myself. Embarrassment. Greg L (my talk) 23:33, 22 March 2008 (UTC)
Tom is correct that your rhetoric and dismissive attitude are not helpful (nor are they new to discussions of this issue). It would be much more productive if you put less effort into denigrating opposing viewpoints and more effort into explaining how advertising is "scientific literature". Additionally, not having worked on any computer articles in the project is not the positive you seem to think it is; if anything, it suggests you may not have had to actually deal with this issue in a meaningful way and undercuts your "I am absolutely right" stance. There are certainly people whose opinion on this issue is colored by their experience editing a subset of Wikipedia's computing articles, but that does not somehow turn a complete lack of experience into a positive. — Aluvus t/c 01:55, 23 March 2008 (UTC)
  • We’ll have to agree to disagree on how one argues a point Aluvus. There is a big difference between attacking the individual and exposing the shortcomings in his arguments. That how debate is accomplished; you know that as much as I do. I can’t help it if my choosing to avoid authoring Wikipedia articles on computers doesn’t impress anyone; the point is, I’m am not coming into this debate with an agenda to protect prior work. As for “scientific literature”, as I stated before, I withdrew that verbiage from the proposal over 24 hours ago (see the below post) as it just detracts from the obvious point: the IEC prefixes are unfamiliar terms to the typical reader and only adds to confusion. Efforts to obfuscate on this, as others have done by rhetorically asking “what magazines?” (what parallel universe?) will prove unsuccessful in overcoming this obvious reality. Trying to “have it both ways,” as you pointed out, doesn’t work and continued use of terms like “GiB” doesn’t make Wikipedia look proffesional, it makes Wikipedia look foolish IMO. You previously stated that my latest proposal was the “least bad” one you’ve seen advanced so far. Does that mean you are disposed to support it in an up-or-down vote? Greg L (my talk) 03:19, 23 March 2008 (UTC)
You do not get to "agree to disagree" your way out of WP:CIVIL. And rhetoric about kazoos does not expose the shortcomings in any argument but your own. The entire thrust of your argument appears to be "it is obvious that using IEC prefixes is bad", but the very fact that there is disagreement here should indicate to you that that is not obvious to all parties. Numerous variations of that same tactic, often combined with the same dismissive attitude, have been tried on this page in the past, and they have consistently served to lower the level of the discourse. One of the primary reasons that discussions of this issue have repeatedly broken down is that people have dismissed out of hand any viewpoints they don't agree with, which makes it difficult or impossible to build genuine consensus. And as a final note, it is inappropriate to attempt to pressure me into voting in your favor, and also inappropriate to call a poll on a proposal that by all appearances is still in flux and has not been discussed. — Aluvus t/c 08:15, 23 March 2008 (UTC)
  • I don’t think I am being uncivil at all. If one reads WP:CIVIL (as I just did). It says “[Incivility] is calling someone a "crank" "moron" or "POV-pusher", among others.” It further specifically defines it as “personally targeted” attacks. I have consistently only gone so far as to ridicule some others’ arguments, not them personally. That is “debate.” I am now taking deep interest in your lack of taking a balanced position throughout this debate. It comes across as having a strong position on this matter while trying to appear that you aren’t. Tom94022 in his 00:17, 22 March 2008 (UTC) post said “[pushing a] POV of a rabid editors such as Fnagaton” (my emphasis). That was so clearly a prohibited and over-the-line, direct personal attack on another editor and was absolutely unwarranted. And I responded (politely) to it. Yet you attacked me for admonishing Tom about his personal attack! In fact, you wrote “And while I am not singling anyone out, it is a factual statement that a number of editors have been pretty rabid about…”. In other words, you called another editor rabid and hid behind the apron strings of civility by stating “I am not singling anyone out…” when it was obvious that the individual being targeted was Fnagaton.

    I’ve noted a strong tendency on this page for you to 1) argue about “how” issues are debated here rather than the substance, and 2) for you to attempt to craft language that looks balanced and measured but really isn’t. You accused me of being uncivil when I clearly haven’t and yet defended and even participated via proxy in the most extreme practices of it. I am done arguing with you about "how” people make their points here—particularly when you try to use slight-of-hand to engage in personal attacks yourself. I will no longer be baited by your diversions. Please stick to arguing on the points. 19:43, 23 March 2008 (UTC)

The nutshell description of WP:CIVIL is: "Participate in a respectful and civil way. Do not ignore the positions and conclusions of others. Try to discourage others from being uncivil, and be careful to avoid offending people unintentionally." When you write "P-u-h-l-e-e-ze! Who are you trying to kid? Do you think the rest of us here are not going to read and parse the logic of what you just wrote?" or "metric ton of weapons-grade bullonium" or "come back when you have something to add that can remotely be considered as constructive", that is not a civil or polite way to deal with people. Ridiculing arguments is not civil and is the crux of why this issue has still not been resolved. I do not claim to be balanced or neutral on this or any other disagreement, but I am damn tired of seeing months-long angry arguments about it that go nowhere or turn into wars of attrition. I will think on what I have previously written; I would encourage you to do the same. — Aluvus t/c 23:02, 23 March 2008 (UTC)
  • Aluvus, I am quite disgusted with all this arguing with you about "how” people make their points here since recent history on this page shows that you have a far from stellar record in the “civility” department by flat out stating that other editors have been “rabid” and pushing a POV. That is absolutely prohibited behavior. As I stated before, I am entirely disinterested in engaging in all this name calling and I will no longer be baited by these diversions. Please stick to arguing on the merits of the proposal. You seem to be suggesting that the only reason people oppose deprecating IEC prefixes all boils down to “Greg L is a poopy-head in his advocacy and if it weren’t for that, I might support his proposal.” While that might be the case, mark me down as a skeptic on that one. Why don’t you just admit that you like the IEC units and don’t agree with any policy that calls for them to be removed from Wikipedia? I would very much appreciate a statement like that than have to endure any more of your games. Now please take your parting shot below and make it a good one as I will no longer be responding directly to you; it is clearly unproductive and fruitless. Goodbye. Greg L (my talk) 01:49, 24 March 2008 (UTC)
Attacking me does not accomplish anything. People oppose deprecating the IEC prefixes because they think the IEC prefixes are useful. Your berating those people seems to have done little to change their minds. If anything, it has simply made them oppose you more strongly, just like every time someone else has tried the same tactic before. I have tried to encourage you to actually engage people in discussion, or at the very least to direct your rhetoric at me instead of other participants, so that this might finally be settled. Because I can assure you, if you continue to belittle opposing viewpoints and tell people they should "admit" that they disagree with you, the people with opinions on this issue will just dig in their heels yet again (some of them already have; did you notice?). That usually leads to a protracted, mostly silly fight that in turn leads (at best) to a problematic compromise built on weak consensus, and then a few weeks or months later there is a new spark and the cycle begins again. This has happened over and over again.
  • I for one am tiring of this belligerent, hysterical, personalised approach, Alvulus: I must ask you to desist and to show more collaborative spirit. As far as I can see, Greg has done a fine job in trying to find a solution to this issue, and has shown a good deal more talent and imagination than you, who've done nothing but inject negative sentiment into the discourse. I can't find a scrap of your previous post that I agree with, and as for the edit summary—that's just ramping up you hysteria. STOP IT, please. Tony (talk) 06:03, 24 March 2008 (UTC)
You have an opportunity to build a stronger consensus and a lasting policy; stop wasting it. — Aluvus t/c 05:34, 24 March 2008 (UTC)
I agree with Tony that the behaviour of Aluvus is completely unacceptable and disruptive. Aluvus is being uncivil. Aluvus should strike his comments from the page and apologise to everyone and specifically to Greg, Tony and myself. Fnagaton 15:37, 24 March 2008 (UTC)
  • All: I realize now that quoting “use what’s in the scientific literature” is an ineffective and unpersuasive argument so I deleted it from the proposal. It was pulling the thrust of the predicate off track. As you can see, it now focuses exclusively on the “well… duhhhh” point, which is that if it’s not recognized by the typical reader and isn’t used in the real-world popular press and even the computer industry, it’s clearly of no value in an encyclopedia.

    I also added proposed text that specifically addresses how to deprecate the IEC 60027-2 prefixes from existing articles. I added this because I saw in some of Fnagaton’s writings that this has been a hot-button issue. I don’t know if this is of any real value, but it’s a starting point to build off from (or discard) and is now available for comment. My main motivation for adding it was so that this proposed policy directly addresses the subject of deprecating existing articles and can’t be interpreted as applying only new ones.

    Aluvus: I much enjoyed your 06:29, 22 March 2008 (UTC) post in which you stated…

I consider this proposal less bad than many that have come before, and an improvement on the current guidance. That is only because I dislike the "split the baby in half" solutions, and because there is a possibility that this would actually settle the matter for more than a month.
Although a damn funny way to characterize it, I can see the truth behind it; I suppose that’s what makes it humorous. I hope you still support helping to put this issue to rest once and for all. Greg L (my talk) 09:53, 22 March 2008 (UTC)
Checking the history Omegatron made this change to the guideline without talking about it here first. The changes are clearly pushing his POV and ignores consensus. The user then started to edit war when his change was reverted and made personal attacks and threats about using blocks to try to make sure his version stayed on the project page. These bad faith actions are documented in the ANI report. (talk) 12:35, 22 March 2008 (UTC)
I'm allowed to change the guideline without talking about it first. If you have a problem with my edits, you can then change them to make a compromise version that you agree with, or you revert them and then discuss your reversion. After discussion, we understand where each other is coming from and try to change it again, while working together. This is how editing works. See Wikipedia:Be bold and Wikipedia:Consensus. You don't just revert war everything back to your preferred version and claim that it represents community consensus.
And I seriously hope no one's trying to "deprecate" the standardized prefixes on WP. Please don't start that endless argument again. Sarenne was banned for pushing his POV all over the project, and you will be too. There is no site-wide consensus on use of prefixes, so they are decided on an article-by-article basis. In some articles they are appropriate and serve a useful purpose (List of device bandwidths, Floppy disk); in others it is acceptable to use the ambiguous units because they are familiar to that particular field (Commodore 64). — Omegatron 15:11, 22 March 2008 (UTC)
What do Omegatron's edits more than 3 months ago have to do with anything, Fnagaton? — Aluvus t/c 17:44, 22 March 2008 (UTC)
What do Omegatron's pathetic threats about "should have been blocked with Sarenne" have to do with this topic? They do not have anything to do with the topic of course, they are personal attacks and must stop.Fnagaton 11:58, 23 March 2008 (UTC)
Two wrongs do not make a right. — Aluvus t/c 23:02, 23 March 2008 (UTC)
Now you are using misrepresentation because what I have done is not wrong by any stretch of the imagination, so I demand you retract what you posted because it is untrue insinuation. What I have done is show why Omegatron is wrong and produced a better argument which helped result in the guideline being changed to something he doesn't support. You and Omegatron are in the wrong here as demonstrated by Tony's comment above. Fnagaton 15:34, 24 March 2008 (UTC)

The current wording is good, but needs some fixes

I made some small changes to neutralize the section, but there are bigger changes to be made that I'm obviously not going to be bold about:

There is consensus that editors should not change prefixes from one style to the other, especially if there is uncertainty as to which term is appropriate within the context

The consensus is not to be a fanatic and change everything to your preferred style. There is no reason why you can't change one style to another if there's a good reason to do so, as long as you aren't revert warring or going against the consensus on the talk page of the article you are editing.

and follow the lead of the first major contributor.

Should be removed. This is inherently anti-consensus, as I've explained elsewhere. Without global consensus, you decide based on good reasons, not on arbitrary precedent.

These replacements for the historical units have gained only limited acceptance outside the standards organizations. Most publications

This is biased ("only limited acceptance") and implies a level of certainty that can't exist. How do you define "acceptance"? Most what publications? Do multiple copies of the same publication count? I'm pretty certain most hard drive manufacturer publications use the standard prefixes. Who tallied them all up? What about software? What about the myriad Linux servers hosting all of the web pages you view? I see IEC prefixes on a daily basis. Should we go with your personal experience or mine? — Omegatron 15:48, 22 March 2008 (UTC)

I support the changes just made by Omegatron. Thunderbird2 (talk) 16:02, 22 March 2008 (UTC)
I support the changes just made by Omegatron. Tom94022 (talk) 16:47, 22 March 2008 (UTC)
I oppose the changes —Preceding unsigned comment added by DavidPaulHamilton (talkcontribs) 06:29, 23 March 2008 (UTC)
I oppose the changes for the reasons stated below by DavidPaulHamilton: The changes to the main page split the baby in half by using prefixes that are unknown to the general population. Greg L (my talk) 18:42, 23 March 2008 (UTC)
And what specifically do you disagree with? Bearing in mind that the points he raises here and the changes he actually made are quite different, of course. — Aluvus t/c 07:22, 23 March 2008 (UTC)
The changes to the main page split the baby in half by using prefixes that are unknown to the general population. DavidPaulHamilton (talk) 09:29, 23 March 2008 (UTC)
The existing wording already did those things, so you have not provided any reason to oppose Omegatron's changes specifically. Omegatron's changes were minor wording tweaks that had no impact on what the guidance actually tells people to do. In what way was the previous wording superior? — Aluvus t/c 22:10, 23 March 2008 (UTC)
  • Omegatron: You’re setting an unnecessarily high hurdle when you challenge opponents of your view to enumerate every single publication on Earth that uses one style or another. Clearly, a common-sense test should suffice and any quick inventory of the general-interest on-line and in-print computer publications demonstrates that “GiB” isn’t currently being used and that’s why such terms are unfamiliar to most readers. The test shouldn’t be what policy best makes a few Wikipedia editors happiest, it should only be whether or not Wikipedia is best serving the interests of its readers.

    And please leave the formatting and style of the proposal alone. When you deleted all the horizontal rules and made sections within the proposal their own “====” article sections, it made it even harder to discern where the proposal starts and ends, which is already hard enough given the length of some of these threads. Greg L (my talk) 18:48, 22 March 2008 (UTC)

Oppose - Omegatron's changes to the guideline and those talked about here (as Aluvus said they are quite different) are not an improvement. A couple of direct questions to Omegatron. 1) Why do you make changes to the guildeine without talking about them first? 2) Are you interested in seeing a fair and balanced guideline? Fnagaton 11:00, 23 March 2008 (UTC)
Let’s try to find some common ground, and tackle this one step at a time. The first edit Omegatron made was an attempt to remove the POV from the opening sentence, which right now reads:
  • In computing, where binary numbers (powers of 2) are often more useful than decimal (powers of ten), the de facto standard for binary prefixes and symbols (collectively, "units") is to use kilo-, mega-, giga-, with symbols K, M, G, where each successive prefix is multiplied by 1024 (210) rather than SI's 1000 (103). To attempt to resolve this ambiguity …
This gives the unwarranted impression that the binary use is more common than decimal. I don't recall Omegatron's precise phrasing, so I'm going to attempt my own. How about:
  • In some fields of computing, binary numbers (powers of 2) are more useful than decimal (powers of ten). In others the reverse is true. As a consequence there are two different de facto standards for defining symbols K, M, G (representing prefixes kilo-, mega- and giga-, respectively), one in powers of 1024 (210), and the other following the SI prefixes convention using powers of 1000 (103). To attempt to resolve this ambiguity …
Constructive comments are requested from, and a happy Easter is wished to, all at MOSNUM. Thunderbird2 (talk) 13:36, 23 March 2008 (UTC)
I want Omegatron to answer the questions put to him first before I respond to your points Thunderbird2 because the conclusion from what you posted is contingent on what he writes. Fnagaton 16:40, 23 March 2008 (UTC)
Your first question is a red herring, as it always has been; no one is obligated to get permission on Talk before making minor changes. The second one should be easy for you to resolve yourself by assuming good faith. So no, nothing of any importance really hinges on those questions. If you have an opinion on what Thunderbird2 has written, you may as well get on with it. — Aluvus t/c 22:10, 23 March 2008 (UTC)
I'm asking Omegatron, not you Aluvus. Unless you are now admitting you and Omegatron are one and the same? As for your answers they are easily demonstrated to be wrong. Firstly look at the header of the project page "Edit this page through consensus, discussing proposed changes on the talk page first." Since the changes he made were not minor, as demonstrated by the fact that you are wrong when you say "no one is obligated to get permission on Talk before making minor changes". Secondly, assumption of good faith is not the catch all answer to everything and my question is a valid concern about what Omegatron wants to see in the guideline. So my questions stand for Omegatron to answer. As I say my responce to Thunderbird2 is contingent on what what Omegatron writes in reply. Depending on what he writes another couple of questions may also be needed. Fnagaton 15:13, 24 March 2008 (UTC)
Unless you are now admitting you and Omegatron are one and the same?
Is this seriously how you think? You really think that anyone who opposes something you do on Wikipedia is part of a vast conspiracy of sockpuppets? What if they are merely individual people who disagree with you on a particular issue for good reasons? Aluvus and I are not the same person, and do not have the same opinions on this subject, as far as I can tell. Please try to understand where we're coming from, listen to our rationales and reasoning, and suggest ideas that find a common ground between all of us. Work together. Stop treating Wikipedia as a battleground. — Omegatron 14:47, 30 March 2008 (UTC)
That's rich coming from you who has previously made pathetic threats about "blocking" and personal attacks just because I have presented a better argument than yours. It is a reasonable assumption to make since you and that user were both using personal attacks and the user answered on your behalf, also it is reasonable since there has been widespread sock puppet activity around this subject which is currently under investigation. When you apologise for those threats and personal attacks then you will show you are interested in reasonable debate. Fnagaton 14:52, 30 March 2008 (UTC)
For reference, this was actually the second edit he made, and the appropriate diff is here. — Aluvus t/c 22:10, 23 March 2008 (UTC)
Omegatron made three quick changes, not just the one you linked. Fnagaton 15:13, 24 March 2008 (UTC)

I was just reading the latest issue of the IEEE Computer Society magazine, Computer. They don't use the IEC binary prefixes. They still use Mbyte and Gbyte. I did a search of the IEEE publication database last year and the IEC binary prefixed are rarely used in IEEE magazines and journals. (I mean RARELY.) -- SWTPC6800 (talk) 04:12, 26 March 2008 (UTC)

Is there a link so that I can buy a copy? DavidPaulHamilton (talk) 08:40, 26 March 2008 (UTC)

Mebibytes are not hard to find:
Thunderbird2 (talk) 17:31, 26 March 2008 (UTC)
And also Intel, Dell, Freescale, IBMWoodstone (talk) 20:03, 26 March 2008 (UTC)
With Google even the -rarest- examples are "easy to find". That does not say anything. Only the ration of acceptance is an indication Mebibytes gives just 20.800 hits, Megabytes gives 5.840.000 hits. Mahjongg (talk) 12:59, 27 March 2008 (UTC)
Mahjongg, it is unbelievable how someone could be completely missing the point after this lenghty debate. The Megabytes you are counting include a lot of text that uses the term "Megabyte" in accordance to IEC 60027-2 and that includes even text that was written long before this standard was published. Nobody here argues whether we should use "Megabyte" or "Mebibyte", we are discussing whether Wikipedia should use all of these terms in accordance to IEC 60027-2 and pre-IT standards or whether it should use the IT industry slang meaning of "Megabyte". Therefore it is absolutely pointless, useless and at best misleading to count use of "Megabyte" against "Mebibyte". It is also obvious that "Megabyte" and "Mebibyte" will be rarely spelled out, so that MB and MiB will be far more common in written language whereas "Megabyte" and "Mebibyte" are mostly used in verbal language hence Google statistics won't get you anywhere. -- (talk) 14:07, 27 March 2008 (UTC)
That argument has already been refuted in the top most section of this page. It is fallacious to say "Megabyte" and "Mebibyte" when it is obvious that only Mebibyte is rare compared to Megabyte. As for trying to search for MB and MiB that generartes too many false positives with "Men In Black" for example. Fnagaton 14:13, 27 March 2008 (UTC)
The argument is kinda valid in that those measurements are all meaningless either way. I suspect that our real-world experiences, if we're all honest, lean heavily towards not generally using the IEC prefixes. I'm going to try a new approach to trying to figure out where everyone stands shortly... in doing so, I'm not going to try and change the proposal above or claim there's overall consensus about anything, just figure out what is actually still left to argue about. SamBC(talk) 14:27, 27 March 2008 (UTC)
Fnagaton, I wasn't suggesting to use some magic search terms that will show us the real reality. I was merely stating that Google can be used to bend the perceived reality in either way. For that we certainly don't need to show that the IEC prefixes are more common, something only a loony would claim, but simply that their use is increasing seemingly indicating progressive adoption. In other words, I was just pointing out that Mahjongg's comment was neither constructive nor helpful but showing a very superficial view on this issue despite the fact that he wouldn't have to read beyond this page to get the bigger picture and lot of pointers to verifiable information...if you catch my drift. -- (talk) 14:29, 27 March 2008 (UTC)
Tell everyone here exactly how posting "it is unbelievable how someone could be completely missing the point" is constructive? Mahjongg's post is concise and to the point. Fnagaton 14:40, 27 March 2008 (UTC)
I'm not "missing the point", (and for the record, I just re-joined this discussion yesterday, since I have given up on it half a year ago, but was curious about it's status quo. Can't believe the Mebibyte pushers are still at it. This issue should have been resolved by now) I just tried to illustrate the futility of giving a few examples you can find on the net of publications that use this term. As you can find -anything- with Google on the net, even the most obscure subject. And if you use these finds to try to claim the term is "in common use", then you are intellectually dishonest IMHO. Did I make a "futile remark at this stage of the discussion"? Perhaps, but it seems most of the discussion here seems futile, as it seems to go on and on without resolving the issue.
By the way, it seems telling to me that the given examples are seemingly not coming from mainstream technical journals etc, but rather from "legal departments", and such, at least if I look at the names ending with "et al" etc. this seems clear to me. Which makes kind of sense, as only a legal department would use terminology that overrules common sense for the wish to use "watertight language". But most of the worlds "normal" population uses the more common terminology. Perhaps you could claim that in articles with a "legal" background the use of Mebibyte (etc.) is more or less acceptable, but not in the much more common articles of a technical, (not legal) nature. Mahjongg (talk) 14:57, 27 March 2008 (UTC)
In contrast to your previous comment this one was indeed helpful in certain way. -- (talk) 15:11, 27 March 2008 (UTC)

This whole debate fatigues me. Sigh. Fnagaton, you said above that you asked me some questions that I did not answer. I'm not sure what you're referring to. Can you restate them here in a list so that I can answer them? — Omegatron 14:39, 30 March 2008 (UTC)

Omegatron: You’re setting an unnecessarily high hurdle when you challenge opponents of your view to enumerate every single publication on Earth that uses one style or another. Clearly, a common-sense test should suffice and any quick inventory of the general-interest on-line and in-print computer publications demonstrates that “GiB” isn’t currently being used and that’s why such terms are unfamiliar to most readers.

I'm not setting any hurdle. I'm saying that the current wording implies that there is certainty about relative usage that can be interpreted a huge variety of ways depending on personal viewpoints. Are there more publications using "KB" to mean 1024 than those that use "KiB" to mean 1024? Probably. But what about software? Are there more publications using "MB" to mean 1,000,000 than "MB" to mean 1,048,576? Probably. Are there vastly many more publications using "k" as a prefix to mean 1,000 than those that mean 1,024? Definitely. So which do we mention? Why does this even need mentioning here?
All I'm saying is if something like this appears in the guideline, it should be reworded either to say something more specific "The majority of Commodore 64 manuals use the traditional prefixes", or something less specific "Despite the standardization, SI prefixes are still used to refer to binary multiples in many contexts", without claiming more certainty than actually exists. I'm not sure why it's even necessary to have this sentence, though. — Omegatron 15:03, 30 March 2008 (UTC)

And please leave the formatting and style of the proposal alone.

I don't know what you're referring to. — Omegatron 15:03, 30 March 2008 (UTC)

Standardisation is a good thing

I was asked to come here and comment by Greg L, probably since I am an old software engineer and not a mathematician. But he is not going to like my answer:

I grew up and still live in Sweden, where as far as I know we switched to SI units for most measurements already in the 1800s, an switched to SI notation for date and time somewhere mid 1900s or so. Since they are not ambiguous and can more easily be handled by computers, such as sorting and converting them. The Japanese and many other countries did the same modernisation long ago too.

So when I write date and time I write 2008-03-23, 18:00. I do not write 23 March 2008, 6 p.m., or should that be 6 a.m.? And if I use the old Swedish notation 6 f.m./e.m. then I guess most of you have no idea of what that means, right? See, a.m./p.m. is even language specific! See the benefits of standardisation?

So basically, I disagree with Greg. I like the new SI units GiB for 230 bytes. Since it is less ambiguous. I just wish we had a non-ambiguous unit for 109 bytes too, but currently we don't.

When we use GiB in articles then we need to put a footnote or paranthesis explaining that 1 GiB = 230 bytes = 1,073,741,824 bytes, since as you people have mentioned here many readers do not know what it means. But if we use GB then we have to put a footnote anyway to tell which GB we mean. So there really isn't much difference. So I say we are an encyclopaedia and should go with the best possible notation, at least as long as it is reasonably easy to understand. (I would not like to see room temperature written in Kelvin...)

--David Göthberg (talk) 2008-03-23, 02:16 (UTC)

Okay, with the risk of starting a flame war: I feel I have to say something very unpleasant. I think Greg L or someone posing as him [is trying to "stuff the ballot" on this. (Perhaps to make Greg look bad? So it might be someone opposed to him. Conspiracies galore! Now it suddenly feels more like fun than unpleasant. #:))
Seemingly while he still thought I preferred GB over GiB "he" left this edit on my talk page. Note that he is asking me to come vote in his favour when the time comes, and asks me for others that might have the same point of view. And especially note that the edit is done by an IP user so will not be visible in Greg L's "user contributions", still he went to the trouble of manually adding his full signature.
Please try to stick to logical arguments instead of trying to stuff ballots on this, whoever you were that left that edit. Okay?
Although GiB vs GB is perhaps more a matter of personal taste than logical arguments so it is a tricky case. (Since the logical arguments there are goes both ways, so no side has a clear advantage.)
--David Göthberg (talk) 03:50, 23 March 2008 (UTC)
David, the issue here is context. If I told you my computer has 3 GB of memory, would you (or anyone on the planet, for that matter) believe that I am saying that I have 3,000,000,000 bytes of memory? No, of course not. "GB", in the context of memory, is always, 100% of the time, understood to be in powers of two.
Wikipedia's responsibility to the world is to present the world as it is, not as we as individuals want it to be. We as people with the goal of writing a great encyclopedia are naturally idealistic, but the content we produce must be pragmatic and realistic. This means that we don't push our views on our readers; this extends to what we as individuals may think the "right" unit of measurement is, or what the "right" spelling of a word is. The guidelines, as it stands now, encourages the use of the unit of measurement that's common within its field. In our computing articles, we tend towards KB, MB, etc.; this is very similar to how articles on American topics are going to present measurements gasoline in gallons and distance in miles. In terms of serving our most likely readers in any given topic, this works out very well, because it has the highest chance of understanding. -/- Warren 04:48, 23 March 2008 (UTC)
While that may be correct, in principle, I don't think it applies here. Advertising and other informal communications often use slang and loose, ambiguous terminology. On the other hand, we are a formal reference work, and owe it to our readers to be technically correct. We use words according to their dictionary definition, not according to their informal usage. In the case of technical terminology, the major standards organizations write the dictionaries, and they have stated unequivocally that "mega" means 106 and only 106, and that the only technically correct way of representing 220 with a prefix is to use the prefix "mebi". As with many words, informal and technically incorrect usage of words is common, but it is not appropriate for a formal reference work, even if common. "Representing the world as it is" would simply mean stating in the binary prefix article that in common usage, the decimal terminologies often are used to represent binary values. But we should not follow this example, because of the simple fact that it is not correct, based on the definitions of the terms set by those authorities tasked with defining them. We should not knowingly present incorrect or ambiguous information, even if the practice is a common one. Seraphimblade Talk to me 08:56, 23 March 2008 (UTC)
The "technically correct" argument is not the Wikipedia way of writing articles because "technically correct" is actually decided by real world consensus and Wikipedia does not always follow what a "standards organisation" says. Wikipedia wisely make the choice that real world consensus trumps standards bodies, see my talk page for an excellent example. Also your statement is not accurate because the JEDEC, which is a standards organisation, does state that mega and kilo use powers of two. So the only sensible option is for Wikipedia to use what is commonly used in the real world because that is the most widely understood, this means for the majoirty of articles that IEC prefixes should not be used. This is a natural function of real world use and wanting to be widely understood by the majority of people and has nothing to do with being for or against any particular "standard body" per se. Fnagaton 11:11, 23 March 2008 (UTC)
I repeat myself: The question is not whether IEC prefixes are used frequently, but whether they are used frequently in publications that try to be unambiguous, because that is what we (or most of us) want to be. The IEC symbols are certainly the most popular ones in these cases. (I have seen BB once, for billion bytes, but nothing else. Subscripts have been suggested here: GB16.) The IEC names are not as frequent as the symbols in my experience and they sound ugly, that is why I suggested the self-explanatory term binary megabyte (in contrast to decimal megabyte where necessary) instead. The added i does not seriously obfuscate the unit, it just gives more exact information to the knowledged reader while others would (hopefully) read it as if the i was not there.
I have never concealed my preference for simple and straight guidelines even if they were specific to Wikipedia. Allowing a binary meaning for SI prefixes in certain contexts – i.e. semiconductor storage capacity and perhaps file sizes – was already a step of mine towards compromise. I suggested to specify defaults in MOSNUM, so not every page needed explanatory footnotes. Footnotes, by the way, are not a desirable thing in neither encyclopaedias nor hypertext, (hyper)links are more appropriate and sufficient.
I started off by criticising Fnagaton’s proposal, but soon found that to be less explicit than writing a counter proposal, because his one was too anti-IEC, although overall sounder than I expected based on his hostile, uncompromising attitude exposed in Talk several times (with like responses of course). I tried to make it more general (to also apply to tons, pounds etc.), too, which probably lead to it being perceived as not precise enough. — Christoph Päper 20:29, 23 March 2008 (UTC)
You repeat yourself yet you are still wrong for exactly the same reasons I have already given. My proposal is not anti-IEC. Not once does my proposal say "do not to use IEC prefixes". I'll give you a chance though, if you can prove where my proposal is specifically "anti-IEC" then you'll get an apology. If you cannot then you admit you are wrong and I expect you to retract what you just wrote. My proposal uses the most clear and precise way to disambiguate prefixes which is to use mathematics to make it completely unambiguous about the exact number of bytes being used in a particular situation. It also does it in such a way without advocating any particular prefix system, therefore removing any chance that personal bias of the editor is in the guideline. Also I demand you retract you untrue allegations about "hostile, uncompromising attitude" because I present a clear argument and I do not resort to the kind of personal attack you have just demonstrated to try to score points. Fnagaton 15:54, 24 March 2008 (UTC)
You are right, your proposal does not say “do not use IEC prefixes”, because it does not mention them at all! That is pretty anti in my book. Instead you write “Editors must use the units/prefixes employed in the sources used by the article.” As I have shown that would only be a good recommendation, if those sources were disambiguating (inline), which most do not.
I repeated myself, because neither you nor anyone else has countered that point. The section #General IT prefix discussion ended with my notion that “real world consensus is that IEC prefixes are one way to achieve disambiguity where needed”.
Your constant calling people wrong and demanding retreats constitute the hostility, your unwillingness to accept IEC prefixes at all – direct quotes you cannot touch – constitutes the uncompromising attitude. It does not matter for the reception whether you perceive or mean it like that. — Christoph Päper 18:32, 24 March 2008 (UTC)
Not mentioning something is not "anti" that something, it is also not "for" that something. It is neutral, it is unbiased. So you admit you are wrong. When are you going to retract what you wrote? You are also wrong in your last paragraph, when people are wrong it is more polite to tell them as opposed to the kind of personal attacks you resort to. Also when someone writes something that is not true, like you have been doing, then they should retract it and there is no harm in pointing that out. You are making the mistake of confusing "uncompromising position" with "not accepting rubbish". It is easy to demonstrate you are wrong here because the fact is I made a proposal, then Greg comes along and makes another proposal. What do I do? I support his proposal. If I had an "uncompromising position" as you claim then I would not have supported it. You are therefore claiming and writing rubbish. Q.E.D. By repeating such baseless accusations you are also being uncivil, disruptive and making personal attacks. What you wrote has been countered or has nothing relevant to start with, for example when you wrote the "least space-taking solution..." post that uses entirely irrelevant POV which ignores what was already said before. It is easily demonstrated that the -bi prefixes are not the "most established solution" as you claimed. You are also using misrepresentation because what you wrote about "your unwillingness to accept IEC prefixes at all" is completely untrue, so I also demand that you retract that. I have no problem with IEC prefixes being used if that is the consensus as shown the the sources used in an article, the fact is for the majority of articles that is not thee case. Stop with the personal attacks right now, this is not the first time I've had to point this out to you. If you cannot be civil and debate the actual topic then I suggest you take a break and do not post here on this topic. As the flow of that discussion shows your "solution" is not a viable solution, for the reasons given at the time and by the fact that you made a comment 18 days (!!!) after the last post in the thread had been made and you did not even think to put a note on the involved editors talk pages, so there was little point going over what you kept on repeating. You will not get your own way on every single post that you make. Just so you are clear, telling someone they are wrong is not a personal attack. If you react badly to being told you are wrong then you have two viable choices. 1) Don't be wrong. 2) Don't post on Wikipedia. The choice you make, using personal attacks, is not a viable option. Fnagaton 19:03, 24 March 2008 (UTC)

There is a theme throughout this whole discussion that everyone knows that:

KB can mean 210bytes = 1,024 bytes, and that
MB can mean 220bytes = 1,048,576 bytes and that
GB can mean 230bytes = 1,073,741,824 bytes.

I doubt that this is true for the vast majority of the users of Wikipedia.!

  • The first point I would make is that I am an experienced programmer/electrical engineer and could not write the decimal equivalents of MiB or GiB without a calculator - to answer -/- Warren's 04:48, 23 March 2008 question above, when u tell me you have 3GB, I, absent a calculator, can only say you have about three billion bytes. IMO, most people don't have a clue how many bytes are in your PC, just a large number!
  • The next several points I will make, comes from an informal survey I did this weekend at a friend's Easter Brunch. There were 20 of us, ranging from a High School Junior to several retired teachers. All were computer users, about equally split between under 35 years old and over 55 years old. There was one other programmer and one MD. All were college graduates (the HS junior will do so). Here are the results of my informal poll leaving out the programmer:
  • About 5 of the attendees thought kilo = 1,000, even they were uncertain.
  • No one thought a kilobyte or a KB was 1,024 bytes. The five thought it meant one thousand bytes.
  • Exactly 2 (the MD and a high school principal) thought mega meant 1,000,000, neither was certain!
  • No one thought a megabyte or an MB was 1,048,576 bytes. the two thought it meant one million bytes.
  • My friends use a MAC, none of its three users could tell me how many bytes of memory nor the HDD size.
  • The programmer when asked in front of several members of the group, said mega meant million. When prompted, he said, of course, it could mean quote 220 unquote; however, he like me, could not state the precise decimal amount.
  • The final point I will make is that in settling the litigation Western Digital said in its motion supporting the settlement, "As noted above, Plaintiff admits that Western Digital is using the historically correct, industry standard definition of gigabyte"[1] In Wiki words, we have here a reliable source as to the common, historically correct definition of GB and by implication MB! While lawyer's are advocates, IMO, it would be unusual for there to be any gross misstatement in a defendant's motion supporting such a plaintiffs motion, so I find this statement compelling in resolving the many arguments posted hereinabove.

Standardization is a good thing and now that we have standards that eliminate ambiguity it seems to me we foolishly give up the opportunity to educate the Wiki public as to the issue when we deprecate (or ignore) them. To argue that we shouldn't use them because they are relatively unknown is a rather peculiar argument, how else can they become known? Following that argument to its end, we shouldn't use MB or GB either because most people don't know precisely what they mean. We the practitioners all understand the difference between MB and MiB, why should we withhold this? As David G says at the beginning of this section, why shouldn't we go with the best possible notation? Tom94022 (talk) 20:14, 24 March 2008 (UTC)

Except the JEDEC standards organisation and the memory chip industry have standardised KB/MB/GB using powers of two. I can also tell you what KB/MB/GB are in decimal, hex, binary and octal. No problem at all without using a calculator. Also without using a calculator I can perform arithmatic for those memory sizes with other typical power of two numbers that are very common in computer programming. I'm also a very experienced software engineer, I can for example tell you that 0xc000 is 49152 off the top of my head or for example 0x10000 / 0x400 = 0x40 = 64. To answer your first question: By the real world (not us as editors) adopting them over time if that is the will of the people producing the sources we must use without bias when writing articles. To answer your second question: Following on from your point we shouldn't be using the -bi prefixes because introducing extra terms will only go to confuse the people you polled even further. To answer your third question, because the real world doesn't use the notation very much, therefore it is not the best notation. It is POV to state what is the best notation. It is not POV to say "the real world uses this so we follow that example". Fnagaton 20:28, 24 March 2008 (UTC)
What is JEDEC? (That question is rhetorical, I know what it is.) It's no IEEE, no ANSI, no ISO. It certainly does not overrule a position that all of those major standards bodies have agreed to. Nor does advertising. We would not say that something is "smoking hot", even if advertising or other informal communication commonly and routinely did. To bring up JEDEC in the face of consensus among the major world standards organizations is ridiculous. JEDEC is just an industry consortium at best, the standards organizations have decades to centuries of history and experience. Also, JEDEC itself states that the usage is deprecated: "The definitions of kilo, giga, and mega based on powers of two are included only to reflect common usage. IEEE/ASTM SI 10-1997 states "This practice frequently leads to confusion and is deprecated." (emphasis added) JEDEC is not advocating improper use of these terms, only acknowledging that it happens. They are, with this statement, implicitly accepting IEEE's standard and agreeing that the terminology in common use is ambiguous and undesirable. I believe, as above, that you are also confusing us with the average person. Upon asking several people today how many bytes were in a kilobyte of RAM, they replied "Well, a thousand, of course." A reasonably intelligent person who does not have a strong computing background would naturally assume that, because kilo means one thousand. On the other hand, asked how many bytes were in a kibibyte, they replied "Well...I don't really know." That's GOOD! Instead of automatically assuming they had the right answer when they actually had the wrong one, using the correct term made it clear to them that they don't know. If they needed to know, that would trigger them to look it up. Better to use a term someone is unfamiliar with (and thus will look up) than a term with which they are familiar with but that in this specific case they are likely to interpret incorrectly. Seraphimblade Talk to me 11:38, 26 March 2008 (UTC)
Your post still does not mean you can ignore the fact that the JEDEC defines those prefixes as powers of two. Also the JEDEC is a standards organsiation and trying to misrepresent it as "just an industry consortium at best" is your POV. Quoting the JEDEC main page it says "JEDEC is the leading developer of standards for the solid-state industry." which refutes your claim. Also the JEDEC does not state the standard is deprecated as you claimed. In actual fact the word deprecated is part of a quote from the IEEE in the note attached to the standard and is not part of the main standard text, this is because the notes are there for context and not actually part of the main standard definition. The main standard and the notes to the standard are two separate things and you are applying undue weight to the notes and not applying that weight to the main standard definition. The fact is the JEDEC still expect memory to use KB/MB/GB in the binary sense. Just so you are clear, the main JEDEC standard definition refutes what you wrote above. Fnagaton 12:02, 26 March 2008 (UTC)
The blog from the British Computer Society's web pages gives a computer professionals' perspective on the IEC standard [1]. Pyrotec (talk) 19:53, 26 March 2008 (UTC)
Fnagoton, yes, JEDEC's definition disagrees. So what? JEDEC is, compared to the venerable ISO, IEEE, ANSI, etc..., an upstart. You wouldn't care if I personally said "Binary prefixes are the right way to do this", and for good reason—I lack such authority. JEDEC might be able to claim some authority, but certainly not enough to overrule such a clear consensus among the major standards makers. We should footnote in the binary prefix article that JEDEC disagrees, but that's all. Your suggestion is much like those who argue that we should not use the mainstream scientific position as the major opinion because there are a few dissenters or the general public tends to misunderstand. ANSI, and IEEE, and ISO, are the largest, top standards bodies in the world. They have stated clearly that decimal prefixes are to be used for decimal numbers and binary prefixes to be used for binary numbers. Whether or not most people do that yet, the authorities on the matter have clearly and unambiguously spoken. We don't question that, we follow it. We use terms as they're defined, and the top authorities responsible for defining them have clearly and unambiguously stated what they mean. It's not our place to second-guess that or argue over it, it doesn't matter what we think. Seraphimblade Talk to me 21:21, 26 March 2008 (UTC)
Wikipedia doesn't always follow standards organisations even if you think they are "the top in the world", it follows real world consensus. See my talk page for an excellent example. Real world consensus in this case is to not use IEC prefixes in the majority of situations, you cannot refute that. So Wikipedia should not do so either, unless it is in a very small minority of articles which is what it says in the proposal. Just so you are clear, what you think is "a clear consensus among the major standards makers" has been overruled by the real world and their standard has failed to get widely adopted. The top authority is real world consensus, that is how Wikipedia operates. These facts neatly refute what you have been posting. Fnagaton 21:26, 26 March 2008 (UTC)
Of course I can refute that. (Really.) Last time I checked IEEE and ISO and ANSI are very real organizations, and really the ultimate authority on what technical terms mean. They quite really have defined decimal prefixes as decimal and binary prefixes as binary. So kilo really means one thousand, and kibi means 1024, and that really gets misused a lot. Most technical terms really get misused a lot. That really doesn't mean in real articles we shouldn't use the real meanings. In the real world, the authorities on things, not popular usage, define terms, in terms of their formal definitions for use in formal publications. And that's based entirely on real things from the real world. Really. Seraphimblade Talk to me 21:41, 26 March 2008 (UTC)
No they are not "the ultimate authority" they don't even enforce what you claim of course so since real world consensus trumps what they think then sometimes when they propose something which isn't followed then Wikipedia does not follow the organsiation either. Wikipedia follows real world consensus, this has been demonstrated before. I'm glad you posted WP:NOR since NOR refutes what you posted above because it says "Wikipedia does not publish original research (OR) or original thought". This means Wikipedia articles have to follow WP:Verifiability which says "The threshold for inclusion in Wikipedia is verifiability, not truth." This means an article that has sources which use the terms KB/MB/GB should not use the IEC prefixes since the article sources do not verify that those IEC prefixes are used. Just so you are clear, it is against WP:NOR to write force using IEC prefixes in articles when there is virtually no real world consensus to use those prefixes. These facts neatly refute what you have been posting. If you really think the standards bodies trump real world consensus then prove it, since you have not proven it so far. Fnagaton 21:26, 26 March 2008 (UTC)

When it comes to measurements, the government is the authority. If you offer goods for sale, and label them with the wrong units (as determined by the goverment) the weights and measures inspectors might come into your place of business, seize the offending goods, and write a citation requiring you to account for your misdeeds in court. If you engage in a transaction with another private party that involves measurements, and the other party thinks he has been cheated through improper measurements, he can sue you, and the court will decide who was right about the measurements. The voluntary standards from the IEEE, JEDEC, ANSI, and other non-profit private organizations have no authority unless they are adopted in a law or legally binding regulation, or they become so widely used that they become part of the language, just like all the other words that are used in contracts. --Gerry Ashton (talk) 21:56, 26 March 2008 (UTC)

"The voluntary standards" says it all and since they are not widely used that then means Wikipedia shouldn't use them either. Fnagaton 21:58, 26 March 2008 (UTC)
If we then want a government source, that is easy enough as well, given that the NIST has adopted the use as well, see [2]. So, at least in the United States, the official government body responsible for setting standards has also spoken. (This may be true in other countries as well, I live in the US so am most familiar with its government. Anyone with insight into other governments, please speak up.) Seraphimblade Talk to me 22:46, 26 March 2008 (UTC)
The NIST page cited by Seraphimblade describes what the IEC does. It does not create any requirement to use the IEC prefixes. It does not even say that NIST has adopted the prefixes for use in NIST's own publications (I don't know what, if any, prefixes NIST uses when describing the size of RAM.) --Gerry Ashton (talk) 22:53, 26 March 2008 (UTC)
Seraphimblade for the reasons given by Gerry posting that link is not proving your point of view that standards bodies trump real world consensus. Please try again. Fnagaton 22:55, 26 March 2008 (UTC)
Fnagoton, please kindly do not misrepresent me. My position is not that standards bodies trump "real world consensus", my position is that they are real world consensus. They are the bodies tasked with creating and defining technical terms. A consensus among standards bodies is "real world consensus", it is not separate or distinct or something different from it. It does not need to "trump" it, it is it! Widespread misuse of technical terms is common with almost all such terms. That does not mean that the terms do not actually mean what they were defined as, it simply means misuse is widespread. Seraphimblade Talk to me 23:57, 26 March 2008 (UTC)
I'm not misrepresenting you, actually you are misrepresenting what I wrote. This is because in this section at "11:11, 23 March 2008 (UTC)" I first used the term "real world consensus" and clearly stated "real world consensus trumps standards bodies" and "does not always follow what a standards organisation says". From this you may deduce that I mean "real world consensus" is not the same as a "standards organisation". Your later posts then attempt to contradict that by trying to claim the standard bodies are better (i.e. trump) what I said about real world consensus. So what I wrote is using the terminology I set down for this debate and what you wrote is self contradictory. So instead of "standards bodies trump real world consensus" you really mean "standards bodies are real world consensus", if you want to try to redefine real world consensus then so be it, but your POV is still not proven by your statements. And you've not spelt my name correctly either. Fnagaton 00:11, 27 March 2008 (UTC)
Apologize for the misspelling. I do indeed believe that the standards bodies set the real world consensus for what technical terms mean, as they are a definitive source tasked with making such definitions, just as other definitive sources such as a dictionary set the real-world consensus for what words actually mean, even if they're commonly used in a slang or other manner. I'm not stating that the standards bodies can override some type of "real-world consensus", I'm stating that they set it, and that trying to speak of "real world consensus" as something apart from the official definition of the terms by the authoritative bodies which do so is nonsensical. In the real world, standards bodies set the official definitions of technical terms. That is the real-world consensus, however widespread the misuse may be. Even you know what "kibibyte" means, so there's no lack of consensus as to its meaning. As to whether "kilobyte" means the same thing, since there's an ambiguity, we look to the authorities on the subject. What do they say? "No, it doesn't, kilobyte means one thousand bytes." That's not original research, as you accused earlier, it's quite well sourceable. It is not original research to state that 1 inch = 2.54 cm, that's perfectly sourceable and an entirely valid conversion. Similarly, it's not original research to claim that 1 KiB = 1.024 KB = 1024 B. That's simply a measurement conversion, and a perfectly source-based one, no original research is involved in that. Measurement conversions have never been considered original research, even if the original source used a different unit. Seraphimblade Talk to me 00:29, 27 March 2008 (UTC)

(Outdent) The entire computer industry follows the IEEE/ANSI standards for computer storage units. That is the standards adopted in the 1980s and 1990s that codified kilo, mega and giga as either decimal or binary depending on context (ANSI/IEEE Std 1084-1986). They have totally ignored the new IEC binary units.

If you look at the technical specifications on the Intel website for microprocessors you will find MB for megabyte. If you look at the technical specifications on the Samsung website for DRAM you will find Mb for megabit. When company like Dell signs a contract to purchase millions of dollars of microprocessors and memory, the specifications have the standardized binary units the industry has used for decades. I doubt that the contracts are written in Esperanto. -- SWTPC6800 (talk) 01:26, 27 March 2008 (UTC)

And I would also add that "the standards bodies set the real world consensus for what technical terms mean" is not true and would be better phrased without the word consensus - consensus would be a pretty subjective term in this context. Standards bodies set standards and guidelines in order to try and establish common usage and gain consensus. They do not start with consensus (unless you're referring to consensus within the group of people that form that particular standards body). In fact, its usually a lack of consensus on a particular topic that's driving them to create and promote their standard in the first place. --Marty Goldberg (talk) 01:45, 27 March 2008 (UTC)
Seraphimblade wrote "I do indeed believe that the standards bodies set the real world consensus for what technical terms mean, as they are a definitive source tasked with making such definitions, just as other definitive sources such as a dictionary set the real-world consensus for what words actually mean, even if they're commonly used in a slang or other manner." I disagree. Dictionaries generally document the meaning well-edited sources ascribe to words; they generally don't tell their readers how the language ought to be improved, or what words should be dropped because they're confusing. Standards bodies may very well pass standards that advocate change in the meaning of words, or they may invent brand new words; dictionaries don't do that.
Also, nobody is in charge of voluntary standards bodies. Anybody can start one, and it's really just a popularity contest that determines which ones are influential, which ones are marginal, and which ones go out of business. (In that sense, they are like dictionaries.)
Oh, and just to disclose my interests, I'm a member of IEEE and have contributed one sentence to an IEEE standard. --Gerry Ashton (talk) 02:54, 27 March 2008 (UTC)
Anyone can start a standards organization. The Secure Digital card memory for your camera is standardized by the SD Card Association. [3] The Universal Serial Bus ports on your computer are standardized by the USB Implementers Forum. [4] Neither is controlled by a government agency. -- SWTPC6800 (talk) 04:02, 27 March 2008 (UTC)
Seraphimblade you wrote "I do indeed believe that the standards bodies set the real world consensus for what technical terms mean" and "In the real world, standards bodies set the official definitions of technical terms. That is the real-world consensus". The two things are not quite the same. I will explain. The first quote means they set what the technical terms mean, it doesn't mean the world has to follow those standards. Sometimes the real world consensus, through common language use, is that the standards bodies are advocating something that is not popular. For example with IEC prefixes you cannot refute these prefixes are not popular despite the standards bodies advoating them for the last ten years. Real world consensus therefore is against what some of the standards bodies advocate. So while you are partially correct, they do set the terms, you are incorrect to try to claim they reflect real world consensus. In your second quote you again try to equate setting a definition with real world consensus, where no such link exists in the case of IEC prefixes. Lastly, when you wrote "It is not original research..." you're the forgetting "The threshold for inclusion in Wikipedia is verifiability, not truth." part of WP:NOR. Anyway I do have to agree with Gerry's point about dictionaries, they do not create new words at all, they do reflect common use in language. So Seraphimblade if after reading Gerry's comment you still believe "standards bodies set the real world consensus" then I want you to demonstrate exactly why. You can demonstrate you are correct by removing the feet and yards from the American Football article and getting your change to stay for at least a month, this is because according to the standards bodies you advocate the terms yards and feet are not to be used. If, as you claim, the standards bodies are real world consensus then you should face little problem is making this change to the article and seeing it stay that way. If you do not make the change then you're not going to demonstrate why your point of view is correct and I'm afraid you'll have to concede the point you've been making. Fnagaton 09:25, 27 March 2008 (UTC)

(outdent) I support and agree with the positions and arguments expressed by David Göthberg and by Tom94022.

For me this issue is very simple: There are some things in the computer industry for which decimal prefixes are used when the meaning is binary, and there are some for which decimal prefixes are used when decimal is intended, and this is confusing and ambiguous. The notion expressed above by e.g. Fnagston that "everyone knows what is meant" is false: The ambiguous usage of e.g. "GB" for both meanings leads some to believe that "everything in a computer is in powers of two". For example, I've had numerous people try to tell me that "10 megabit Ethernet" means 10 times 2 to the 20th bits/second, or that a "100 gigabyte tape" must hold 100 x 1024 x 1024 x 1024 bytes. (No, I don't have the decimal equivalent of 0x40000000 or even 0x100000 memorized either, even though I've been programming for decades, much of the time in various assembly languages.)

Heck, I even think that wherever confusion is possible (e.g. just about everyplace in computer contexts except RAM sizes), I think the standard "metric prefixes" should be replaced with something like "gidebytes" (GdB) where the original number was divided by a power of 10. Of course it isn't WP's place to push or even suggest such a thing.

In the meantime, though, the binary prefixes already exist, having been proposed and adopted by major, widely recognized standards organizations and professional societies. Using them wherever appropriate is less confusing in the long run, even if they are unfamiliar upon first encounter to most readers. QED (for me), using binary prefixes is what WP should do, even if it does mean that WP is taking a leading rather than a documenting role with respect to real world consensus. For me, that concern is less important than accuracy and precision. Jeh (talk) 16:13, 27 March 2008 (UTC)

Not when they are virtually unused and unknown and Wikipedia doesn't ignore real world consensus for the sake of accuracy or precision. It is more accurate to not use IEC prefixes in most articles and instead use some other form of disambiguation like stating the exact number of bytes. Fnagaton 16:18, 27 March 2008 (UTC)
One, regarding "virtually unusued," you could have said the same thing about metric measurements in the U.S. in the 1960s outside of scientific fields. Yet the right thing for science textbooks and encyclopedias, even those intended for the general public, at that time was to use metric measurements. Two, regarding "more accurate", that is simply wrong. You can argue a lot of points but you cannot say that it is "more accurate to not use IEC prefixes" where they are appropriate -- for example, 4 GiB RAM is exactly 4 times 2 the 30th bytes, is it not? Third, a usage such as "4 GiB" is no doubt seen by you as more obtrusive than "4 GB", but having to write "4 GB (4,294,967,296 bytes)" is far more obtrusive and awkward than either. Your suggestion that one should in such cases disambiguate by stating the exact number of bytes is also an admission that "4 GB" by itself was inaccurate, or at best ambiguous! Jeh (talk) 16:30, 27 March 2008 (UTC)
If Wikipedia was around in the 1960s then I'd expect it to reflect the real world consensus as my Physics and Mathematics textbooks do from that period of time and they use feet, inches and pounds. I can say it is more accurate not to use IEC prefixes because as I have shown before IEC prefixes can be confusing and mistaken that they are using decimal and not binary. Therefore to state the exact number of bytes is completely unambiguous. Stating the number of bytes need not mean writing out all of the numbers because it is common to use power notation instead. This is shown in the binary prefix conversion table. Fnagaton 16:51, 27 March 2008 (UTC)
Fnagaton, I wonder why you insist on using the term "real world" as if it was something in strong support of your position. As other people already wrote and you know: "The threshold for inclusion in Wikipedia is verifiability, not truth." The word "reality" is synonymous to the word "truth". Therefore "real world" is no valid argument. So it doesn't even matter that some people apparently have a different perception of reality than others. -- (talk) 17:11, 27 March 2008 (UTC)
  • Jeh, terms like ppb and ppt are ambiguous so IUPAP proposed adoption of the “uno”. Thus, ppm would be 1 µU and ppt would be 1 pU. Notwithstanding that the uno was absolutely unambiguous and was a great idea, it didn’t catch on. What point would there be to using the unit of measure here on Wikipedia if the reader is expected to remember a unit of measure that will likely only ever be encountered here on Wikipedia? Manufacturers of memory don’t advertise “2 gibibyte SODIMM card” and general-circulation magazines don’t use the IEC prefixes either. No matter how meritorious and wonderful a new unit of measure is, encyclopedias and dictionaries never add them to their lexicon until they achieve a certain level of ubiquity in real-world use. There are sound reasons for this: to communicate with the intended readership with minimal confusion. Wikipedia is not immune from this common-sense rule. Greg L (my talk) 16:41, 27 March 2008 (UTC)
The cases are not at all parallel. We're not talking about a "new unit of measure" here -- the "unit" in question is still the byte.
The "uno" is not only a new unit, its expression in text looks completely different and so presents a visual stumbling block, or "speed bump." IMO, the beauty of the binary prefixes is that they look close enough, and numerically are close enough, to the decimal equivalents that the casual reader will just read them as if they were the decimal versions, and go on, and will get the close-enough meaning; no one has to "remember a new unit of measure" in this case.
In fact, if Fnagston is correct and "everybody knows" what's meant, they'll get the exact meaning! Of course, I don't think an encyclopedia should depend on "everyone knowing what is meant," and I believe that any arguments from that position are completely specious.
Otoh, if binary prefixes are used, then someone who wants technical precision will recognize that the use of the binary prefix indicates that, yes, the numbers preceding those prefixes were thought about and the right numbers are present (e.g. 4 GiB rather than 4 GB when the true number is 4.3 GB).
Regarding "minimal confusion": The computer industry's persistent use of decimal prefixes when the actual meaning is binary (and that is exactly what you are supporting, Greg, is it not?) is rather far from "communicating with minimal confusion". Indeed this ambiguity actively increases confusion, as demonstrated by the recent lawsuit against WD (the actual offender there was the operating systems that insist on reporting e.g. an 80,000,000,000 byte drive as "74.5 GB").
WP should not promulgate a confusing convention, no matter how widespread it is. Particularly not when a better alternative is considererd correct usage by recognized organizations in the field. Jeh (talk) 17:07, 27 March 2008 (UTC)
Jeh, the "computer industry" is not supporting either potential meaning of "mega" at all. There is no consensus about it in the IT industry. Even those who are members of JEDEC are only a fraction of the whole IT industry. It is therefore reasonable to follow a standard that is less well-known but accurate and in compliance with the rest of all other industries and sciences, as represented by IEC 60027-2, than following a so-called "de-facto standard" that isn't unanimously accepted, inaccurate and misleading in its own area. -- (talk) 17:25, 27 March 2008 (UTC)
In the IEEE 100 standard kilobyte etc are defined as binary and decimal. Fnagaton 17:59, 27 March 2008 (UTC) I agree completely. (I also added an extra indent colon to your graf, as is the convention here; hope you don't mind.) Jeh (talk) 17:47, 27 March 2008 (UTC)
  • To If you aren’t going to register, why don’t you at least use an alias or “handle” (like “Gunther” or something); your arguments will carry greater weight that way (v.s. when you weigh in entirely anonymously).

    Let me ask you this: If you wrote an article on how the change in frequency on a crystal oscillator was 200 ppb/°C, would you approve if I went in and changed it to 200 pU/°C? Wikipedia could have a whole article dedicated exclusively to the uno. That’s because terms like “ppb” are ambiguous because “billion” denotes a different value in different countries. Uno, which is entirely unambiguous, can therefore be used to disambiguate terms like ppb and ppt (which usually means parts per trillion but occasionally means parts per thousand). We can start routinely using the uno and its unit symbol to denote all high-ratio (low value) proportional quantities in an “oh… didn’t-cha know?” fashion and link each use to an article on the uno. Thus, we will be in a position of helping to promote the adoption of a unit of measure that simply didn’t catch on.

    We shouldn’t now, should we? The uno is a good idea that failed to catch on and isn’t widely recognized. If it weren’t for the fact that I myself added the “uno” section to the Parts-per notation article, there would be even fewer people who know about it. Routinely using the uno in articles would mislead some impressionable readers who would run about with that ‘space-cadet glow’, using a term that only confuses others because it is so poorly recognized. The same goes for “kibibits” and the other IEC units of measure. They didn’t catch on. Our continued use of them here is a mistake. It’s been eight years and the IEC proposal still hasn’t caught on. Encyclopedias and dictionaries simply do not add new terminology to their lexicon until it has reached a certain level of ubiquity. Wikipedia is not immune to this common sense rule. Doing otherwise subverts the primary objective in technical writing: to communicate to the intended reader with minimal confusion.

    The only reason it’s been such an uphill battle is some of the “oppose” authors have been really, really prolific and there is a lot of hard work that must be undone. Either that, or they let Fnagaton do it for them. Both options are unpalatable to them so they’re fighting like hell to preserve the continued use of units that the typical reader doesn’t recognize and is unlikely to encounter anywhere else. Greg L (my talk) 18:01, 27 March 2008 (UTC)

Greg L, your mention of "space-cadet glow" is incredibly ironic. There was no need for kibibits in the first place but because we cared so much for the needs of our space-cadets, we added a few words to the vocabulary, so that they could stick to their habits and we get on with our work. If you prefer to count your shiploads of memory in 1024 large pieces, so be it, but please pick a proper name. I almost get the impression that people don't care so much about confusing the average reader as annoying our beloved space-cadets. Regarding "Uno", Jeh already said all I could have said about it. -- (talk) 21:06, 27 March 2008 (UTC)
Anonymous editor from Germany: What are you talking about? Words like kibibits are real but uknown to the typical reader. Since Jeh and you’d don’t agree with this common-sense fact (or you do know but refuse to admit it), I’m not going to argue the point further with you. Greg L (my talk) 21:28, 27 March 2008 (UTC)
Pseudonymous editor from the USA: I wasn't denying that kibibit is a real word. I didn't even claim it's widely known or the opposite actually. It's probably not as unknown as you'd like to believe. A lot people writing here and elsewhere know the kibit but don't use it because they despise it. I'm surprised you know the typical reader. He's probably from the real world, right? I wonder does this typical reader only know the word "megabyte" or does he (or she?) also know what it means in any given context? What I'd also like to know, is the typical user a space-cadet or not? Once, I thought the typical reader reads Wikipedia articles to learn something but maybe the emphasis should be on "something" rather than "learn". -- (talk) 22:12, 27 March 2008 (UTC)

To GregL:

1. Your assertion that the "typical reader doesn't recognize" the IEC prefixes is, one, unproven, and two, uncompelling to me even if true. My (yes, unproven) counter-assertion is that if someone is familiar only with "GB", "GiB" is very UNlikely to be very confusing: it looks similar enough to the imprecise "GB" that at worst they will likely think "oh, it means GB" and get approximately correct information... which is what they're getting from the decimal prefix anyway! Whereas, if correctly interpreted, the IEC prefixes provide precise meanings, something that cannot be said of the decimal prefixes when used with e.g. RAM sizes. Even to those unfamliar with the IEC prefixes, they will therefore provide no worse confusion than continued use of the decimal prefixes where binary would be correct.

2. The uno is a completely different and nonparallel issue. Reason: Unlike "GB" vs. "GiB", there is nothing in an uppercase U that suggests it means 1 part in 10 to the 9th, i.e. no obvious relationship to the "ppb" it tries to replace. This is NOT the case with GiB vs. GB. That is, in my opinion and as I have said, part of the genius of the design of the binary prefixes.

3. What you think encyclopedias and dictionaries "simply do not" do is your perspective, and not necessarily applicable here. Most reference books do not allow anonymous members of the general publicto make changes that are then instantly distributed to their readers; the rules here are very, very different.

4. The bottom line for me (and a point you have not at all addressed, and which I said before, but since you feel free to repeat points I have addressed, I see no reason not to repeat points you have not addressed) is this: The widespread usage of e.g. "GB" to mean 1,073,741,824 bytes is KNOWN to be confusing. Witness (again) the recent lawsuit against WD, in which even the court came to the wrong answer. An encyclopedia should strive to reduce confusion, not perpetuate it, let alone require it via its manual of style.

5. I think "the only reason it's a battle at all" is just about the opposite: you, Fnagston, and a very few others here are trying to force a change in the status quo. My view is that the guideline could stand exactly as it is; but I think it should be improved by actively encouraging the use of IEC prefixes, for reasons I have described already. Jeh (talk) 20:21, 27 March 2008 (UTC)

  • Given that you don’t even agree that “the typical reader doesn't recognize the IEC prefixes”, there is no further point debating with you. You’ve voted and I accept that. Greg L (my talk) 21:06, 27 March 2008 (UTC)

I agree that the IEC prefixes are probably (but not proven) initially unfamiliar to most WP readers. The reason I find this uncompelling -- even if proven -- is this: I strongly doubt that for most readers this unfamiliarity survives about half a second into the first encounter. The approximate meaning (that is, close to the same thing as the corresponding SI prefixes and units, with which they share two out of three letters) is almost always completely apparent from context. That is the beauty of their design, a point that the "Uno" does not share. Jeh (talk) 21:59, 27 March 2008 (UTC)

I wonder how many people noticed that The Pirate Bay, many other BitTorrent index sites and related software have been using the new prefixes from IEC 60027-2 for quite some time. It has to be millions of people, maybe hundreds of millions over the years. -- (talk) 02:52, 28 March 2008 (UTC)
The gender neutral pronoun s/he or (s)he also shares 2 letters. This monstrosity was more widely used than MiB before it was flushed down the toilet. -- SWTPC6800 (talk) 00:50, 28 March 2008 (UTC)
Your point being...what? -- (talk) 01:30, 28 March 2008 (UTC)

I would also like to reply to Fnagaton, above, regarding verifiability. It is trivial to verify the definitions of the binary prefixes. "Verifiable" simply means that one can verify the information through a reliable source, and that a citation can be provided allowing for anyone with reasonable access to reference material to do so. One need only cite the standards bodies' definitions to fulfill this requirement. If the meaning of the binary prefixes were unverifiable, I'd be the first to stand against their usage, but this is simply not the case. Even those who are against their usage are clearly aware what they mean, and we can certainly provide verification of that fact through reliable sources. Such verification is indeed provided in the binary prefix article. Seraphimblade Talk to me 09:52, 29 March 2008 (UTC)

No, adding new sources that most readers don't understand is not promoting verifiability because this involves citing the sources used in for the article. What you are advocating is pushing virtually unknown prefixes into an article which then causes more confusion to the reader. So adding a cite to the relevant standards body is not all that is needed to saitisfy verifiability. Fnagaton 16:15, 29 March 2008 (UTC)
I think you need to go read the verifiability policy again. I daresay most readers would not understand the main source I used in writing salt tectonics, but that does not render the information unverifiable. "Verifiable" simply means "one can go verify it." In this case, the conditions (the average person would have reasonable access to the sources cited, and the source is reliable) is satisfied. You can argue that we still shouldn't use it, but verifiability is easily satisfied here. Seraphimblade Talk to me 08:23, 30 March 2008 (UTC)
No, you are wrong. The goal of disambiguation and verifiability is to make it easier to understand what something means in an article and that goal is not helped by using terms that the majority of readers don't recognise and don't understand. You need to read the verifiability policy again in the context of disambiguation because it is clear it means that you should make sure the article sources can verify what is in the article by using the sources relevant to the article. This does not mean adding extra sources (to IEC or SI sites for example or wikilinks to IEC prefixes) to push your POV that the (virtually unknown and little understood) IEC prefixes are somehow better used for disambiguation. The IEC prefixes are not the best choice for disambiguation and that is clear from the comments on this page. Fnagaton 09:51, 30 March 2008 (UTC)
Are you trying, then, to state that the average reader does understand the term "megabyte", including that in common usage (and currently, our usage) it is ambiguous and has no exact value? The conversation below certainly seems to indicate that the majority of our readers quite understandably presume that the metric prefixes "kilo", "mega", and "giga" mean exactly the same thing as they mean in kilometer, or megaton, or gigavolt; that is, thousand, million, and billion. Who can fault them for that? If I didn't already know differently, I'm sure I would make the same assumption. So in stating we're using terms that readers are not familiar with, well, they're not really familiar with the ones you wish to use! They may have heard them more frequently, but that leads to an even worse situation, in which our reader believes (s)he understands the term and really does not. Using a genuinely unfamiliar term will lead that reader to say "What's this mean? Better look." And then our reader learns something and gains understanding of the matter which (s)he did not have before, which, well, after all, is exactly why we do this in the first place! As to your arguments regarding verifiability, one may certainly verify that memory sizes are binary and hard drive sizes are decimal, etc. You seem to be claiming that the use of additional sources compromises verifiability, a claim which frankly is mind-boggling. I see nothing in verifiability that even hints that the use of multiple sources violates verifiability, and I would venture a guess that such a change would fail to gain anything approaching consensus. The use of multiple sources is allowed so long as it does not come to original research, and simple mathematical or deductive conversions have never been considered original research. Seraphimblade Talk to me 19:56, 30 March 2008 (UTC)
No, it's not a case of "the ones you wish to use" it is a case of me saying to be consistent with the sources relevant to the article. You on the other hand do want to push virtually unknown prefixes into articles where the sources do not use those terms. That is the crucial difference. I am being unbiased and saying defer to the choices made by creators of the sources we should be using. You on the other hand are trying to force your own POV into articles. It is against original research because you are using your own bias. Your point about "I see nothing in verifiability that even hints..." is fallacious because you're forgetting that the prefixes you are wanting to use are virtually unknown. The whole point about disambiguation and verifiability is to use terms that help the reader and the IEC prefixes do not help the reader enough to require their use because other better options exist. Fnagaton 20:29, 30 March 2008 (UTC)
  • I've just noticed that the templates for seaching, page edits etc are have switched back to using KB/MB etc. During the height of the push for IEC prefixes these were changed to use the IEC prefixes. It looks like consensus is now back the other way. Fnagaton 15:32, 30 March 2008 (UTC)
Neither the change to IEC prefixes nor the change back mean much of anything. Can you link to the templates and discussions about them? — Omegatron 16:08, 30 March 2008 (UTC)
Are you trying to say there was a change without meaning? Fnagaton 16:47, 30 March 2008 (UTC)

There is no common usage among readers

I see a lot of people saying "we should not use IEC prefixes because readers are unfamiliar with them" or "we should use the ambiguous units because readers are familiar with them". But I don't think there's any merit to this argument.

Wikipedia is a general-purpose encyclopedia, not a computer science textbook. The vast majority of our readers don't know that "KB" is supposed to mean 1,024 in the first place. "KB = 1024" is just as unfamiliar as "KiB = 1024". This is especially true in metric countries (which is pretty much all of them except the US). Ask friends who are not from technical backgrounds or look through Yahoo Answers:

  • Resolved Question - how many bytes are in a kilobyte? how many kilobytes are in a megabyte? how many megabytes are in a gigabyte?
    • Best Answer - Chosen by Voters: "1000 1,000,000 1,000,000,000"
  • "1,000,000 kb in 1 g"
  • "1000 Bytes in a Kilobyte, 1000 KB in a Megabyte, 1000 MB in a Gigabyte, 1000 GB in a Terabyte - So the answer is a million."
  • "1000 bytes to make a kilbyte, 1000 kb = 1 mb, 1000 mb = 1 gb"
  • "a byte is a size measurement of memory stored by a device. a kilobite (KB) is 1000 bites. a megabite (MB) is 1,000,000 bites (1000KB). A gigabite (GB) is 1,000,000,000 bites (1,000,000KB, 1,000MB) and a terabite (TB, probably) is one trillion 1,000,000,000,000 bites."
  • "a hundred thousand approximately... i guess..."
  • "10,000 kilobytes are in a gigabyte"
  • "kilobyte 1000 bytes / megabyte 100,00 byte / gigabyte 1,000,000 bytes"

Regular people have no clue.  :) There are certainly people on there who know the Microsoft definition, but there are many more who don't. I don't think we're confusing anyone any more by using IEC prefixes than we would using Microsoft prefixes. — Omegatron 16:04, 30 March 2008 (UTC)

The vast majority of encyclopedias use the terms that are commonly found in the real world. i.e. They don't use IEC prefixes because IEC prefixes are not common. Fnagaton 01:02, 31 March 2008 (UTC)
We are not talking about "common" terms, but about uncommon technical terms used in articles about technical matters. Nor do I believe it is true that the standard is "common" usage, whatever that means! Since IEC prefixes have been adopted by IEEE and are unambiguous their usage in appropriate technical articles should be encouraged - an encyclopedia is about teaching, isn't it? Tom94022 (talk) 17:34, 1 April 2008 (UTC)
An encylopedia is about teaching what is in the real world and not in the business of supporting failed standards. Fnagaton 18:03, 1 April 2008 (UTC)

The "IEEE" does not use the IEC binary prefixes. The IEEE Standards Association is just one part of the IEEE. The IEEE Publications have not adopted the IEC binary prefixes. The April 2008 issue of the flagship publication, IEEE Spectrum magazine has this article.

Tools & toys: Hacking the Nokia N800 "A lot can happen in a decade. You can hold the Nokia N800 in your hand, yet it’s a near-exact match for a high-end desktop PC from 10 years ago. It has a 320-megahertz processor, 128 megabytes of RAM, and a few gigabytes of available mass storage."

Wallich, Paul (April 2008). "Tools & toys: Hacking the Nokia N800". IEEE Spectrum (IEEE) 45 (4): 25. doi:10.1109/MSPEC.2008.4476441. 

I guess the 385,000 technology professionals in industry, government, and academia that read the IEEE Spectrum each month[5] can deal with the ambiguous "128 megabytes of RAM". -- SWTPC6800 (talk) 01:33, 2 April 2008 (UTC)

I dare say that IT professionals can work out from the context what kind of megabyte is meant. I can do so myself sometimes, but certainly not always. Either way, it is a red herring - the majority of WP readers don't even know that MB has more than one meaning, let alone how to differentiate between the different meanings. That's why there is a need to disambiguate. Thunderbird2 (talk) 19:14, 2 April 2008 (UTC)
But not to disambiguate by using IEC prefixes since that way doesn't have consensus. The consensus at the moment looks like disambiguation by stating the exact number of bytes is acceptable. Don't you agree? Fnagaton 19:45, 2 April 2008 (UTC)
It's not a red herring since it shows quite nicely that even the IEEE don't enforce their standard in their own publications, which just goes to show you can't always work against real world consensus for common use. ;) Fnagaton 19:46, 2 April 2008 (UTC)
Three points:
  • The red herring is that the experts can work it out for themselves. The experts can work it out for themselves, but the point is an irrelevant one here.
  • No evidence has been presented to support the statement that IEEE journals do not follow their own standard.
  • I have always agreed that an exact number of bytes is an acceptable disambiguation method, but Swtpc6800 implied with his statement that disambiguation was not necessary. It is.
Thunderbird2 (talk) 22:09, 2 April 2008 (UTC)
No, just because experts might be able to work it out does not preclude the fact that the IEEE journel cited above uses prefixes in a non-IEC way. What is a red herring is you mentioning "experts can figure it out" since that is an irrelevant point and does not alter the presented facts. The evidence is the cite from the IEEE journal, since if the IEEE do follow IEC prefixes then the journal entry cited above would not show those prefixes being used in a non-IEC way. Q.E.D. What you are doing is denying that the cite has been made when it is a fact that the cite has been made and then trying to waffle about "no evidence" when that is the evidence, which is silly. It's like trying to deny a man is standing outside your window by closing your curtains. The third point you made about disambiguation is also irrelevant to the point that the IEEE journal cited uses prefixes in a way that are non-IEC. Also your third point is a straw man logical fallacy since you misrepresented what the editor wrote and then presented it in your own words. Fnagaton 22:49, 2 April 2008 (UTC)
The members of the "Institute of Electrical and Electronics Engineers" are not just IT professionals. The typical issue of Spectrum covers power generation and transmission, electric automobiles, satellite communications, radio systems, electronic sensors and the occasional article on computers. The readers have more expertise with megawatts and megahertz than megabytes. Thunderbird2, could you provide some samples of wide circulation magazines that use the IEC binary prefixes. Technical journals that publish completed papers from authors tend to allow MB, Mbyte or MiB. I am thinking about magazines that have complete editorial control of the articles. -- SWTPC6800 (talk) 00:50, 3 April 2008 (UTC)
Google: "magazines using iec prefixes" First link User:Swtpc6800/Adoption second link User:Swtpc6800/Standards. Congrats ;) Fnagaton 01:02, 3 April 2008 (UTC)
I do not claim MiB is in widespread use, nor have I ever done so. It is used by those publications that strive to be unambiguous. A magazine is free to make that choice and most prefer the ambiguous MB. Thunderbird2 (talk) 10:43, 3 April 2008 (UTC)
Published by the IEEE. It speaks volumes about the lack of acceptance of the IEC prefixes when an IEEE publication doesn't enforce their use. Fnagaton 11:12, 3 April 2008 (UTC)

The journals published by the IEEE and the ACM predominantly use MB and Mbyte. (You also see the occasional paper using MiB.) The current IEEE Computer Society Style Guide has this entry. MB: megabyte; use Mbyte (40-Mbyte hard disk, 12 Mbytes of memory) [6]

A common claim by the IEC advocates is that the IEC prefixes are use by organizations and "publications that strive to be unambiguous". Could someone name a few? (Other than standards groups.) Who is striving to be "ambiguous"? -- SWTPC6800 (talk) 00:57, 4 April 2008 (UTC)

FWIW, The current IEEE Computer Society Style Guide also has this entry.
M: SI prefix for million or mega (40-Mbyte hard disk, 12 Mbytes of memory)[7]
This defines 12 Mbytes of memory = 12,000,000 bytes and we all know that is incorrect. Doesn't this further illustrate the need to get it right in Wikipedia, even the IEEE Computer Society doesn't get it right. IMO, IEC Binary prefixes are a good way to get it right. Tom94022 (talk) 19:05, 5 April 2008 (UTC)