Template talk:Quantities of bits

RFC: Column name/position/content for binary computing units

The following discussion is an archived record of a request for comment. Please do not modify it. No further edits should be made to this discussion. A summary of the conclusions reached follows.

No Consensus to change header - There was consensus for better clarity over the issues of standards, and industry usage, but not as to what replacement header(s) should be (if any).
Feel free to add a sub-note (as was suggested in the discussion) to the column headers as appropriate. For example linking to: https://www.jedec.org/standards-documents/dictionary/terms/mega-m-prefix-units-semiconductor-storage-capacity (which apparently uses the phrase "common usage") - or whatever other internal or external reference is deemed appropriate, per normal editorial discretion. - jc37 16:45, 1 December 2022 (UTC)

---

This is a multiple question RFC:

1. What should the column name for the column shown in this revision as "JEDEC" be going forward?

• Also common
• Common
• Common Usage
• Computing
• Customary
• Deprecated
• Historical
• JEDEC
• Legacy
• Memory
• Metric
• Traditional

Please indicate your choice from the list above, or if you have an alternative choice not listed, state that.

2. Should the above named column be positioned before the "IEC" column? Please answer yes or no.

3. Should the full list of entries be provided up to yottabyte/yottabit? Please answer yes, no, or the maximum value you would support.

Thank you for your participation. —Locke Cole • t • c 01:19, 10 October 2022 (UTC)

Responses

• Computing, this matches the language used at one of our sister projects, the English-language Wiktionary (see petabyte). Common or Common Usage would be second choice, Traditional would be third choice. Yes/before to the question of position, and Yes to the full list. —Locke Cole • t • c 01:19, 10 October 2022 (UTC)
• Deprecated, as per SI and Binary prefix#IEC prefixes, second choice Memory, Customary, Common Usage or Legacy; No/after for position; No for yottabyte/bit, overdone and the bottom link covers that for now. --Zac67 (talk) 05:12, 10 October 2022 (UTC)
• Customary this is in line with the naming of miles and pounds; "Common" would be a second choice. Memory is too limited, because files sizes use it often. "Metric" is nonsensical. "Computing" makes no sense because what else are Bytes used in? "Deprecated" is correct but too controversial. JEDEC is too little known. Historical, Legacy and Traditional put too much emphasis on time. Keep the decimal column first. Extend binary to TB. −Woodstone (talk) 12:59, 10 October 2022 (UTC)
• My responses are
  • Deprecated, JEDEC and Legacy are all acceptable headings as these all seem concise and accurate. I can also live with Memory because computer memory seems the main use of this notation. (Note added 2022-10-17: Memory doesn't work as a heading because IEC prefixes are also used for computer memory) 'Customary' or 'Traditional' don't work for the binary meaning because (for hard drives and communication speeds) the decimal meaning is also both customary and traditional. 'Historical' implies they are no longer used when we know they are. None of the others make sense to me (all prefixes are for Computing, Common is too vague, and ... why 'Metric'?!?)
  • No
  • It depends. If 'JEDEC' is selected, prefixes listed should be up to giga (not tera) because K, M and G are the only prefixes defined by JEDEC. If 'Deprecated' is chosen, the column can go up yotta because all prefixes are deprecated. If 'Legacy' or 'Memory' is chosen, I can see the table going up tera, and perhaps peta.
• Dondervogel 2 (talk) 21:19, 10 October 2022 (UTC)

  'Customary' would become acceptable if qualified to clarify the column does not apply to hard drive storage. For example Customary (RAM). Dondervogel 2 (talk) 00:26, 15 October 2022 (UTC)

  (see threaded responses below in discussion section)

• Deprecated and legacy are not acceptable, because these are neither deprecated nor legacy units. These are the mainstream units used by the vast majority of sources and operating systems. Headbomb {t · c · p · b} 00:22, 13 October 2022 (UTC)
• Customary, second choice common usage, third choice common (only if we wish to take up less space than "common usage"), The idea that usage must be in a standard to be listed here is at best a distraction, terms like "legacy" and "traditional" give the misleading impression that the usage is not a current norm, and "deprecated" is thus doubly misleading. Before IEC to reflect general usage and reading experience. TB at least, open to PBand EB. NebY (talk) 17:57, 13 October 2022 (UTC)
• Customary seems least worst, as the analogy with imperial weighs and measures has merit. I think there could be room in the template for an asterisk leading to a footnote saying Used by Windows and Android. Keep the column positions as they are. Go up to YB because why not? Leaving the PB+ entries blank just raises questions. Barnards.tar.gz (talk) 21:37, 17 October 2022 (UTC)
• 1 - Jedec is good, the wikilink shows the source; 2 - current position is good; 3 - yottabyte is ok for table length. Cheers Markbassett (talk) 20:29, 19 October 2022 (UTC)
• Deprecated looks most accurate to me. "Historical", "Legacy", "Customary", "Traditional", "Common" etc. imply that there was a de facto standard for these units at one time, but there wasn't. No, the IEC column should stay to the left of the "Deprecated" column. No the "Deprecated" column should only go up to giga-. In fact, why does this column even exist? Most uses of bits are 1000-based, right? It was only *bytes* that were sometimes measured in multiples of 1024? Oh, EEPROMs use "1-Kbit" or "1-Kb" for 1024 bits — Omegatron (talk) 18:05, 23 October 2022 (UTC)
  • Nonstandard would be a good column name, too. — Omegatron (talk) 03:04, 18 November 2022 (UTC)
• Deleted (preferred) or Deprecated. Deprecated per Omegatron's reasoning. Deleted (not an option in this RfC, unfortunately) because the vast majority, if not all the mainstream computer industry doesn't use terminology like “gibibytes,” nor does any portion of the mainstream computer press; it’s proposed standard that hasn’t seen anything close to widespread adoption. Having such units mentioned in Wikipedia’s regular articles is contrary to WP:COMPUNITS, which hashed this out 14 years ago, and does our readership a disservice because it sows confusion. Greg L (talk) 04:08, 24 October 2022 (UTC)
• Delete column - there is no common usage, historical or otherwise, to use kilobit to mean 1024 bits. If there is, please provide sources to the contrary. The argument that kilobyte means 1024 bytes isn't enough. Historylikeyou (talk) 10:08, 30 October 2022 (UTC)

  Also used for memory or similar wording. It was pointed out in the discussion below that kilobits are used for memory chip capacities. Historylikeyou (talk) 13:49, 30 October 2022 (UTC)

• Customary is acceptable for the extended scheme. JEDEC also acceptable. Deprecated or Historical are not acceptable, because those are the current units used in the industry. Headbomb {t · c · p · b} 13:51, 30 October 2022 (UTC)

Discussion

I added "Computing" after seeing it in use at petabyte, otherwise this matches the full list above. —Locke Cole • t • c 01:19, 10 October 2022 (UTC)

Computing would mean it is the predominant form there which isn't true (the whole world of networking doesn't use binary prefixes); JEDEC is a trade organization not normative for units/prefixes and not known to all/many readers, binary prefixes likely predate them; Metric is entirely wrong; Historical might be misleading as they're still in common use. --Zac67 (talk)

In so far as most units go that are seen by the public, the overwhelming majority utilize binary prefixes. The terms are used in many computing contexts as described here, networking is really the outlier. —Locke Cole • t • c 05:47, 10 October 2022 (UTC)

128 × 8 TB ((1024 × 1024 × 1024 × 1024) × 8) files = 1 PB (1,125,899,906,842,624 bytes)

131,072 × 8 TB ((1024 × 1024 × 1024 × 1024) × 8) files = 1 EB (1,152,921,504,606,846,976 bytes)

@Dondervogel 2: With regard to where the units should end, wouldn't it be at least Exabyte/EB (see images)? Microsoft Windows, which according to Usage share of operating systems for desktops/laptops represents about 75% of users (for gaming it rises to 96% for Steam desktop users) uses PB and EB in the binary sense. This also addresses your concern that the units aren't used to refer to storage, clearly Microsoft and Google/Android believe the units do. As the IEC units are also unknown to the vast majority of the world, aren't used in the vast overwhelming majority of our sources, it does our readers a disservice to list them first. —Locke Cole • t • c 21:31, 10 October 2022 (UTC)

Yes, I see your point about peta and exa. On this basis I find exabyte acceptable for 'Legacy' or 'Memory', though not for 'JEDEC'. I prefer to settle on a name for the heading before worrying too much about the order. Dondervogel 2 (talk) 21:55, 10 October 2022 (UTC)

On the other hand I'm having second thoughts about 'Memory' because IEC prefixes are also used for computer memory. My preference remains with Deprecated, JEDEC, or Legacy. Dondervogel 2 (talk) 23:46, 10 October 2022 (UTC)

All memory manufacturers that I'm aware of still utilize MB/GB when referring to memory (especially in marketing and support for consumers). Can you provide some reliable sources for your statement that IEC units are used for computer memory? —Locke Cole • t • c 03:51, 11 October 2022 (UTC)

Here are a few examples (there are plenty more available): Ayers 2018, Balkesen 2013, Beniamine 2015, Chen 2020, Chrobot 2011, Deakin 2017, Fichte 2020, Gruhn 2016, Hildenbrand 2021, Hunter 2021, Jouppi 2017, Karkkainen 2017, Rae 2016, Schurmann 2014, Son 2017, Song 2015, Tatar 2018, Uecker 2013, Vano Garcia 2018, Verdejo 2017s, Wardana 2018. Dondervogel 2 (talk) 09:00, 11 October 2022 (UTC)

Those all appear to be from academia, and while there is some evidence they use these symbols, there's also a lot of evidence they continue to use the standard/common form significantly more often. Do you have any sources that use these units for memory outside of academia? Manufacturers, major software vendors, media outlets? —Locke Cole • t • c 15:44, 11 October 2022 (UTC)

• I stated that IEC prefixes are used for this purpose and they are. I found all of these from Google Scholar, which naturally emphasises academic sources, and there are plenty more where these came from. I contend that IEC prefixes are by far the primary or preferred means of disambiguation used by academic sources, with other forms of disambiguation being virtually non-existent.
• Users outside academia include HP, IBM and Ubuntu. It's not hard to find sources.

Dondervogel 2 (talk) 20:58, 11 October 2022 (UTC)

HP appears to still use standard binary units (site:hp.com "gibibyte" returns zero results; site:hp.com "gigabyte" returns thousands of results). Also, it doesn't matter if they are used for disambiguation, it matters whether they are used at all. And by and large, they are not. —Locke Cole • t • c 21:26, 11 October 2022 (UTC)

• Of course it matters whether they are used for disambiguation. That is what IEC prefixes are for, and both HP and IBM use them for this purpose.
• Why does your reply not mention Ubuntu?

Dondervogel 2 (talk) 00:31, 12 October 2022 (UTC)

...for disambiguation. That is what IEC prefixes are for... Do you have a source for that claim? My reply also didn't mention IBM, but as you only took issue with Ubuntu, I'll start by stating the obvious: Linux is used in only a fraction of consumer-facing systems. Wikipedia presents the world as it is, not as individual editors or standards bodies would like us to present it. That it is used in a handful of places by HP or IBM, or at all by Ubuntu, is irrelevant when these same manufacturers/vendors utilize the mixed usage in public-facing areas such as HP in their online catalog and marketing. These are the things our readers see and are familiar with, and this is what our sources support. —Locke Cole • t • c 17:21, 12 October 2022 (UTC)

• The source is the IEC itself "Mathematics dictates that the disparities resulting from the mixed and incorrect use of decimal prefixes will become increasingly significant as capacities and data rates continue to grow. In IEC 80000-13:2008, all branches of the IT industry have a tool with which to iron out this inconsistency. It eliminates confusion by setting out the prefixes and symbols for the binary, as opposed to decimal, multiples that most often apply in these fields."
• I did not read the 2 links you gave carefully enough and assumed one was for HP and one for IBM
• It seems to me that HP, Ubuntu and IBM use IEC prefixes where they wish disambiguate a binary meaning. In their marketing literature they often do not disambiguate. The purpose of WP is not to market computer products but to educate interested readers.

Dondervogel 2 (talk) 19:41, 18 October 2022 (UTC)

On extending the JEDEC column to PB and EB

@Woodstone: With the images from Windows 10 (see right/above) showing uses of PB and EB in Windows Explorer, would you support extending to those? Headbomb also indicated these units are used by Android as well, which is a large installed base as well for mobile devices (phones, tablets, etc). See also this Intel support page for storage. In so far as "Computing" goes, as I explained, petabyte refers to this definition as "Computing" but the IEC definition as "SI". For these column headers we would have "IEC" as we have now and "Computing" to cover the widespread use of the binary meaning in the larger computer (hardware and software) industry. —Locke Cole • t • c 03:51, 11 October 2022 (UTC)

If these examples are real, I do not object to extend to higher values. However "computing" as a header makes little sense, because where else would these unit be used?−Woodstone (talk) 09:31, 11 October 2022 (UTC)

I think that's the crux of the issue here, we have IEC units attempting to be used widely on the basis of being a "standard" but being largely dismissed by the "computing" industry at-large (hardware, software, media, etc). —Locke Cole • t • c 15:44, 11 October 2022 (UTC)

I don't see anyone attempting to use IEC units widely (are you referring to a WP editor? who did you have in mind?). Even if they were, this has little or no bearing on the template, the purpose of which is to provide information where needed. Dondervogel 2 (talk) 21:07, 11 October 2022 (UTC)

Even if they were, this has little or no bearing on the template Of course it has bearing, our articles must be well sourced, verifiable, and not present undue weight to topics. I won't discuss conduct issues here, but there are editors who are using these units in articles where neither the sources nor our MOS support it. —Locke Cole • t • c 17:22, 12 October 2022 (UTC)

I agree with Locke Cole and share the same concerns. Having this table used in any other article than Binary prefix (and perhaps some closely related articles) induces other editors—many of whom are novices to Wikipedia and don’t know about WP:COMPUNITS—to use the binary prefixes in wholly inappropriate ways; this isn't news to anyone here.

Since probably only about one milliuno of the computer industry uses the binary prefixes, it makes no sense to have these units in articles as doing so only seeds confusion. I think we need another version of this table without the binary prefixes for most ordinary articles about computers. Greg L (talk) 23:39, 12 October 2022 (UTC)

It is clear that these are just attempts at censoring the use of standards-based unambiguous units on WP, because some people just don't want to see them used and resist the change, which is well in progress. These forces have systematically tried to eradicate them here, despite ever increasing usage in new software. New writers come to WP all the time too, introducing new units into articles, but this guard removes them because at one time they and their sock-puppets were successful in writing them out of the style manual. A long time has passed since and there is simply no justification for this. The dishonesty in this is exemplary, incredible. Any competent software engineer working in open-source software these days encounters these units daily, many are required to use binary and metric units per official definitions, or else the software cannot be distributed in a variety of operating systems. The train has long left the station for acceptance and it will simply take time for the old-timers to die out. There is no standards org left that promotes or defines old usage. Some large companies are the laggards, sure, the reasons are more or less clear, but this should not be justification for censorship. The 'rules' in WP policy need to be reexamined and changed. Everywhere else on WP we can can observe a drive to disambiguate unclear terms and such, but not here? I have never encountered an author, IIRC, that is as militant about using the new units, as these refuseniks are about censorship. Open-minded, competent people just walk away and shake their head, knowing where things are headed. kbrose (talk) 00:42, 13 October 2022 (UTC)

This incredible notion of eliminating confusion by censoring the use of the official binary units is so outrageously stupid, because there always was confusion before their definition, and this was the very reason for their introduction. The industry had come to a point where they needed to be unambiguous, just like many other fields had to convert to metric units for compatibility, trade, safety. It is true today, that one can never be certain anymore which system of storage units is in use in many places. It is unfortunate, surely, but the way to get past this is not to go backward–this is definitively impossible–but to be progressive and eliminate every ambiguous use, except perhaps in the discussion of historical facts and contexts. kbrose (talk) 01:08, 13 October 2022 (UTC)

You clearly feel, kbrose, that the "unambiguous" units are swell and should be used more and more. Unfortunately, your opinion, which is predicated on arguing the virtues of the binary prefixes, is against MOSNUM (specifically WP:COMPUNITS), which was the product of a well-established consensus. And, per WP:CONLEVEL, a consensus among a limited group of editors, at one place and time, cannot override community consensus on a wider scale. In short, that ship has sailed and won’t be re-prosecuted here.

So now the only thing that matters is how we better align Wikipedia's articles with our policies. Greg L (talk) 01:24, 13 October 2022 (UTC)

@Headbomb: Can you state a position on the 2nd and 3rd questions please? Thanks! —Locke Cole • t • c 20:44, 13 October 2022 (UTC)

@Dondervogel 2: And yet Microsoft Windows utilizes it for hard drive storage. As do other operating systems. —Locke Cole • t • c 04:45, 15 October 2022 (UTC)

@Locke Cole: The place for discussion is the discussion section. Please move your comment there, so we can discuss it without disrupting the Responses. Dondervogel 2 (talk) 09:27, 15 October 2022 (UTC)

Yes, I accept that Microsoft uses MB, GB, TB ... in the binary sense, but such use is relatively new, especially for the larger prefixes, while hard drive manufacturers have used those same symbols with a decimal sense for decades, so the decimal use is customary for hard drives. The customary use of the binary sense is limited mainly to computer memory. Dondervogel 2 (talk) 20:27, 16 October 2022 (UTC)

They've used the binary versions of KB, MB, GB, etc. since as far back as at least Windows '95, and while I'd need to look into it more, I suspect they even used it in MS-DOS. —Locke Cole • t • c 23:28, 16 October 2022 (UTC)

Agreed. Microsoft used kB since the stone ages. There was no MB and GB at the time. The point is moot anyway. Having this table salted in articles in which the binary prefixes aren’t supposed to be used is inappropriate and serves only to keep the dream alive for a handful of the faithful who must learn to abide by a wider and crystal-clear consensus. And that consensus is that Wikipedia will use “gigabytes” like all the mainstream computer industry and computer press does. And in the rare cases where it’s truly necessary to disambiguate, we can do so—again—like the mainstream computer industry and press does. Greg L (talk) 03:46, 18 October 2022 (UTC)

MB was used in the late 1970s/early 1980s in the decimal sense, long before it came into mainstream use in the binary sense. Similarly with GB (though probably a decade later). And this decimal use continues today, 40 years later, so the customary (or traditional) use of these prefixes is decimal, not binary. You are pointing out that the customary use for RAM (which came much later) is the binary one, which is why I suggested Customary (RAM) for the header. Dondervogel 2 (talk) 08:57, 18 October 2022 (UTC)

The use of KB, MB, GB etc in a binary sense was customary and normal for any binary usage before the invention of KiB, MiB, GiB etc., which were created as a reaction to that established use and have still not entered general use. The very fact that KB, MB, GB etc are the customary binary units as well as being the customary decimal units has been what has so upset you and others, and been the driver for your campaign over so many years to replace them as binary units in Wikipedia. NebY (talk) 09:29, 18 October 2022 (UTC)

@NebY: Yes, that's another way of saying the same thing, and is my main reason for opposing the unqualified 'Customary' as a heading for the JEDEC column. Your post suggests a possible alternative heading to be Customary (binary). I think that's better than 'Customary (RAM)', which would receive my support if added to the RfC options. (There could also be a new 'Customary (decimal)' column for good measure) Dondervogel 2 (talk) 11:55, 18 October 2022 (UTC)

A hasty, badly formatted demo of an another layout for Multiple-byte units
unit Symbol values decimal binary SI & customary customary kilobyte kB 1000 1024 megabyte MB 10002 10242 gigabyte GB 10003 10243 terabyte TB 10004 10244 petabyte PB 10005 10245 exabyte EB 10006 10246 zettabyte ZB 10007 10247 yottabyte YB 10008 10248	unit symbol values binary SI & IEC kibibyte KiB 1024 mebibyte MiB 10242 gibibyte GiB 10243 tebibyte TiB 10244 pebibyte PiB 10245 exbibyte EiB 10246 zebibyte ZiB 10247 yobibyte YiB 10248
Orders of magnitude of data

"Customary" applies to both uses, we wouldn't want to repeat "(binary)" immediately below "binary", and we certainly don't need to puzzle everyone by duplicating a column to support another heading. It may be that part of the problem with headings springs from the table design. The layout suits a normative function rather than a descriptive one; it suits telling people what to do (which units they should use) rather than what they're reading. Wikipedia is of course descriptive, so we run into problems. Here's a quck example of a descriptive approach, reshuffling the bytes table; it's very poorly formatted with odd leftovers from the original but I hope it demonstrates the difference. NebY (talk) 15:37, 18 October 2022 (UTC)

This reminds me a little bit of an old version of the bytes template, and I honestly think something formatted similar to this would likely result in fewer issues and misunderstandings. —Locke Cole • t • c 17:20, 18 October 2022 (UTC)

Binary use of decimal prefixes is deprecated by international standards bodies

(Trying to start a new thread here but I don't know how - this'll have to do until someone comes along to fix it). @Headbomb: The claim Deprecated and legacy are not acceptable, because these are neither deprecated nor legacy units is incorrect. The use of decimal prefixes with a binary meaning is deprecated by international standards bodies ISO and BIPM. For example:

• ISO 80000-1:2009 "SI prefixes refer strictly to powers of 10, and should not be used for powers of 2. For example, 1 kbit should not be used to represent 1024 bits (2^10 bits), which is a kibibit (1 Kibit)."
• BIPM SI Brochure 9th edition (2019), p143 "The SI prefixes refer strictly to powers of 10. They should not be used to indicate powers of 2 (for example, one kilobit represents 1000 bits and not 1024 bits)."

Dondervogel 2 (talk) 19:30, 18 October 2022 (UTC)

Of course they are deprecated. Only standards orgs that have a say over standards, have the ability to deprecate old usages or definitions. When users change to a new method as a matter of fashion, it does not constitute deprecation, only disuse or preference, or something. But it is here not anymore about the correct official status, but a hard-necked resistance to accept modern standards terminology, for only personal reasons, and to prolong the confusion by censoring new writers and emerging trends. The picture they want to paint that nobody uses the new units is patently and intentionally false, because it seems impossible to ignore the mass of modern software that has made the transition. This is just endless repetition of the same bullshit, wandering from article to article. It is time to change the matter at the heart and change the obsolete language in WP policy statements, which were simply forced by this kind of reckless disregard for standards, with the help of proven sock-puppets. Serious editors frankly don't participate in these issues as fervently as the refuseniks, because they have more productive things to do. kbrose (talk) 20:37, 18 October 2022 (UTC)

They aren't deprecated by the industry, which is what matters. That's like saying inches and feet are deprecated in construction because meters are around. They aren't. Headbomb {t · c · p · b} 22:36, 18 October 2022 (UTC)

No, that is a false analogy. Industry has no means to deprecate something they have no primary control over. They can only defy standards or adopt them. That is exactly reflected in the language by JEDEC. The standards bodies that define the the metre do not specify that the inch should not be used because of some kind of ambiguity. Certain industries have adopted metric measures because of the cost and danger of confusion, but that is not deprecation per se. Perhaps a government can dictate deprecation for their jurisdiction. An industry can deprecate norms that they created or control. kbrose (talk) 22:52, 18 October 2022 (UTC)

The use by certain companies of the binary interpretation is no valid reason to perpetuate their use in new products, specifications, or documentation. These companies have a long standing investment in their products and the support systems thereof. They are extremely risk-averse because additional confusion by changes in the product line would likely cost enormously. Instead, we see exactly what one would expect, namely new systems and software using the new prefixes and use the metric interpretation of the old ones rigorously. This is a long-term process as one would expect. kbrose (talk) 23:08, 18 October 2022 (UTC)

A spell checker from a computer company thinks that 16 gigibits must be 16 servings of chicken giblets

Why are you still arguing the merits of terminology like “kibibits” and advocating its adoption, Kbrose? This was settled in 2008 by a widely established consensus. Hardly anyone but some odd Sheldon Coopers who program in Linux know what that weird terminology means. My spell checker (from a computer company, no less) just flagged “gigibits” and suggested “giblets”; so much for its “adoption” by the industry in any shape, form, or fashion. It’s time for the die-hard fans to straighten up and get with the program.

Now the task at hand is to get all our computer articles in compliance with WP:COMPUNITS. Tables featuring these odd prefixes only serve to keep the dreams of a handful of hold-out wikipedians alive that by continually salting every computer-related article they can possibly find with these tables, like billboards every 16 kibifeet along a highway (4.9938 kilometers), that less experienced editors who don’t know to read MOSNUM will start *oopsy*-using the units and the grass roots swell of adoption will force the hands of Dell and Apple and the rest of the computer world, who will finally see the light and follow Wikipedia’s way. Well…

Perhaps. Maybe that will happen. But the community spoke and it’s time to cease with fervently wishing for the Pleasant Outcome Fairy to grant our wishes.

WP:COMPUNITS is clear. Mentioning the units at every turn for no other reason than to mention the units at every turn doesn’t cut it and never did. It’s time to use the table of binary prefixes only in articles directly discussing the units. The decision now is what, precisely, a table for regular computer-related articles, as WP:COMPUNITS intended, will look like. Greg L (talk) 00:28, 19 October 2022 (UTC)

Childish diatribes like this just confirm and underscore the unserious, personally opinionated and motivated nature of these characters. Quod erat demonstrandum. kbrose (talk) 17:20, 20 October 2022 (UTC)

@Greg L: No. The task at hand is to decide how to harmonise presentation of the 'JEDEC' column in this and its two sister templates, which after 10 months of stability have become inconsistent. If you have something sensible to say about the RfC, say it. Or if you want to start a second RfC on how this template is used, do that instead. In the meantime, stop cluttering this talk thread with irrelevant nonsense about kibifeet and chicken giblets. Dondervogel 2 (talk) 07:05, 19 October 2022 (UTC)

It may not have been a good idea to start this thread. It seemed Dondervogel 2 wanted to reopen the idea of using "deprecated" as a heading, but it might have been better to acknowledge the opposition to that and consider alternative approaches. For example, it would be possible to add a note to the layout I offered above to the effect that BIPM and IEC deprecate the customary use of decimal prefixes (as in MB, KB) for binary values. NebY (talk) 12:41, 19 October 2022 (UTC)

@NebY: I started this thread to correct a factual error. Do you consider correcting a factual error to be a bad idea? Dondervogel 2 (talk) 13:26, 19 October 2022 (UTC)

I thought you might not like the way it went. Ah well. Would you like to respond to the suggestions I've made for moving forward? NebY (talk) 13:37, 19 October 2022 (UTC)

• The only part I didn't like was Greg_L's childish reaction. At some point some innocent bystander is going to be tasked with closing this RfC, and we can all make her task less complicated by discouraging such nonsensical distractions.
• I fully accept that 'Deprecated' is not favoured by others, but please let's acknowledge the indisputable factual statement that they are deprecated. I am willing to consider alternative layouts, and I appreciate your good faith suggestion to clarify the deprecated status. That would be a step in the right direction.

Dondervogel 2 (talk) 13:58, 19 October 2022 (UTC)

Oh, please desist, Dondervogel 2, with inflammatory rhetoric like declaring other editors are “childish,” which—given your history—is intended to make other editors lash out so you can taken them to ANI with complaints of AGF and impugning your character.

You have been editing against a widely felt consensus that was settled in 2008, and such behavior has many of the hallmarks of tendentious editing, which is disruptive, so you must start trying to find the strength to go with the flow.

Now, in the last month, you got caught with your hand in the cookie jar, are clearly intent on continuing to use the binary prefixes against a wider consensus that is memorialized on WP:COMPUNITS, persist on advancing (re-raising ad nauseam) arguments from the discussions that were raised, heard, and rejected in 2008, and play every game in the book to redirect the conversation to ways to make the binary prefixes *prettier* in scores of articles that shouldn’t have them even mentioned because doing so causes confusion that results in less knowledgeable editors, who are unaware of all the wikidrama that went into the decision to not use those unrecognized units on Wikipedia, using them right in the middle of articles (e.g. The computer had 16 mebibits of RAM). Those bizarre units that our readership hasn’t seen anywhere else except here should only be mentioned in one or two articles on that very subject. Please stop asking everyone to pretend that causing other editors to edit this way was an unforeseeable and unfortunate accident.

It’s time now for someone to go make their own sandbox with a proper table of prefixes that we can use on all regular computer articles. Once that’s ready, we’ll discuss whether the table is well formed, accurate, and complete, and then we'll swap out articles that have this table with the proper one. Greg L (talk) 14:44, 19 October 2022 (UTC)

• I never said you were childish. Your post was childish and I stand by that.
• You have so far made zero contribution to the substance of the RfC (what heading to use for the "JEDEC" column). Instead you make false accusations against me and other editors without a shred of evidence to support your unfounded claims, which in any case are unrelated to the RfC. Do you really have nothing better to do with your time?

Dondervogel 2 (talk) 08:21, 22 October 2022 (UTC)

@Kbrose: Please stop adding your comments wherever you feel like and use threaded conversation. Or don't contribute at all. I don't care. I'm also sorry that 99.99% of the world doesn't use the terms you seem to have a vested personal interest in promoting here on Wikipedia, so much so that you refer to people opposed to using them as "refuseniks" and other colorful names. Maybe in 50 years when I'm buying a 100 yottabyte storage unit to upload my consciousness to I'll think back to you and wonder if I shouldn't call it a yobibyte for old times sake. Serious editors frankly don't participate in these issues as fervently as the refuseniks, because they have more productive things to do. Maybe you should go do those things before you find yourself blocked from editing. —Locke Cole • t • c 15:31, 22 October 2022 (UTC)

My work is to reflect the reality of standards and their adoption, which you intentionally ignore or try to cover up by childish and stupid insinuations, bullying, and such. The most widely used operating systems in the world are already using these standards in one way or another. Get a grasp of reality and do something useful. kbrose (talk) 21:16, 22 October 2022 (UTC)

My work is to reflect the reality of standards and their adoption I wonder if your work might also include accepting that there is little, if any, significant adoption of these units even after 24 years? The most widely used operating systems in the world Microsoft Windows is the most widely used desktop operating system in the world. It doesn't use those terms. Android is the most widely used mobile operating system in the world. It doesn't use those terms. Adobe Creative Cloud is the most widely used graphic design and media production software in the world. It doesn't use those terms. Get a grasp of reality and do something useful. Yes, please do. —Locke Cole • t • c 03:36, 23 October 2022 (UTC)

You really have no clue as to your statements, but I'd rather suspect you do know, and this is just some kind MAGA game. The Linux kernel is documented using binary prefixes for memory blocks and metric units where needed. Just picked a random quote when opening the manual: "The generator will repeatedly access two 100 MiB sized memory regions one by one. " The units are used consistently and pervasively. kbrose (talk) 06:59, 23 October 2022 (UTC)

The Linux kernel The Linux kernel that is used in less than 1% of desktop operating systems? Really? That's your silver bullet? Nevermind the fact that Android, which is derived from the Linux kernel, uses a user interface that uses the traditional KB/MB/GB units and not the IEC units (thus hiding it from users, almost as if... they know their users won't comprehend gibibyte/etc). —Locke Cole • t • c 01:26, 24 October 2022 (UTC)

@Locke Cole: 99.99% of the world doesn't use the terms seems to be the world as you know it = OR. In the consumer world, MS Windows is by far the most popular desktop system (75% of the desktop market[1] but only ~30% overall[2]) and we know it's completely ignoring IEC prefixes. However, in the professional world things may be very different. A clear majority of web services runs on Linux[3] which does use and acknowledge IEC prefixes. But @all, that discussion would need to take place for WP:MOSNUM, here's not the right place. Let's focus on the problem at hand which is deciding about the column header. Also, major changes to the table need a separate thread imho. Currently, those issues and all the bickering water down the focused discussion which we should seek to continue. --Zac67 (talk) 08:47, 23 October 2022 (UTC)

(Is it hard to put your comment in the right place or am I gonna be forced to clean up after editors going forward?) The onus is on those wishing to make the change to prove their position. Not on those defending the status quo. Besides, Google is your friend (well, actually, not your friend as Google shows how infrequently terms like "gibibyte" and "tebibyte" are used in the wild relative to "gigabyte" and "terabyte"). That's not OR. That's looking at a cross-section of sources quickly to determine the validity of the claim that IEC units are somehow widely used and none of us happened to notice. (Hint #1: they aren't widely used, and kbrose protesting the opposite repeatedly won't change that singular fact from being true; Hint #2: and usage in operating systems/software is just one place where IEC units fall flat; broaden your view to hardware manufacturers, system builders, tech news sites, and the media at-large and it gets even worse for gibibyte and friends) —Locke Cole • t • c 01:26, 24 October 2022 (UTC)

Back to the exam question

Unnecessary, the RFC is above, and is still running. This is disruptive and as Neby indicated, RFC closers are more than capable of closing this discussion at the end of the RFC without you coloring or skewing the discussion. —Locke Cole • t • c 18:34, 23 October 2022 (UTC)

If someone tried to close this RfC now they would find it hard to see the wood for the trees. The main exam question is what heading to use for the 'JEDEC' column (for want of a better name - I use the term 'JEDEC' here because it's use was stable for 3 years). A brief review of responses so far reveals support (at least acceptable) for either Customary or Customary (RAM) (5 editors), Common or Common usage (4 editors), Deprecated (2 editors), JEDEC (2 editors), Legacy (2 editors). This suggests we might make some progress with a reduced set of options such as:

• Common
• Common usage
• Customary
• Customary (RAM)
• Deprecated
• JEDEC
• Legacy

Would this help?

• Pinging all editors who have ever contributed to this talk page @Barnards.tar.gz:, @Compvis:, @Cybercobra:, @Greg L:, @Headbomb:, @Jeh: @Kbrose: @Locke Cole: @Markbassett: @NebY: @Omegatron: @Quantling: @Quondum: @Rebroad: @SarahTehCat: @Shreevatsa: @Stepho-wrs: @StuartBrady: @TechControl: @Woodstone: @Zac67:

Dondervogel 2 (talk) 12:04, 23 October 2022 (UTC)

This is a multiple question RFC; Dondervogel 2 has recast it to omit the second two questions and removed options. Dondervogel 2, please don't understimate closers; much lengthier RFCs have been closed before. Also, there are constructive suggestions in the above discussion, in amongst the vituperation and rehashes.

Anyway, the 3 questions are:

1. What should the column name for the column shown in this revision as "JEDEC" be going forward?

Also common

Common

Common Usage

Computing

Customary

Deprecated

Historical

JEDEC

Legacy

Memory

Metric

Traditional

Please indicate your choice from the list above, or if you have an alternative choice not listed, state that.

2. Should the above named column be positioned before the "IEC" column? Please answer yes or no.

3. Should the full list of entries be provided up to yottabyte/yottabit? Please answer yes, no, or the maximum value you would support. NebY (talk) 14:04, 23 October 2022 (UTC)

Just to clarify, I have not recast the RFC. I have suggested recasting it to make it simpler for the closer. The default it to leave the original wording unless there is consensus to change. Dondervogel 2 (talk) 14:48, 23 October 2022 (UTC)

The conflicting views only emphasize the confusion about and the inadequacy of the deprecated use. Delete the column entirely, and only present what the standards are. There can be no confusion or disagreement with that. Deviations from the standards should not be tabulated, but discussed in prose where needed and context can be provided. kbrose (talk) 17:24, 23 October 2022 (UTC)

The wording of the RfC was never discussed, but with hindsight, deleting the column should have been one of the options as there are several editors who favour deletion of the JEDEC column for bits. Should it be added? Dondervogel 2 (talk) 18:28, 23 October 2022 (UTC)

Use of 'kilobit' to mean 1024 bit

@Historylikeyou: there is no common usage, historical or otherwise, to use kilobit to mean 1024 bits – memory chip capacities are generally stated in "kilo/mega/gigabits" referring to powers of 1024. I don't think we need a source for that... --Zac67 (talk) 12:41, 30 October 2022 (UTC)

@Zac67: Several editors have questioned the binary use for bits. A source would be helpful, if only to establish how this non-standard representation of 1024 bit would be written. Would that be kbit, Kbit, kb or Kb? Dondervogel 2 (talk) 13:07, 30 October 2022 (UTC)

@Zac67 you're right, but I didn't know that. Here is an example (PC Hardware in a Nutshell, 3rd Edition by Robert Bruce Thompson, Barbara Fritchman Thompson (O'Reilly)). Most of the discussion on this talk page is about bytes which is not relevant, obscuring the issue. Historylikeyou (talk) 13:42, 30 October 2022 (UTC)

@Dondervogel 2: As per WP:MOSNUM, multiples of 1024 bit are written as Kbit on WP (with capital K=1024 and "bit" not abbreviated as b). Needless to mention that I'd prefer Kibit. --Zac67 (talk) 14:23, 30 October 2022 (UTC)

I'm not asking how they appear in Wikipedia articles, but how they appear in reliable sources. And my question extends to binary interpretations of megabit, gigabit, etc. Dondervogel 2 (talk) 15:11, 30 October 2022 (UTC)

@Dondervogel 2: EEPROMs are specified using "1-Kb" or "1-Kbit" to mean 1024 bits. Only example I can think of, but that's a good enough reason to leave the column in the template.

@Zac67: I, too, would like standard units to be used throughout the encyclopedia, and it was proposed a long time ago, with broad support, but a few users strongly disliked the IEC prefixes and refused to compromise, so there was no consensus. 🤷‍♂️ I don't know what happened after that; I quit Wikipedia for a while and have tried to stay out of it since, for my own sanity. (Someone pinged me to comment on this proposal, though.) — Omegatron (talk) 03:25, 18 November 2022 (UTC)

@Omegatron: Thank you. Can I assume 'Mb' or 'Mbit' would be used for mebibit (and 'Gb' or 'Gbit' for gibibit)? Dondervogel 2 (talk) 14:13, 18 November 2022 (UTC)

@Dondervogel 2: I've never worked with ones that big, but they do appear to exist: "1-Mb" = 131,072 × 8 bits = 1,048,576 bits. "1 Gb" … "128 Mbyte" — Omegatron (talk) 01:00, 19 November 2022 (UTC)

Thank you for clarifying. In that case the JEDEC column seems to make sense in its present form, up to 'Gbit' for gibibit. Dondervogel 2 (talk) 01:07, 19 November 2022 (UTC)

Closure

Requested closure here as the RFC was delisted some time ago and comments have mostly came to a stop. —Locke Cole • t • c 00:05, 29 November 2022 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

New prefixes

Editors might like to join the discussion at Wikipedia talk:WikiProject Physics#Add ronna- and quetta- to units articles?. NebY (talk) 23:22, 18 November 2022 (UTC)

Tera and the JEDEC column

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

---

The question arises at Bit and byte prefixes whether the JEDEC column should include the prefix tera. JEDEC defines the prefix giga as a 'multiplier equal to 1 073 741 824 (230 or K3, where K = 1024)', and similar (binary) definitions are provided for kilo and mega. On this basis I can see the case for including KB, MB and GB in the JEDEC column, but there is no equivalent entry for tera, and therefore no justification for extending the column beyond giga IMO. What do others think? Dondervogel 2 (talk) 01:01, 17 December 2022 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

The discussion you refer to at Bits and bytes took place years ago, with Fnagaton and his band of socks, hell bent on disruption. The discussion has been centralized at bits for more than a year. It does not make sense to hold 3 separate discussions on the same question. Dondervogel 2 (talk) 08:37, 17 December 2022 (UTC)

Where to hold a discussion on the JEDEC column in the ‘bits’, ‘bytes’ and ‘bits and bytes’ templates?

The options I see are

1. hold the discussions separately on the 3 separate talk pages
2. hold one discussion for all 3 templates here at ‘bits’, as has been done for more than a year
3. hold one discussion at a centralised location other than here (please specify where)

Dondervogel 2 (talk) 10:54, 18 December 2022 (UTC)

My preference is #2 (it has worked for a year and I see no reason to change). Are there any other suggestions? Dondervogel 2 (talk) 10:56, 18 December 2022 (UTC)

How about none at all per WP:DEADHORSE? Headbomb {t · c · p · b} 12:10, 18 December 2022 (UTC)

#4 None at all. I am astonished at the idea that discussion here "has worked", at the idea that another "discussion on the JEDEC column" would be fruitful, and at any such question being posed without option #4 (none at all. Please consider that asking "how can I get rid of this thing I don't like, I've been trying for 14 years and it's not gone well" is itself disruptive. NebY (talk) 17:36, 18 December 2022 (UTC)

That's rich coming from someone who just proposed starting a parallel discussion on the same subject at MOSNUM. In any case I am not trying to get rid of anything. My purpose is to identify a suitable place to discuss harmonisation of the 3 templates. Dondervogel 2 (talk) 18:25, 18 December 2022 (UTC)

WT:MOSNUM#Table in COMPUNITS is not a proposal to start a new discussion about the content of the table. You might like to reconsider your response there. NebY (talk) 18:33, 18 December 2022 (UTC)

I've already asked jc37 to re-examine their RFC closure above to see if they see a consensus on the 2nd and 3rd questions, that should address the "Tera" question and beyond potentially. Absent that, I've already pointed you back to the old discussion which was edited as recently as a couple of months ago, and there is a source there for "Tera" at a minimum. There's nothing, IMO, to discuss beyond seeing if the RFC closure is clarified. Otherwise, I agree with Headbomb and NebY on their points. Be grateful the discussion at AN/I appears to have gone your way Dondervogel 2, and stop beating this dead horse lest your actions change any minds there. —Locke Cole • t • c 20:30, 18 December 2022 (UTC)

Hi, Locke Cole. To respond to your request for clarification:

I suppose I can answer each, but they're kind of moot due to the discussion concerns, as I'll explain in a minute.

On the question of "column, before or after", the result was somewhere between "No Consensus to change" and "Keep as is" - Which, essentially, are pretty much are the same result, but I think does allow for a followup discussion, if wanted.

As for the other - Also "No Consensus", but as I noted in the close: "There was consensus for better clarity over the issues of standards, and industry usage". This was the fundamental main crux for the discussion. (Which is part of why I didn't spell out the results of the nom's questions.) Unless/until you all find consensus on clarity between "standards", and "industry usage", the other questions are pretty much moot. One way forward could be adding the reference note, as I mentioned in the close. But, again, as I mentioned, that's up to normal editorial discretion (WP:BRD, etc.) - jc37 17:30, 19 December 2022 (UTC)

Why is this column still an obsession by certain writers here. The only reason it is labeled as it is, is because it provides the only hope for refusenics of unambigous storage units to hang on to a sliver of credibility, when every standards bureau that actually has a voice in the subject matter has deprecated the binary interpretation and stood firm on the new standards, and when new software written these days uses these prefixes accordingly and without which the software may not be distributed in certain environments. And JEDEC in fact agrees with them and refers to their definitions in clear language by pronouncing their ambiguity and deprecation. They state their intent unambiguously in listing them because of ongoing usage. They do not define them in any binding manner. Why is this simply ignored here and why is this column even there? We are listing examples of outdated usage solely because an industry interest group still explains the old usage. Why is it not sufficient to display the standard definitions and discuss deviations in each article in prose? In every other article of units and natural constants or metrics and such the standards orgs are followed rigorously. Why hot here? It's obvious that there still are WP editors who let their personal tastes cloud their mind and deny readers a modern, accurate, unambiguous presentation of these subject matters. Childish belittling the sounds or pronunciations of the units is much like teenager bullying. When the newest SI prefixes got defined this fall, WP editors were eager and quick to add them to every table there is, without regard to actual usage, and without criticism in stupid opinions about their sound or pronunciation. None of these sound any more or less stupid or amusing than giga or pico or any other one. Get rid of the column, or fill it out with a header Deprecated. kbrose (talk) 20:15, 18 December 2022 (UTC)

"refusenics"

Also known as the vast, overwhelming, majority of people. Headbomb {t · c · p · b} 01:56, 19 December 2022 (UTC)

So wrong. The 'vast, overwhelming majority' does not really care, they have no choice in the matter, because a small group of executives likes it this way, and gaslights the majority into believing that essentially every disk drive made in a half century+, longer than any personal computer has existed, is wrong in using proper, unambiguous units displayed on labels and packages. This deception is the real story. kbrose (talk) 03:46, 19 December 2022 (UTC)