Talk:Hard disk drive

Refimprove tag

I'm glad to see the biggest problem the article has now is a lack of references in the "speed" section. My goodness, here I was worried about organization and clarity; maybe even filling in glaring holes in the content. Like, what if anything happened between 1956 and 1998, say? Instead we can just insert "feel good" tags into articles. I should have just inserted a "cleanup" tag, that's as good as actually *doing* the work, right? Or we could grab a mop. --Wtshymanski (talk) 20:51, 25 November 2010 (UTC)

I'm sorry my actions bother you, but part of doing the work when you're added/changing material is providing references. --Nuujinn (talk) 20:56, 25 November 2010 (UTC)

Right, right. Happy tagging. What does the "guild of copy editors" do, again? --Wtshymanski (talk) 21:31, 25 November 2010 (UTC)

Again, I'm sorry to have miffed you, but this article is not exempt from WP:V. --Nuujinn (talk) 23:10, 25 November 2010 (UTC)

• Frankly I don't see per WP:V "any material ... likely to be challenged" in the Speed section so I also wonder what particular value Nuujinn has added to the article. In the absence of specificity I am going to remove the tag. If Nuujinn wants to challenge anything within the section he could do a useful service by challenging specific material or better yet finding a reference for the material or, if it is in error, correcting it. Just sprinkling tags about really doesn't help anyone. JMO Tom94022 (talk) 05:39, 26 November 2010 (UTC)

• There are large sections of this article that are not sourced, and the material here, like all material on wikipedia, should be attributed to a reliable source. You're entitled to your opinion, but using a tag to request improvement in references is a legitimate action--that's what the tags are for. Presumably editors adding material to this article have sources at hand, so why not add them. --Nuujinn (talk) 15:02, 26 November 2010 (UTC)

If you have references, please add them. In my opinion, adding a refrence is more value to the article than adding tags. --Wtshymanski (talk) 15:26, 26 November 2010 (UTC)

I agree, but I don't have sources for everything. I have added some material and a couple of references, and did some copy edits. Tags do serve a purpose in drawing attention to areas that could use some improvement, and there's no need to get wound up about them. --Nuujinn (talk) 15:50, 26 November 2010 (UTC)

RTFP — "in practice not everything need actually be attributed." The Section tagged has one reference, the article has eighty plus references. It seems to me the editors of this article are doing a reasonable job of referencing. If Nuujinn has a legitimate challenge to any statement in the section or article he should identify them or better yet fix them. Sprinkling Refimprove tags indiscriminately about serves no purpose other than to add to the 101,635 existing tags that for the most part are not being worked on. Tom94022 (talk) 17:25, 26 November 2010 (UTC)

Respectfully, I disagree. And, according to WP:BURDEN, any material that is not sourced can be removed. Personally, I think a refimprove tag at a section level is a very polite way to ask for improvements in sources, and is less invasive than extended use of the cn tag, but since you all seem to prefer the latter I'll use them here. Whether or not articles that are tagged are not being worked on is open to discussion. --Nuujinn (talk) 17:43, 26 November 2010 (UTC)

Wikipedia doesn't respect it's contributors, that's why this sort of thing happens. If we trusted our contributors, we could accept that they knew what they were talking about when the sumarized perfectly standard concepts related to a topic. Instead we have the present model where in principle any random drive-by IP editor can challenge "The sky is blue" with a tag, and Stephen Hawking couldn't refute it.

I picked up an older edition of Moeller's handbook on upgrading and repairing PCs at Value Village; it's old, but my packed-away copy was even older. Anyway, it has a lot of really quotable sections on disk drives that I can cite here, though it is not useful on the 25-odd years before the XT. We still haven't done enough in the article, to this date, to explain jsut *why* we went from $2000 and 5 megabytes in 1980 to $100 and 2 terabytes 30 years later. That is a far more interesting story than the incredibly banal list of dimensions that we do have (also unreferenced), though at least someone's put in the biggest capacity of drive made in each form factor with dates. Where's giant magnetoresistance, thin-film heads, oxide vs. plated media, glass disks? Not here, and not in the History of hard disk drives article either. Though I notice one brave soul suggested a History of external disk drives article. Good luck figuring out how a disk drive actually works, from the present text; I hope we can improve this. First make it readable, then make it accurate, then make it comprehensive. --Wtshymanski (talk) 17:51, 26 November 2010 (UTC)

RTFP — "It has always been good practice to make reasonable efforts to find sources yourself that support such material, and cite them." Apparently you disagree with WP:V when you tag facts not likely to be challenged and don't make a reasonable effort to find sources of the facts you think might be in dispute. I have removed two of your tags because are facts that "in practice do not have to be attributed" as follows:

• The published RPMs of all HDDs are mathematically transformed in to milliseconds
• The published average seek times of all HDDs are in milliseconds
• The published data rates of all HDDs mathematically yield data transfer times for any reasonable block transfer of at most a few milliseconds.

There are numerous cites to HDD specs already in the article than make the adding of such cites to these facts superfluous. Tom94022 (talk) 18:02, 26 November 2010 (UTC)

Again, respectfully, you are missing the point. The text reads "These two delays are on the order of milliseconds each. The bit rate or data transfer speed once the head is in the right position creates lesser delays." Yes, drive speeds are measured in milliseconds. Any delay can be measured milliseconds. It would be more useful to have a specific reference with an actual speed. Also, please refrain from making assumptions about whether or not I have made a reasonable effort to find sources.

Wtshymanski, I agree completely with you regarding the lack of history and scope of the article. --Nuujinn (talk) 18:14, 26 November 2010 (UTC)

Geologic time can be measured in milliseconds, but the sentence says the delays are "on the order of milliseconds" - "on the order" usually means " not millions of times larger" but comparable. Does it really serve the reader's interest to know the seek time of a RAMAC was 650 milliseconds and of an IBM XT drive was 83 milliseconds, but the Binford 6100 disk drive has a seek time of 7 milliseconds and the model 6100A (with a Windows logo on it) takes only 6 milliseconds? --Wtshymanski (talk) 19:21, 26 November 2010 (UTC)

(outdent) Many of the 80+ "references" are press releases or data sheets for drives. We can do better than *that*. Let's find some references that don't fall out of Google or at least come from real books, not data sheets. --Wtshymanski (talk) 19:38, 26 November 2010 (UTC)

IMO data sheets are good enough in the sense and spirit of WP:RS. What if a manufacturer cheated? The competition would blame them faster than you can spell l-a-w-y-e-r. They might be a bit optimistic, ie a new drive in perfect condition, not one which barely made the quality check. I do think however that the resulting error is <1ms. The seek time paragraph within the Speed section is well-written IMO; by providing one to two significant figures, it des not trespass into false precision.

Maybe the form factor, capacity, shock resistance, and access time characteristics, along with the data sheets should be moved into the History article, as they are still changing. User.Zero.Zero.Zero.One (talk) 16:26, 22 February 2011 (UTC)

There is a difference between a parts catalog and an encyclopedia article, but less difference between a parts catalog and a Wikipedia article. We love parts lists on Wikipedia because they are so easy to write. --Wtshymanski (talk) 17:18, 22 February 2011 (UTC)

List in History Section

I suggest we all should feel free to add a FEW more highlight measurements and possibly metric conversion to the list of measurements in History Section, but I think the list should be short with all product specific examples in the main history article. I also think the values of the measurements in the list should be generic so that the list does not have to be continuously updated as the areal density progression continues.

Accordingly I backed out some recent changes that reflected the latest products and prices but really didn't add to the point of the section. As it reads now I believe the progression is reasonably accurate for today's state of the art. Comments? Tom94022 (talk) 18:15, 8 December 2010 (UTC)

Technology > Components and overall quality

After having read the whole article, here are my thoughts:

• The overall quality is quite high and very accessible. It is well written and has a good flow.
• The components subsection inside the technology section is a bit obscure to me. I think it needs rewording.
• There are many images showing the same things in the articles. Maybe consider removing some of them?

Kudos to the writers anyway.

#!/bin/DokReggar -talk 09:50, 24 December 2010 (UTC)

I took a first hack at the components section; it still needs some work which I will get to later. Tom94022 (talk) 20:55, 24 December 2010 (UTC)

Regarding the components section: it seems better to me, however needs a further bit of editing; it is really unclear for me, and I think I have more knowledge on physics and mechanics than the lay person who would read this article. #!/bin/DokReggar -talk 09:15, 11 January 2011 (UTC)

You can tell me here what u think is unclear but why don't u take a stab at improving the section - it's usually better when the work is done interactively? Tom94022 (talk) 22:00, 11 January 2011 (UTC)

What is Speed

disk drive speed can be measured by? —Preceding unsigned comment added by 173.178.52.166 (talk) 03:46, 11 January 2011 (UTC)

Just above: Depends on what you call speed: data rate or angular velocity? #!/bin/DokReggar -talk 09:15, 11 January 2011 (UTC)

Speed can also refer to seek times or average access time, it all depends upon context.Tom94022 (talk) 22:00, 11 January 2011 (UTC)

Damaged hard disk photos

hi guys! i have damaged hard disk photos!! could anyone update it in wikipedia? thanks a lot!!! roberto spadim - brazil - são paulo/sp http://www.spadim.com.br/hd%20agra.zip —Preceding unsigned comment added by 200.161.132.95 (talk) 17:21, 25 January 2011 (UTC)

Write Verification

Can anyone confirm whether writes are verified immediately or at all? If so, is this done by the read head being just behind the write head, or is it re-read on the next revolution? Ajoiner (talk) 13:10, 29 January 2011 (UTC)

In the vast majority of drives there is just one head per surface. (A few high-performance server drives have two heads and two separate actuators, but this is used for seek performance improvement, not for read-after-write verification.) They don't do read-after-write on the next rev either. If they did, writing would be markedly slower than reading, and it isn't... if you want read-after-write verification you'd have to program that in the host. Jeh (talk) 13:23, 29 January 2011 (UTC)

Thanks. Ajoiner (talk) 08:57, 30 January 2011 (UTC)

error wiht harddisk

my harddisk when i press power button . doesnotngoto thw BIOS mode and safemode it hanged....... —Preceding unsigned comment added by 113.199.237.222 (talk) 04:09, 5 March 2011 (UTC)

please update revenue

Worldwide revenue from shipments of HDDs is expected to reach $27.7 billion in 2010, up 18.4% from $23.4 billion in 2009[77] corresponding to a 2010 unit shipment forecast of 674.6 million compared to 549.5 million units in 2009.[78]

Can someone update this? Also, it would be nice to get the help requests out of here. This isn't tech support.130.166.41.124 (talk) 18:10, 14 March 2011 (UTC)

"Someone" could be you. Take the existing references and see if you can find later versions, for example. Jeh (talk) 18:15, 14 March 2011 (UTC)

Vandalism

I saw that 194.72.129.29 has made several vandalism edits (all reverted) in this page and others. Maybe a temporary block would be useful? #!/bin/DokReggar -talk 08:27, 18 March 2011 (UTC)

User:194.72.129.29 -talk is a school that seems to have periodic vandalism spasms, I suggest u report it Tom94022 (talk) 16:08, 18 March 2011 (UTC)

Performance Section Vandalism

RaptorHunter has reverted my attempts to consolidate and update the Performance Section of this article. Speed previously a separate section is more accurately a subsection of Performance and is more accurately entitled Access Time. The subsection of Performance on Data Transfer Rate has a "non-sequitor on manipulation of sequential data" which is better off eliminated. The recently added sentences on comparing rotational speed is just wrong. Its hard for me to understand what RaptorHunter disagrees with since he reverts with minimal comment, the latest of which is a personal attack. I would appreciate it if other editors took a look at my proposed edits and tried to make them better. Tom94022 (talk) 22:36, 1 April 2011 (UTC)

Just looked.. They look fine to me., and I agree with your assessment of the discussion of "manipulation of sequential data". (Fact is that most drives spend most of their time operating in a very much NON-sequential manner, so even if there had been some valid, properly RS'd points in that text its real-world applicability is in question.) Jeh (talk) 23:27, 2 April 2011 (UTC)

IEC prefixes and WP:MOSNUM

Per Wikipedia:MOSNUM#Quantities_of_bytes_and_bits: One of the exceptions for binary prefixes is:

" in articles specifically about or explicitly discussing the IEC prefixes. "

This table compares SI and binary prefixes. That counts as "explicitly discussing" it. Saying that 1000 TB = 931 TB is just confusing. For the purpose of this table, it makes much more sense to use binary prefixes for comparison purposes: 1000 TB = 931 TiB

Also, some editors are under the mistaken impression that no one uses IEC prefixes, ergo they are not worth mentioning. Here is a list of all the software using IEC prefixes: Binary_prefix#Software and Timeline_of_binary_prefixes#2000s The second most popular OS: Mac OSX defines SI units properly. (1GB = 1000000000 bytes) The entire linux kernel use the IEC standard. Is it really so terrible to have a table comparing them in this article?

To the guy that keeps changing IPs to edit the table, yes WP:MOSNUM does say that. I have directly quoted it. If you can't refute my point on this talk page, then you don't get to continually revert. --RaptorHunter (talk) 17:11, 7 April 2011 (UTC)

This article is about hard drives not IEC prefixes so that exception in MOSNUM does not apply. The table is to compare kilobyte with numbers of bytes. The consensus on Wikipedia is to not use IEC. WP:MOSNUM says do not use this disambiguation. According to MOSNUM it makes more sense to numbers and not use IEC. Why are you editing against that consensus? —Preceding unsigned comment added by 220.255.2.169 (talk) 17:26, 7 April 2011 (UTC)

I asked for feedback on WT:MOSNUM. Now we can wait to see if the claimed exception applies. —Preceding unsigned comment added by 220.255.2.155 (talk) 17:37, 7 April 2011 (UTC)

The whole point of the table...is lost...if it doesn't show the binary prefixes! Using 6 different IPs in 30 minutes is not usual behavior for a contributor. Assuming it's the same person and that all of Singapore hasn't develped a grudge against the IEC at the same time. Referring the multi-IP to the Manual of Style, Numbers, Quantities of Bits and BYtes, doesn't seem to be sinking in as the IP is misquoting the version we can read. --Wtshymanski (talk) 18:01, 7 April 2011 (UTC)

This ISP uses a shared HTTP proxy for all HTTP requests, that explains the range of IPs changing. I cannot turn off this feature because it is part of the network setup at the ISP. Nothing suspicious there. The point of the table is not lost since it uses the WP:MOSNUM disambiguation method of using the number of bytes. MOSNUM says IEC must not be used. I am asking you both instead of editing against consensus and reverting my changes I am asking you both to leave those changes there and wait for feedback from the MOSNUM talk page where people who know about MOSNUM will be able to tell us is IEC is allowed to be used or not. I think it is not allowed to be used and that is why I think my edits have consensus and improve the article. I ask you both, will you accept the consensus of MOSNUM? 220.255.2.81 (talk) 18:29, 7 April 2011 (UTC)

I encourage you to create an account so we can leave messages to each other more easily. Please see Wikipedia:Why_create_an_account?--RaptorHunter (talk) 20:13, 7 April 2011 (UTC)

The manual of style explicitly says...oh to heck with it. Off to ANI for a rangeblock request....--Wtshymanski (talk) 18:55, 7 April 2011 (UTC)

As this is a content dispute, I have protected the page for a day. Favonian (talk) 19:34, 7 April 2011 (UTC)

It should be noted that this is just one editor who keep hoping IPs so he can edit the table. We already have consensus from the username editors that a table comparing binary prefixes to SI prefixes should use both for clarity. This clearly falls under the exception of Wikipedia:MOSNUM#Quantities_of_bytes_and_bits, quoted above. I hope that the editor who keeps hoping IPs will grow tired of this game by the time the page protection expires.--RaptorHunter (talk) 19:49, 7 April 2011 (UTC)

Duly noted. Hopefully the extra day will bring a clear consensus involving uninvolved editors as well. Favonian (talk) 19:52, 7 April 2011 (UTC)

• FWIW, I agree that this is an appropriate use of IEC Binary Prefixes. Tom94022 (talk) 20:02, 7 April 2011 (UTC)

FWIW, I also agree. It is Just Plain Wrong to have 1000 TB and 931 TB in the same row of the same table representing the same amount, regardless of what WP:MOSNUM says. Jeh (talk) 20:37, 7 April 2011 (UTC)

• This is an appropriate and useful application of IEC prefixes, per WP:COMMON, and arguments by Wtshymanski, RaptorHunter, and Jeh. I believe the IP editor means well, but an inflexible application of the guideline WP:MOSNUM leads in this case to the logical absurdity of saying that 1 TB is equivalent to 0.9095 TB, using exactly the same unit of measurement. It should be obvious that any guideline cannot possibly cover all conceivable circumstances, and this use does seem to fall within the spirit of the exemption noted there and above. The argument that using both prefixes improves clarity seems to be the best for our readers. Although I have edited this article in the past, I'm uninvolved in this particular edit war or issue. Good call on the page protection. — Becksguy (talk) 21:30, 7 April 2011 (UTC)

Note that WP:MOSNUM proclaims itself to be a guideline, which is a different thing from policy. And it even says at the very top: "Use common sense in applying [this guideline]; it will have occasional exceptions." This looks like one of them. Even if there was no credible argument that the exceptions in the WP:COMPUNITS section applied here (and there is), there's certainly credible argument that the global "escape clause" applies. Jeh (talk) 23:33, 7 April 2011 (UTC)

• I do not think this is an appropriate use of IEC prefixes for a few reasons.
  1. Is the column of extra values needed? I think not because there are already two columns of values to disambiguate, one in the decimal sense and one in the binary sense. It looks overly messy to include fractional values with the IEC prefix.
  2. As far as I know there have been no instances of hard drive manufacturers using IEC prefixes and certainly none cited for the use in the table so it is against WP:OR and WP:NPOV to try to make it look like they are used for this topic. Since the sources for this topic do not use IEC prefixes then this article should not do so especially since WP:MOSNUM gives alternatives.
  3. Even assuming the extra column of values is needed with the fractional values the question then becomes one of how to write these values and on this WP:MOSNUM is clear where it says "Disambiguation should be shown in bytes or bits, with clear indication of whether in binary or decimal base." There are some exceptions "when the article is on a topic where the majority of cited sources use the IEC prefixes" but this is not relevant since the sources do not use IEC prefixes. Or "when directly quoting a source that uses the IEC prefixes" again this is not relevant for the same reason. Lastly there is "in articles specifically about or explicitly discussing the IEC prefixes" which does not apply because this this article is not specifically about or explicitly discussing IEC prefixes.
• Instead the table can be edited in this way which makes it smaller and easier to read as well as following the guidelines.

SI prefixes (hard drive)	Decimal equivalent	In the binary sense
1 TB (Terabyte)	1 * 10004 B	0.9095 * 10244 B
1000 GB (Gigabyte)	1000 * 10003 B	931.3 * 10243 B
1,000,000 MB (Megabyte)	1,000,000 * 10002 B	953,674.3 * 10242 B
1,000,000,000 kB (Kilobyte)	1,000,000,000 * 1000 B	976,562,500 * 1024 B
1,000,000,000,000 B (byte)	-	-

• As everyone can see the table still makes it clear what numbers are in the binary or decimal sense. So as you can all see when written like this it is superfluous to include the third column. Keep in mind the text in WP:MOSNUM regarding IEC prefixes was written to give guidance on how to reduce the use of IEC prefixes to places where it is strictly necessary to use them. In this case it is obviously not necessary. It follows WP:OR and WP:NPOV by specifically not including any conversions to IEC prefixes. It follows WP:MOSNUM by using powers.
• Now then, does anyone have any disagreements about using that edited table? Glider87 (talk) 00:50, 8 April 2011 (UTC)

For some strange reason you are doing everything in your power to avoid using these binary prefixes. You even changed the words binary prefix to the very strange "in the binary sense". Now how is that vague description better than using the actual term. "in the binary sense" is nothing more than a euphemism for binary prefix.

Furthermore, your argument that HDD manufacture's do not use binary prefixes is invalid here. The table is comparing how the OS measures space to how the hard drive manufacturer's measure space.

It's as if some Wikipedia editor's want to do everything they can to censor any Byte measurement that has a little "i" in it. The "i" is accepted by an international standard's body. It's used any many tools, open source programs and websites. It adds clarity to the article when distinguishing SI versus binary units. I know the computer industry has been using MB and GB to refer to non-SI units of data for a long time now. That doesn't make it correct. The SI units have existed long before the computer age and it seems that finally computer software is catching up to the fact that they have been doing it wrong for all this time. Accept it.--RaptorHunter (talk) 01:19, 8 April 2011 (UTC)

It does no good to waste one single line of this article to using units of measure (gibibytes) when such units and terminology are not used by one single computer manufacturer in any materials directed to their customer base—not in their advertising, brochures, packaging, or instruction manuals. Because of this fact, no computer-related magazines that are directed to a general-interest readership use such terminology. It is not the proper role of Wikipedia to try to lead the world by using such language here in hopes the rest of the world will see how way-cool and logical the IEC prefixes are; we follow the way the real world works. Greg L (talk) 01:25, 8 April 2011 (UTC)

So if it doesn't appear in a magazine it must not exist? Here is a nice list of a whole bunch of software that all uses binary prefixes: Binary_prefix#Software. Binary prefixes are here. Get used to it.--RaptorHunter (talk) 01:41, 8 April 2011 (UTC)

I did not say "if it doesn't appear in a magazine it must not exist". The consensus is that if it is not how the real world works then Wikipedia does not report it. Binary prefixes are not commonly used, so theyt are not to be used in this article.Glider87 (talk) 01:58, 8 April 2011 (UTC)

See also Timeline_of_binary_prefixes#2000s. The tide of history is turning against you.--

On the contrary it shows that despite ten years adoption is still low and that it has not turned yet. So it is not up to Wikipedia to try to advocate any change it is up to Wikipedia to report significant points of view in a neutral way. This is another nail in the coffin for using IEC prefixes. Glider87 (talk) 09:31, 8 April 2011 (UTC)

RaptorHunter (talk) 02:56, 8 April 2011 (UTC)

RaptorHunter replying to your comment: The article about binary prefixes uses the phrases "in the decimal sense" and "in the binary sense" so it is not my use, nor is it "very strange" or vague. It is very relevant that HDD manufactures do not use IEC prefixes because to try to use IEC prefixes in an article about hard drives is against WP:NPOV. That is the consensus that was agreed for WP:MOSNUM. As for "is accepted by an international standard's body" WP:MOSNUM says "Wikipedia follows common practice regarding bytes and other data traditionally quantified using binary prefixes" and then says "Despite the IEC's 1998 guideline creating several new binary prefixes". It is not relevant what a "international standard's body" thinks because what is relevant is Wikipedia showing what is in common use. A question for you, keepng in mind Wikipedia follows common practice is it common practice to use IEC prefixes for the subject of hard drives? The answer is, of course not. Another question, has consensus changed to show a common practice of using IEC prefixes? Answer, no of course not. The sources do not use IEC prefixes for this topic so to be neutral this article has to follow that example. Do you have any argument against the proposed table that is not WP:IJDLI? Glider87 (talk) 01:40, 8 April 2011 (UTC)

Again, you seem to be under the impression that no one is using this or that it's at least not in common usage. I've have posted a link to a long list of people using it in software everyday: Binary_prefix#Software. All the table does is compare to different methods of measuring data. Deal with it.--RaptorHunter (talk) 01:45, 8 April 2011 (UTC)

That list is not relevant to this topic. Can you provide a long list of sources for hard drive manufacturers using IEC prefixes? The answer is you have not. Since you have not then IEC prefixes are not to be used in this article since that would be against WP:NPOV WP:OR and WP:SOAP. The edited table I provided above still compares different methods of measuring data and does this without going against WP:NPOV. Glider87 (talk) 01:58, 8 April 2011 (UTC)

Greg yes we do we "follow the way the real world works". Wikipedia is not a soap box for IEC prefixes. Glider87 (talk) 01:43, 8 April 2011 (UTC)

It occurs to me that the goal of the table is to contrast correct numbers with how they are displayed by a very commonly used operating system (among other software). And Windows does say e.g. "500 GB", but using "GB" in the binary sense. One solution: in the table proposed above by Glider87, change the "binary sense" column head to "As displayed by Windows and some other software", use the TB or GB notation in that column, and not use it at all elsewhere in the table? That is after all how Windows displays hard drive and file sizes - with SI-like prefixes but used in the binary sense. And it sidesteps the "1000 GB = 931 GB" issue. It does not however address the software that exists that does use IEC prefixes, or SI prefixes in their decimal sense. Do we just pretend that no such software exists? That HD and file sizes are never displayed with either SI prefixes used in the decimal sense, or with IEC prefixes? Jeh (talk) 02:06, 8 April 2011 (UTC)

Another solution is to punt the issue entirely and refer to the Binary prefixes article, which no one seems to have a major problem with (at least not yet). Jeh (talk) 02:10, 8 April 2011 (UTC)

I think we have to consider what most software uses in the real world and not what a minority of software uses. In this case this article is not about the minute details specifically about IEC prefixes used by some software. That kind of topic is already discussed in the binary prefix article. This article is not on the same subject and is not meant to rehash detailed information in binary prefix. This article is a level or two removed from such minute details. Reference WP:UNDUE and ask is it relevant to this article that a minority of software uses IEC prefixes or is it relevant that the majority of software does not use IEC prefixes. So I think for this article it is best not to use IEC prefixes and instead adopt a neutral disambiguation advocated by WP:MOSNUM. This stance is one of the five pillars of Wikipedia. Remove the table completely and just link to binary prefixes would be an alternative. Glider87 (talk) 02:18, 8 April 2011 (UTC)

Did you read what I wrote in my 01:25, 8 April 2011 post, Glider87? Yes, I agree with you 100%. You are correct and this debate is moot. This was settled a long time ago and it’s final. See WT:MOSNUM#Avoiding confusing IEC prefixes. This attempt to use the IEC prefixes in a computer-related article is precisely what MOSNUM’s guidelines are intended to prevent. There isn’t a snowball’s chance in h-e-double hockey sticks that there would ever be a consensus to allow the IEC prefixes to be used in this fashion. The admin might as well unlock the article for this isn’t a close call and the argument that the IEC prefixes can be brought back is without any foundation.

To RaptorHunter: Please take two readings of Complete rewrite of Units of Measurements (June 2008) and call me in the morning if you still don’t *get it*. We don’t use the IEC prefixes like you are trying to do because it doesn’t matter if they are way-cool. The consensus after long debate was that they are clearly unfamiliar with our readership and will not be used on Wikipedia unless the article is squarely on the issue of discussing the units of measure—not by merely using them in a table.

User:Jeh offers a perfectly satisfactory solution to this, none of which relies upon using a stuttering unit of measure (gibibytes) that precious too few readers have ever heard of. Greg L (talk) 02:17, 8 April 2011 (UTC)

You seem to be under the assumption that because you and Glider have written so much more text that anyone else, that you have a consensus here. You do not. Even if we all don't have the ability to vomit text forth like some other editors here, the fact remains that you don't have a consensus. It's a hotly debated issue.

Futhermore, I don't want IEC prefixes because they are "way cool". In fact I think they sound kind of stupid. (kibibyte?) They need to stay because they clarify the difference between SI units and computer units. The software industry has been misappropriating SI prefixes to mean something they don't for decades. The ideal solution of course would be for software makers to actually use SI prefixes correctly, just like hard drive manufactures are now. 1 GB = 1 billion bytes and 1 MB = 1 million bytes. Then we can forget about the kibibytes.--RaptorHunter (talk) 02:24, 8 April 2011 (UTC)

Consensus is defined "by the quality of the arguments posted" and good quality arguments are those that cite relevant policy and guidelines and also cite existing consensus. Consensus is not defined by a small group of editors appearing at a certain point in time. Existing consensus in the archive that Greg posted says not to use IEC prefixes because of many good reasons which was discussed for a long period of time by many editors. You have not posted anything that refutes those very good reasons in the archive or the very good reasons I have posted here on this talk page. What you have written looks a lot like WP:IJDLI which is a "weak and feeble an argument" and "hold little to no water at Wikipedia". Glider87 (talk) 02:34, 8 April 2011 (UTC)

Well if you had actually read any of my arguments (which appear at the top of this scetion), then you would know that I am not arguing that WP:MOSNUM ought to be repealed. I am simply stating that the rules provide a clear exception in this one case because the table is explicitly discussing the difference between binary and SI prefixes.--RaptorHunter (talk) 02:37, 8 April 2011 (UTC)

Well, I don’t agree with you at all and find your arguments unconvincing. MOSNUM’s guidelines are clear on this and your continuing to harp on it and say that up is down is merely tendentious. That this article had to be locked down because of this shows that this is bordering on disruptive. Sorry. Goodbye. Greg L (talk) 02:40, 8 April 2011 (UTC)

RaptorHunter I have read your arguments and they are refuted by the arguments I just posted. That is to say the exception you cite is not valid given the consensus found in WP:MOSNUM WP:NPOV WP:UNDUE WP:IJDLI and WP:SOAP. Glider87 (talk) 02:43, 8 April 2011 (UTC)

I agree with Greg L's stance. Let's avoid units that are so far from common usage as to be derisible. If there must be a note to another article that explains a tiny issue, well perhaps; but it isn't the job of this article to "clarify the difference between SI units and computer units".  GFHandel.   02:59, 8 April 2011 (UTC)

• So given that RaptorHunter's arguments have been refuted. I agree with Jeh that we should "punt the issue entirely" and refer to the Binary prefixes article. Remove the "Following are the several ways of reporting one Terabyte." and the table. Instead replace the text "some operating system utilities" with some operating system utilities. That should solve the issues. Anyone have any problems with that? Glider87 (talk) 02:51, 8 April 2011 (UTC)

Everyone but you and Greg has a problem with that. You are attempting to censor the existence of a unit recommended from an international standards body. We are not trying to force it on every article in the pedia. Just explain the difference with a handy chart.--RaptorHunter (talk) 02:54, 8 April 2011 (UTC)

The chart makes more sense without IEC prefixes because the WP:MOSNUM consensus says IEC prefixes "are not familiar to most Wikipedia readers". I am not attempting to censor anything. I am attempting to find the middle ground by agreeing with Jeh that the issue should be punted to the relevant article. It helps to organise and simplify this article and delegate the superfluous information to the text already in the other article. Glider87 (talk) 02:59, 8 April 2011 (UTC)

• RaptorHunter, quoting you: You are attempting to censor the existence of a unit recommended from an international standards body. “Censored” “Banned”. “Matter of style”. Call it what you will. That argument was the very foundation of the IEC-prefix proponents over on MOSNUM when this issue was settled. Ultimately, that argument was rejected by a consensus. So you are re-raising arguments that were carefully considered and soundly dismissed by the community. Do you understand this point? Or do you understand that this is a sad fact but think that by convincing Glider87 and me you can turn MOSNUM on its ear?

  The BIPM (the SI gods) say that a space must be used between the numeric value and the unit symbol. Thus, it is supposed to be 75 % and not the 75% the rest of the world (including Wikipedia) uses. It matters not that there is a standard from a widely respected organization saying we are supposed to do otherwise. If you want to debate the issue here, then—by definition—an outcome per your wishes on this article will be contrary to MOSNUM. So if you want to change MOSNUM, go argue your point there and see how far you get. Sorry. Greg L (talk) 03:04, 8 April 2011 (UTC)

Once again: I am not arguing that WP:MOSNUM ought to be repealed. I am simply stating that the rules provide a clear exception in this one case because the table is explicitly discussing the difference between binary and SI prefixes.--RaptorHunter (talk) 03:13, 8 April 2011 (UTC)

• Splendid. Well then, the crux of the dispute is what MOSNUM means. So this discussion clearly belongs there then. Your argument that you may merely mention the units in a table so that means they are de facto “being discussed directly” is totally absurd and everyone else at MOSNUM would be able to see that.

  Forgive me, but your manner of posting here and your clear intent reminds me of Thunderbird, who would be intimately familiar with all the goings-on at MOSNUM back in the day and would know full well what MOSNUM currently means. By any chance, are you the editor who was behind the old Thunderbird of MOSNUM fame account? Please flatly declare that you did not operate that account so that no Check User will need to be performed anytime soon. Greg L (talk) 03:37, 8 April 2011 (UTC)

RaptorHunter, going back and editing [1] old comments that have already been replied to (and also refuted) is generally frowned upon because it disrupts the ebb and flow of the discussion. Now then as explained a couple of times already the exception you cite does not allow what you claim in this particular topic. The conclusion formed by the WP:MOSNUM consensus is that IEC prefixes are not to be used in this article. The alternative, Jeh proposed and I just fleshed out, is to punt the entire detail to the existing text in another article. Glider87 (talk) 03:22, 8 April 2011 (UTC)

I didn't change my post. I appended to it, to clear up some confusion. It's an important fact that: The second most popular OS: Mac OSX defines SI units properly. (1GB = 1000000000 bytes) and that The entire linux kernel use the IEC standard.--RaptorHunter (talk) 03:26, 8 April 2011 (UTC)

You added text claiming "Also, some editors are under the mistaken impression that no one uses IEC prefixes" which misrepresents the ebb and flow of the discussion. It also misrepresents the argument I actually posted further on down the talk page. Now back on topic, since your arguments are refuted what is left is the proposed change to remove the table and modify the article text to include a link to another article. Glider87 (talk) 03:35, 8 April 2011 (UTC)

You've refuted nothing. Just because you and your buddy Greg got together and decided everything I said is wrong doesn't make it so.--RaptorHunter (talk) 03:39, 8 April 2011 (UTC)

Claiming that your arguments have not been refuted when they have been shown to be refuted by the WP:MOSNUM consensus is not a strong argument, see WP:IJDLI for the reasons why. Can you refute the stronger arguments I've presented above? Glider87 (talk) 03:46, 8 April 2011 (UTC)

Well, there’s your problem. Two other editors disagree with you so you must be right! (Niiiiice) Try arguing this on WT:MOSNUM and see how far you get with your logic about what MOSNUM prescribes and proscribes. Slam dunk – adios muchachos. Greg L (talk) 03:46, 8 April 2011 (UTC)

I have explained myself time and time again refuting everyone of your arguments. Everyone else, but you two and an ip-hopping vandal agree with me. Goodnight everyone.--RaptorHunter (talk) 03:52, 8 April 2011 (UTC)

You have not refuted that arguments I posted, you have not tackled them at all except to restate the incorrect assumption about the MOSNUM exception which is basically nothing more than "I just don't like it" which is a weak argument. The clarification from MOSNUM is that your interpretation of MOSNUM is incorrect. How about you ask other editors involved in the MOSNUM consensus if your assumption is correct? I bet you they will say roughly the same as Greg and myself have been saying. Glider87 (talk) 03:56, 8 April 2011 (UTC)

The point is RaptorHunter for the exception in MOSNUM to be used you have to first show how the IEC prefixes are being discussed by the reliable sources relevant to the article not by you discussing IEC related text. Your desire to discuss or add IEC prefixes in the article is not good enough to use the exception. You need to support what you want to put in the article with reliable sources. Unfortunately you are not a reliable source (neither am I by the way). That is one reason why your interpreation of MOSNUM is incorrect. Glider87 (talk) 04:16, 8 April 2011 (UTC)

I salute my fellow editors. Not since my days reading USENET have I seen so much effort expended on such a...recondite...topic. With any luck this will end up with WQA or ANI or even a user ban. We should at least nominate this for WP:LAME. --Wtshymanski (talk) 13:15, 8 April 2011 (UTC)

Arbitrary section break

When the article is unlocked I will add a reliable source to the first sentence of Capacity measurements. The second sentence has a reliable source. Together they make this section about the differences between a drive's claimed capacity and its capacity as reported by most operating systems and IMO makes this section about IEC binary prefixes (they have been introduced and are being used to eliminate this confusion). I expect that Glider87 and Greg L will disagree but I think the consensus is that usage here is a valid exception. Tom94022 (talk) 16:57, 8 April 2011 (UTC)

Here is the quote:

Drive displays a smaller capacity than the indicated size on the drive label

Decimal vs. Binary:
For simplicity and consistency, hard drive manufacturers define a megabyte as 1,000,000 bytes and a gigabyte as 1,000,000,000 bytes. This is a decimal (base 10) measurement and is the industry standard. However, certain system BIOSs, FDISK and Windows define a megabyte as 1,048,576 bytes and a gigabyte as 1,073,741,824 bytes. Mac systems also use these values. These are binary (base 2) measurements.

Western Digital

Tom94022 (talk) 17:15, 8 April 2011 (UTC)

That's not true of MacOS X any more. They are now using SI prefixes. Jeh (talk) 20:53, 8 April 2011 (UTC)

Good point, I will truncate appropriately. Tom94022 (talk)

BTW, when u say they are now using SI prefixes do u mean they report drive and file usage in a decimal sense or using IEC binary prefixes for memory or ??? Tom94022 (talk) 22:48, 8 April 2011 (UTC)

They report that a 1GB file = 1000000000 bytes.--RaptorHunter (talk) 23:16, 8 April 2011 (UTC)

The consensus is against using IEC prefixes because the are not familiar to most Wikipedia readers. The table from Glider87 shows how it is clear without using IEC prefixes. You cannot say they are using IEC prefixes unless they specifically use -bi or -iB. MacOS X is not using IEC prefixes. What does have consensus is Jeh's proposal to remove the content and link instead. 220.255.2.93 (talk) 23:14, 8 April 2011 (UTC)

This is the same ip-hopper from yesterday that forced the page lockdown. See Wikipedia:Administrators'_noticeboard/Incidents#Users_in_the_range_220.255.2.XXX--RaptorHunter (talk) 23:18, 8 April 2011 (UTC)

The fact my ISP has a dynamic HTTP proxy does not make my point less valid. The fact is the consensus is against using IEC prefixes because that is what is said [2] [3] [4]. Greg confirms this consensus and he was one of the people involved in building that consensus. If you disagree with what you think MOSNUM means then as Greg said this becomes an issue for MOSNUM to debate not this article talk page. I already asked for feedback on WT:MOSNUM. The feedback from Greg is that consensus is against what you say MOSNUM means. The proposal from Jeh to remove the content and link has consensus and it makes sense. As GFHandle says it isn't the job of this article to clarify the difference between SI units and computer units. At the very least the IEC prefixes should be removed from the table and any disambiguation or comparison of sizes should only use byte values. RaptorHunter or Tom94022 do you have any argument that refutes the facts as described above? 220.255.2.72 (talk) 23:46, 8 April 2011 (UTC)

• I support Tom's RS addition to the section, together with the IEC UoM (Units of Measure) version of the table. I oppose Glider87's version of the table. Despite relabeling the secondary column as "In a binary sense", it's still IEC's binary UoM without the actual UoM terms, and in scientific notation that's not intuitive for the average reader. The header even links there. — Becksguy (talk) 01:28, 9 April 2011 (UTC)

No WP:MOSNUM says that using to disambiguate numbers is to be used instead of IEC because IEC is not familiar. So using IEC is not intuitive for the average reader. Even if a number is in the binary sense MOSNUM says to use numbers not the IEC prefixes. Why are you going against that consensus? 220.255.2.81 (talk) 01:46, 9 April 2011 (UTC)

Even if something uses a binary number of bytes it still has to be disambiguated using numbers not IEC prefixes. That is what WP:MOSNUM says. That is consensus. Saying that you support something that is against consensus without giving a good reason does not change that consensus. Consensus is not a bunch of editors turning up and saying they like something without reason. Consensus is made of stronger ideas than that. 220.255.2.77 (talk) 01:55, 9 April 2011 (UTC)

Please Create an account, so we know who is posting what messages. It will make these discussions much easier.--RaptorHunter (talk) 01:56, 9 April 2011 (UTC)

I don't need an account to point out mistakes that have been made. Do you any any argument that refutes the facts and consensus as described above? 220.255.2.46 (talk) 01:59, 9 April 2011 (UTC)

If you actually cared to read, what I wrote. You would see that I have laid out my arguments over and over again. Also, when you IP constantly changes, it looks like the messages are coming from different people. It also makes it impossible to leave messages on your talk page. This is very abnormal behavior. Please Create an account. It's free, quick and easy.--RaptorHunter (talk) 02:02, 9 April 2011 (UTC)

But as discovered the arguments you made are not correct. Do you have any new argument to make that refutes the facts and consensus above? 220.255.2.47 (talk) 02:05, 9 April 2011 (UTC)

The messages you left on the user talk page were false accusations of vandalism. If false accusation are all you have to write on a talk page then it is better that you do not write at all. 220.255.2.62 (talk) 02:08, 9 April 2011 (UTC)

We won’t have the standard handful of IEC fanatics violating the clear intent of MOSNUM. The consensus finding was that the IEC prefixes are unfamiliar to readership because not one single manufacturer of personal computers when communicating with their customer base uses such terminology. Consequently, no general-interest computer magazine uses such terminology to describe capacity—that is, unless it is a special article talking about the IEC’s proposal that colossally failed to catch on in the real world. The normal routine, when faced with communicating drive capacity, merely state something like “2 terabytes” and add an asterisk leading to a footnote that “1 terabyte equals 1,000,000,000,000 bytes.” MOSNUM provides guidance on how to do this. It’s not difficult. Any proponent of the IEC prefixes who participated in those past MOSNUM discussions and who now attempts to exploit this flare-up by trying to use measures that are entirely unfamiliar to our readership in contravention of the guidelines will end up the subject of an ANI because they know better and are just being tendentious and disruptive. That applies to Tom94022 and to Thunderbird (a.ka. RaptorHunter, I suspect). Greg L (talk) 03:26, 9 April 2011 (UTC)

Don't be so paranoid as to assume that every user that disagree with you must be a sockpuppet. Futhermore your threat to use checkuser against me [5], is in clear violation of the Check user policy which forbids a "threat against another editor in a content dispute."--RaptorHunter (talk) 03:49, 9 April 2011 (UTC)

Third time of asking: Do you have any new argument to make that refutes the facts and consensus above? 220.255.2.46 (talk) 04:21, 9 April 2011 (UTC)

• I have made an edit that has consensus, this removes the table and makes a reference to binary prefixes where the difference is explained in more detail. RaptorHunter and others do not keep on reverting to insert IEC prefixes because as GFHandel said it isn't the job of this article to clarify the difference between SI units and computer units. 220.255.2.79 (talk) 04:30, 9 April 2011 (UTC)

  Please do not portray your edits as somehow more important ("has consensus") than others. WP:MOSNUM is a guideline and is subservient to the actual consensus here. If someone would like to provide a brief explanation of what is wrong with the current article without posturing, please do so. Johnuniq (talk) 04:38, 9 April 2011 (UTC)

  The problem is that some editors want to use this article to discuss IEC prefixes and that is not within the scope of this article. 220.255.2.40 (talk) 04:43, 9 April 2011 (UTC)

  Nonsense. If the IEC prefixes are not banned from Wikipedia, this is the one article that is entitled to use them. Johnuniq (talk) 04:51, 9 April 2011 (UTC)

  The IEC prefixes are banned from Wikipedia for general articles. Only those articles that specifically discuss them are allowed to use them. See WP:MOSNUM for the reasons why. This is a general article because the scope of this article does not specifically cover IEC prefixes. 220.255.2.22 (talk) 05:26, 9 April 2011 (UTC)

  If you would not make so many posts here, you would not make so many errors. A guideline cannot "ban" anything. Johnuniq (talk) 05:34, 9 April 2011 (UTC)

  You used the word "banned". WP:MOSNUM says "The IEC prefixes are not familiar to most Wikipedia readers so are generally not to be used" and so on. The guideline does give very good reasons not to use them. The consensus talks leading up to that guideline give even more good reasons. To use them here on this page can you give one good reason that is better than the consensus talks? As it says on WP:MOSNUM "Consistent standards make articles easier to read, write, and edit." I ask you why do you want to make this article depart from consistent standards and make the article harder to read, write and edit? 220.255.2.20 (talk) 05:38, 9 April 2011 (UTC)

• I have just made Wtshymanski and RaptorHunter aware of the three revert rule. I hope this encourages edits within the consensus formed and to not add IEC prefixes or the table back again unless a clear consensus has been formed on this talk page. 220.255.2.28 (talk) 04:41, 9 April 2011 (UTC)

There is no "consensus formed" as we are in an ongoing process of developing consensus here on the talk page as you noted. Although it appears to me that consensus is leaning toward inclusion of the IEC units. Reverting from the protected version of the table was just done without consensus and should be restored. — Becksguy (talk) 05:15, 9 April 2011 (UTC)

@ User:Greg L: Please try to avoid comments like "standard handful of IEC fanatics". Even if it is true (which I don't believe), it doesn't help to develop consensus here, as it tends to put the focus on the contributors rather than on the content. Thank you. — Becksguy (talk) 04:46, 9 April 2011 (UTC)

The consensus to not use IEC prefixes is on WP:MOSNUM and that consensus applies as a basis to all relevant articles. I disagree that consensus leans towards using IEC prefixes on this talk page those who want to use IEC prefixes have not given a good reason to use them when considering the arguments made for the WP:MOSNUM consensus. Since there has not been a good reason to use them then the consensus formed in WP:MOSNUM which says to not use them takes priority. 220.255.2.76 (talk) 05:22, 9 April 2011 (UTC)

I cannot believe this is still going. Here are some reasons for not including the IEC prefixes in this article:

1. They are not recognised by the average readership of WP.
2. They serve to make the article more complex, and we should be aiming to explain things as simply as possible.
3. Try walking into your local computer store and asking for a 500 gibibyte hard-disk drive. The guy will stare at you until you come back to reality.
4. Just because a unit can be used in an article doesn't mean it has to be.
5. Their use here is simply elitist box-ticking—not because they deepen the understanding of the topic.
6. It is not up to this article to provide conversion factors between different units. The most commonly used unit is sufficient for the average reader to get their bearing/scale.
7. There has been previous discussions on the use of the IEC units, and the consensus has been not to use them.
8. I've worked in the IT industry for quite a while now, and I've never (that's right, never) come across the use of a single one of the IEC units. (That's just the opinion of one expert witness in case one was necessary.)
9. If some mention need be made of such irrelevant terms, a footnote is more than sufficient.

The article is fine without mention of the terms. Those wanting to include the terms are doing so simply because they can, not because there is any compelling reason to do so. How about moving on to tackle the 100-kibi more important issues needing attention on WP?  GFHandel.   06:02, 9 April 2011 (UTC)

Yes indeed GFHandel, so much for the claims of consensus being for IEC prefixes. The problem is ask someone who wants to use IEC prefixes why they was to be inconsistent with other articles or the guideline. Then there is no reason given except "I want to" and here at Wikipedia that is not a good enough reason to add content because it goes against WP:OR and WP:NPOV. Johnuniq asked for a summary of what was wrong with the article before, this is why. 220.255.2.87 (talk) 06:58, 9 April 2011 (UTC)

Comments on MOSNUM and CONSENSUS My arguments for including IEC units, or for rewriting that part of the section, apart from this posting, are separate, as this is just about consensus. After a rereading of the internal references, I have these comments on the the relationship of MOSNUM and the ongoing consensus discussion on this talk page:

1. WP:MOSNUM is a guideline, not policy.
2. In the hatnote at the top of the MOSNUM page, it says: "Use common sense in applying it; it will have occasional exceptions." and links to WP:COMMON and to WP:IAR.
3. WP:COMMON says in part: "Wikipedia has many rules. Instead of following every rule, it is acceptable to use common sense as you go about editing. Being too wrapped up in rules can cause loss of perspective, so there are times when it is better to ignore a rule. Even if a contribution "violates" the precise wording of a rule, it might still be a good contribution."
4. WP:COMPUNITS, the section of MOSNUM that specifically deals with computer related units of measurements says in part: "Specify if the binary or decimal meanings of K, M, G, etc. are intended as the primary meaning.". The table at question uses the SI units as primary, with the IEC units as secondary, therefore it's clearly permitted in the guideline.
5. The archived discussion, referenced in MOSNUM, compete rewrite of UoM, is almost three years old. The last three years is a blink of the eye relative to the field of Elizabethan literature, for example, but it's generations relative to computer technology.
6. That archived discussion had about a dozen participants, not an atypical number when compared to the number of participants in AfDs. How can a claim be made that an extremely small subset of the community can set up a guideline for the entire community, when it can take that number of people just to delete or keep one single article. That is not a meaningful or reasonable community wide consensus.
7. Regardless of how much weight one places on the discussion by a group of twelve editors from three years ago, consensus can change, per WP:CCC.

Conclusion: The subject table with IEC units in a secondary position does not violate WP:MOSNUM, per WP:COMPUNITS and WP:COMMON. — Becksguy (talk) 06:53, 9 April 2011 (UTC)

No that is not the conclusion because using IEC prefixes in the secondary position still violates WP:MOSNUM. The correct method to disambiguate is to use numbers not any prefixes. Claiming MOSNUM says it is "clearly permitted in the guideline" to use IEC prefixes in the secondary position is wrong. It does not allow that at all. If you don't believe me then ask someone at WP:MOSNUM. Oh wait, that has already been done, the conclusion is not to use IEC prefixes. Wanting to use IEC prefixes here also violates WP:COMMON because common sense tells anyone that using prefixes that are not familiar and "have seen little use by the press or the computing industry" is not good for any article. In summary Becksguy your claims about what WP:MOSNUM allows are wrong. 220.255.2.87 (talk) 06:58, 9 April 2011 (UTC)

I do not see anything on this page that you have written which is correct. It is obvious to anyone with knowledge of standard operating procedures here that IEC prefixes are not "banned", so this article is entitled to use them if there is a reason to do so. I assume (per AGF) that you are simply misunderstanding what a guideline is, despite its clear "generally followed" wording, but there is nothing at MOSNUM or its voluminous discussion pages which enforces usage here. Sure, it will be taken into account, and we get it: you don't want IEC prefixes. Indeed I would be speaking up against them too if anyone were suggesting that all the disk sizes be quoted in gibis or other gobbledygook, but that is not the topic of discussion. Johnuniq (talk) 07:48, 9 April 2011 (UTC)

I woke up this morning to a couple of emails asking to comment on this issue because I was one of the editors who worked on the original guideline text. As someone who worked on it for so long I have a long memory and unique knowledge of the intention behind the text which is all archived in the 16 or so binary prefix talk archives. It is correct to say IEC prefixes are in effect "banned" for this article. They are "generally not to be used" on Wikipedia and this also applies to the table that RaptorHunter wants to include in the article. The guideline clause which does allow IEC prefixes in an article does not apply to this article because it is not "specifically about or explicitly discussing the IEC prefixes". Accordingly the table should use "bytes or bits, with clear indication of whether in binary or decimal base". I prefer the style "64 × 1024² bytes" but any notation style can be used as long as IEC prefixes are not used. I have not read anything in this talk page that convinced me that there should be a special case for using IEC prefixes in this article. There is no value to using IEC prefixes in this article. To Becksguy and some others on this talk page who have made claims about the guideline: Speaking personally from the point of view of someone who put a lot of effort into the guideline over several years I find it alarming how so many incorrect conclusions can be made about the guideline text. At the time of writing we all thought it was very clear IEC prefixes are not to be used except in the most rare situations. Let me be clear, this article is not one of those rare situations so to claim the guideline allows IEC prefixes in this article is utterly wrong. Fnagaton 09:21, 9 April 2011 (UTC)

Where is the policy to back your opinions? I actually don't particularly care about the table under discussion, and in general I fully accept the idea that MOS is wonderful, but the policy misinformation and advocacy against this minor usage are irritating. Johnuniq (talk) 09:51, 9 April 2011 (UTC)

I'm glad you asked. One relevant policy is WP:CON which describes the way consensus is reached. Applying that to WP:MOSNUM it can be seen from the archives how many editors took part in the debate over several years to form the consensus. Sanctions can be applied to editors who refuse to accept consensus (WP:disruptive editing). Another relevant policy is WP:NPOV because as explained in the talk archive an editor's personal desire to use IEC prefixes (when they are very obviously not familiar to the average Wikipedia reader) is not a neutral point of view, it is a biased point of view. From reading this talk archive what I see are a couple of editors who have a desire to use IEC prefixes for the table when others have already proposed alternatives (power notation style) that are not biased towards either the old or new prefix systems. Obviously the least biased system is more aligned to the goals of WP:NPOV which is why power notation style should be used to replace the IEC prefixes. There are a couple of other policies that are relevant and these are also explained in the talk archives. The general thrust of the argument is that while the IEC prefixes are not widely used (and they are still not, this is a fact); Then any attempt to force them to be used in general articles (of which this is one example) against the WP:MOSNUM consensus is in violation of WP:NPOV policy and can be seen as disruptive editing behaviour. Fnagaton 12:10, 9 April 2011 (UTC)

As I explained above, the table does not violate WP:MOSNUM because it is explicitly discussing the prefixes. The consensus of editors on this talk page agrees with my interpertation of the rules. In fact it is you who are being disruptive by reverting against the consensus of this talk page.--RaptorHunter (talk) 14:48, 9 April 2011 (UTC)

Such an assertion is obviously not in the slightest bit true. The IEC prefixes aren’t being discussed, they’re being used to describe the capacity of hard drive capacities. Wikipedia always follows the practices of most reliable sources directed to a general-interest readership. Do we see “GiB” generally being used by computer manufacturers to describe hard drive capacities to their consumer base and other general-interest readerships? So why would Wikipedia do so then? Because we… are way cool and forward thinking? Your arguments are fallacious. You, RaptorHunter don’t have to concede on this point; we can simply have a nice quick RFC here to see if the IEC prefixes are “specifically about or explicitly discussed”as MOSNUM requires (as is done on Megabyte) or if it is the obvious reality: using them in an “Oh… didn’cha know?”-fashion to describe capacity.

You can save your wind arguing that up is down; a simple RFC here will settle the issue. Enjoy… Greg L (talk) 15:18, 9 April 2011 (UTC)

RaptorHunter as I explained above your interpretation of the rules is utterly wrong. I say this as someone with a lot of experience on MOSNUM and as someone who helped to write the rules that you mention. The consensus of editors on this page does not agree with you. I do not know how you can possibly think I am "being disruptive by reverting" because as of right now I have not recently reverted any changes on the article so I cannot be editing against consensus. Your edit was reverted not by me and with the comment "there is no consensus for your change; there are (at the very least) credible arguments against it". Obviously you are wrong and consensus is against your edit. Fnagaton 22:39, 9 April 2011 (UTC)

Caution - I strongly suggest that everyone stop reverting. It's a resurgence of the edit war and the article is going to get locked down again, probably longer this time, and maybe some editors will be blocked. Edit wars are disruptive, regardless. — Becksguy (talk) 15:24, 9 April 2011 (UTC)

My thoughts on the matter:

• Point 1: WP:MOSNUM allows the use of IEC prefixes where they themselves are, if not the primary topic of discussion, at least a significant topic. But this article is not about IEC prefixes. The section in question is not about IEC prefixes. It is about the meanings of true SI prefixes as used by hard drive makers, compared to identical-looking prefixes with binary meanings ("customary binary prefixes", per Binary prefixes) as used by operating systems and by many other programs to show hard drive capacities. It is attempting to explain the discrepancy between these two measurements. And neither of those measurements (at least as seen by the vast majority of computer users) uses IEC prefixes. There is therefore no justification for their use here under the "specifically about or explicitly discussing" provision.
• Point 1A: I agree that if the IEC prefixes were widely recognized, used, and understood, then they'd be appropriate to use in this context. Then they would be an aid to understanding the discrepancy. But that is not the case.
• Point 1B: And if the IEC prefixes were widely recognized, used, and understood, then MOSNUM wouldn't say what it says about them and we wouldn't be having this discussion at all. In fact, we might not have any apparent discrepancies to explain. But in the meantime, WP is supposed to document the status quo, not promote change.
• Point 2: The fact that IEC prefixes were added to this little table (or even the paragraphs surrounding the table) does not provide an "out" for the previous point. That would be self-referential, using the disputed usage as justification for itself. (And I think claims in that direction border on the ludicrous.) To satisfy MOSNUM's requirement of "specifically about or explicitly discussing" requires context other than the usage in question.
• Point 3: MOSNUM specifically calls out two other cases where IEC prefixes may be used: Where a majority of the references cited in the article use them, or (of course) when directly quoting a source that uses them. The usage under discussion here does not meet those requirements either. The "sources" here would be operating system displays and hard drive makers' marketing. Except for one Linux distribution (that has literally about 1% market share) and perhaps a double handful of applications and utilities, none of the above use IEC prefixes, so they cannot be used as a "source" that supports their usage here. So why include the IEC prefixes in the table? To do so does not help explain anything; rather it adds a point that needs further explanation!
• Point 3A: Mac OS is also not using IEC prefixes for displays of e.g. memory sizes, not that that would be relevant to this article anyway.
• Point 4: It is true that the lede at WP:MOSNUM allows for exceptions to the whole page, citing the "guideline" nature of the page as opposed to "policy." But I don't think that gives excuse to ignore the guideline just because some editor wants to. In my opinion, given the strong position at WP:MOSNUM against the general use of IEC prefixes, a compelling reason to use IEC prefixes must exist to properly invoke this very nonspecific "exceptions clause." Since the section can say everything it needs to say without using IEC prefixes, such reason does not exist here.
• Point 5: I still don't think "1000 GB" and "931 GB" should appear in the same table row to represent the same number, at least not as the table is (was) annotated. But that issue can be fixed other ways (better annotations). Fact is, HD boxes do say "1000 GB" and the OS will say "931 GB" for the same drive, and that needs to be explained. I think the answer is not binary prefixes, but better annotations. In fact, I'm creating a differently-formatted table to put in the Binary prefixes article specifically to address this issue.

Please note that I am not at all an opponent of the IEC prefixes. In the past, in fact, I argued against the adoption of MOSNUM's current stance on this topic. Since then, though, I have realized that MOSNUM's position is the correct one for Wikipedia, given IEC prefixes' almost complete lack of adoption in the real world. It is not the job of an encyclopedia to promote what some editors think are good ideas; it is the job of an encyclopedia to document information that exists in reliable sources. Yes, exceptions to MOSNUM are permitted, but such exceptions must have justification and consensus. I haven't seen any compelling justification for an exception here, and there clearly is a lack of consensus in that direction. The proper action is therefore to comply with MOSNUM. (p.s.: Yes, I opined otherwise previously. After further thought, I've concluded that I was wrong.) Jeh (talk) 20:34, 9 April 2011 (UTC)

RFC on the use of the IEC prefixes

With regard to the table here in “Hard disk drive”, is the use of the IEC binary measures and their symbols such as GiB “specifically about or explicitly discussing the IEC prefixes” as is required by MOSNUM:Quantities of bytes and bits, or are they being used to describe the magnitude of binary quantities? Is such use compliant or in violation? If you find it to be in violation of MOSNUM but to still be a sound technical writing practice worthy of exemption from MOSNUM’s guidelines, please elaborate. If you find this use here in this fashion to be compliant with MOSNUM, please elaborate.

• In violation This isn’t “explicitly discussing the IEC prefixes” as exemplified at Megabyte and Binary prefixes, and is simply using them to describe a binary quantity, which is specifically prohibited by MOSNUM precisely because terminology like “gibibyte (GiB)” isn’t used in the real world (computer manufacturers and computer magazines) when communicating to a general-interest readership. The overwhelming consensus of the Wikipedia community when the applicable MOSNUM guideline was debated (a three-month-long process) was that the IEC prefixes were unfamiliar to our readership. They still are. Since Wikipedia simply follows the practices of RSs, the use of these units in this fashion is poor technical writing practice and is—not surprisingly—entirely contrary to the guidelines of MOSNUM. Greg L (talk) 15:18, 9 April 2011 (UTC)
• Compliant

Per Wikipedia:MOSNUM#Quantities_of_bytes_and_bits: One of the exceptions for binary prefixes is:

" in articles specifically about or explicitly discussing the IEC prefixes. "

This table compares SI and binary prefixes. That counts as "explicitly discussing" it. Saying that 1000 TB = 931 TB is just confusing. For the purpose of this table, it makes much more sense to use binary prefixes for comparison purposes: 1000 TB = 931 TiB

It has even been suggested that we replace the words binary prefix with in the binary sense, which of course links back to binary prefix. This euphemism is nothing more than obtuse language to avoid talking about binary prefixes. It sounds like little more than a case of Wikipedia:I just don't like it This issue has been brought up again and again in the courts and the tech media. The table clarifies for our readers why a 1000 TB hard drive only has 931 TB. It would a shame for wikipedia not to cover it because of the bias of some editors.--RaptorHunter (talk) 15:38, 9 April 2011 (UTC)

• Deceptive

To simply state that binary quantities may be reported as a certain number of GiB is deceptive in that few sources (including operating systems) actually report hard disk capacities using that symbol, especially when one considers that consumers of such information will read the used or unused capacity of a drive much more frequently that the capacity of a new drive, and (so far as I know) operating systems do not use the IEC symbols to report used or unused capacity. Jc3s5h (talk) 16:14, 9 April 2011 (UTC)

• Compliant

The section of the article under dispute is all about the differences between the two usages of prefixes and therefore using the third prefix is in the interest of both avoiding confusion and informing the reader. Tom94022 (talk) 16:19, 9 April 2011 (UTC)

• Compliant

As it deals explicitly about the difference and confusion created by the ambiguously used prefixes. −Woodstone (talk) 16:37, 9 April 2011 (UTC)

• Or, we do as is done everywhere else on Wikipedia to address the dual meaning of “GB”: we write that in some contexts, such as the capacity of RAM, it means 1,073,741,824 bytes (1024³) and in others “GB” means 1,000,000,000 bytes. What is not done is to have Wikipedia expressing these quantities using terminology that is soundly being ignored by the computer industry; that much is part of Technical Writing 101. Our three-year-long experiment trying to be all futuristic and use the terminology here on Wikipedia in hopes that the idea would catch on and everyone will become familiar with the units of measure and the computer industry will follow suit was proven to be a pipe dream by idealistic, wide-eyed futurists. Get with the game plan please. You lost on this a few years ago and it’s a lost cause this time around too. Greg L (talk) 18:31, 9 April 2011 (UTC)

• Compliant

I agree with User:Woodstone. Martinvl (talk) 16:47, 9 April 2011 (UTC)

• Compliant but not appropriate

It adds to the confusion. In other words, I agree with User:Jc3s5h. — Arthur Rubin (talk) 17:00, 9 April 2011 (UTC)

Perhaps the table could be accompanied by a short paragraph to explain the difference.--RaptorHunter (talk) 17:19, 9 April 2011 (UTC)

Perhaps the IEC-advocates could just comply with the clear intent of MOSNUM, which was to relegate IEC prefixes to “articles specifically about or explicitly discussing the IEC prefixes” such as Megabyte and Binary prefixes and to not use them to describe the magnitude of binary quantities. Your argument that by using them to describe a binary quantity like this amounts to “directly discussing” the units is laughable wikilawyering and isn’t remotely compliant with the widespread will of the community on this issue. Greg L (talk) 18:02, 9 April 2011 (UTC)

Let me check I have this right: you want to introduce terms that aren't used to describe hard disk drives (which can only serve to confuse the average readership) and then introduce an explanatory paragraph to try and clear up the confusion? Sigh.  GFHandel.   21:05, 9 April 2011 (UTC)

So far, that’s Greg L objecting to the use here, as well as Jc3s5h thinking it’s deceptive to use them in the article. And Arthur Rubin agreeing with Jc3s5h and adding that it’s inappropriate to use them. We’ll let this run for a few more days. Greg L (talk) 20:00, 9 April 2011 (UTC)

• Not Compliant

...because I think "in violation" is too strong a wording for something that doesn't follow a guideline (as opposed to a policy). Nevertheless I do not feel the usage of IEC prefixes is justified here. Since an RFC should just contain brief responses I've put my long-winded arguments in the section above. Jeh (talk) 20:18, 9 April 2011 (UTC)

• Which was a thoughtful and logical post. Thanks. Greg L (talk) 21:06, 9 April 2011 (UTC)

• In violation (and not appropriate). The MOSNUM explicitly states "...decimal definition in an article on hard drives...". The IEC prefixes are not used in practice to describe the size of hard disk drives, and their use in this article can only serve to confuse the average readership. There are many links in the article (e.g. SI prefix, Gigabyte) that lead the interested reader to tables of conversions between the different units (and this article doesn't need to provide another one). For other reasons why using the IEC prefixes in this article is not a good idea, please see my post above.  GFHandel.   20:27, 9 April 2011 (UTC)

• In Violation. Per Greg L but with a few comments of my own. If one is truly interested in the difference between a megabyte and a mebibyte, or any other permutations of a byte, then there is more than a sufficient number of places that the reader can go to find it. They don't need it in their faces at every turn, especially when they can't go to their local tech store and ask the sales person there for a one Tibibyte hard drive.SteveB67 (talk) 21:36, 9 April 2011 (UTC)

LOL! “One tibibyte hard drive”. Indeed, Steve, they’d be laughed clean out of the store and they’d joke when the guy left about how me must have gone to Wikipedia to learn how to talk Klingon. Greg L (talk) 22:15, 9 April 2011 (UTC)

It's already "in their face", everytime someone plugs in their new 1TB harddrive and only sees 931 GB and then wonders why. Wikipedia is here to answer that question. --RaptorHunter (talk) 21:39, 9 April 2011 (UTC)

But a table that says "931 GiB" won't answer that question. Because that's not what they're seeing in their OS. Jeh (talk) 21:42, 9 April 2011 (UTC)

False. If there OS is ubuntu, it is what they will see.--RaptorHunter (talk) 21:54, 9 April 2011 (UTC)

That'd be less than 1%?  GFHandel.   22:05, 9 April 2011 (UTC)

Exactly. Ubuntu is written by space cadets, for space cadets. Those people already know this stuff anyway. This article is written for a general-interest readership, not for Wesley Crusher in the year 2350. With the exception of space-cadet OS, Jeh is 99 times more correct than RaptorHunter. All the major computer OSs don’t mention KiB, MiB, GiB, or TiB. They either typically say 1 TB* (* 1 TB = one terabyte = 1,000,000,000 bytes, actual formatted capacity is less) or they say Actual formatted capacity: 986,710,016 bytes. No computer manufacturer or hard drive manufacturer when communicating to a general-interest readership (consumer) is using the IEC prefixes. Arguments that “931 GB” is somehow explained by saying “It’s 931 GiB” is either galactic cluelessness over how to do technical writing or, more likely, grasping at logical straws in hopes no one here will notice. The actual way one would explain this supposed “931 dilemma” is along the lines of One terabyte equals 1,073,741,824 bytes (1024³ bytes) in some contexts (typically memory such as RAM), and in other contexts can mean 1,000,000,000 bytes (typically storage such as hard disk drives. It doesn’t really matter much except to Wikipedians who don’t have a life and argue on talk pages. All but that last sentence is clearly covered on MOSNUM. RaptorHunter should just go edit our Ubuntu article if (big “if”) the IEC prefixes are commonly used in that discipline by the RSs for that article. Greg L (talk) 22:15, 9 April 2011 (UTC)

Ubuntu? That's what you're hanging your argument on? The vast majority of users (Windows users) will see "931 GB." About 7% (current Mac OS X users... actually likely fewer than that, as many will not have upgraded to the newer versions) will see "1000 GB" and need no further explanation. Linux has less than 2% user share and Ubuntu is only about half of that... this doesn't constitute a significant "source" to be cited. I think we will also not be including a table entry quoted in "blocks" for the benefit of VMS users (fan of VMS though I am). Jeh (talk) 22:27, 9 April 2011 (UTC)

I started this RfC because debating this with RaptorHunter was like some sort of electronic feedback loop that could go on forever. I’ve seen that when editors insist “up” is “down” and “mentioning the IEC prefixes in a table amounts to directly discussing them”, there no point arguing. I’ve seen that when some editors don’t get their way in an AfD (article for deletion) and when it fails, they go tag-bomb the article, the best thing is to not argue or editwar over the tag. The best thing is just to (*sigh*) and do an RfC and when the results are in, revert to the community consensus. Sometimes the smart ones go on to something else. Others continue to editwar and get blocked or banned. But the only way these things are ever settled is by just having an RfC and be done with the nonsense.

I note that Fnagaton hasn’t even weighed in and we all know full well how he will !vote. Then the community consensus will be quite clear: MOSNUM correctly has the proper guideline and this stunt was an attempt at circuitous, self-referential logic to circumvent the obvious.

The only common ground RaptorHunter and I have appears to be that the F‑22 Raptor is an awesome, way-cool fighter. But that doesn’t make up for all this hassle-factor. Greg L (talk) 22:39, 9 April 2011 (UTC)

• In violation I have explained at great length why IEC prefixes should not be used. [6][7][8] I do think that if RaptorHunter continues to edit against consensus then administrative sanctions are appropriate. This RfC is useful to the extent that it clearly demonstrates WP:MOSNUM still has consensus and that the case against using IEC prefixes is still valid. I hope that those who were calling for IEC prefixes to be used look at this and decide to accept the consensus that IEC prefixes should not be used. Fnagaton 22:52, 9 April 2011 (UTC)

Actually Fnagaton can't vote since he was canvessed [9] via email. — Becksguy (talk) 23:02, 9 April 2011 (UTC)