Wikipedia talk:Manual of Style (dates and numbers)/Archive 104

From Wikipedia, the free encyclopedia
Jump to: navigation, search
Archive 100 Archive 102 Archive 103 Archive 104 Archive 105 Archive 106 Archive 110

Year ranges... spaced or unspaced? – or –?

Should the endash in date ranges be spaced or unspaced? It says unspaced in the "Dates" section of the MoS but both occur in the "Dates of Birth and Death" section (including the Darwin example which is the basically the example everybody actually looks at to see what the standard is!). Also is there a preference for the "–" character vs the control code "–" (–)? It seems to me that the control code is better. The MoS uses both so it seems like there's no policy. I think there should be one. Jason Quinn (talk) 01:34, 29 May 2008 (UTC)

En dash in ranges is unspaced, so I don't see why it should be different for year ranges. Either – or – are fine (just make sure the – is a – and not a hyphen - or an em dash — or a minus sign −) Headbomb (ταλκ · κοντριβς) 03:46, 29 May 2008 (UTC)

Then the MoS Dates and Numbers source about this needs to be edited to conform to it. A sentence should also be added to the "Dates of Birth and Death" section about the endash being unspaced because many people will not read the whole section and just the birth and death sub-section. I also propose that a new guideline explicitly states that the character and control code are both acceptable for the endash in date ranges. Jason Quinn (talk) 15:31, 29 May 2008 (UTC)
Either spacing is fine; either presentation is fine (one reason to use &endash; is to be sure you have the right dash; on the other hand, the character is easier to understand and edit in edit space); and any editor who goes about flipping them en masse should be warned not to be disruptive. Septentrionalis PMAnderson 16:44, 29 May 2008 (UTC)
Either spacing is not "fine": the guidelines are quite clear that en dashes representing "from -> to" are unspaced where both items themselves contain no internal spaces (1980–85), but are spaced where there's a space within one or both items (3 August 1980 – 13 June 1985). Otherwise, the visual effect is worse than untidy, and in the second case, possibly misleading. TONY (talk) 03:45, 1 June 2008 (UTC)
  • Tony is correct. What he is saying is perfectly consistent with the dominant manuals of style and is widely considered to be the correct way of doing things. Greg L (talk) 04:08, 1 June 2008 (UTC)

automatic archiving?

This page is now humungously big, and must be hell to navigate for anyone on a dialup connection. Dank55 has kindly added an archiving robot to MOS talk, and I've asked him whether he'll do the same for FLC talk, where they like the idea a lot. Does anyone object to this? At MOS talk, it automatically archives any section that hasn't been touched for 10 days. Is that the right duration for here? TONY (talk) 03:49, 1 June 2008 (UTC)

I thought MiszaBot_II already did archiving here? Talk page is unusually big because of the rewrite of section 4. I think it's about to be resolved so a lot of it should be archived soon. Headbomb (ταλκ · κοντριβς) 04:10, 1 June 2008 (UTC)
Tony: thanks for shortening the time to archive. I was unable to load the page at all earlier today, but it seems OK now.
To avoid a recurrence, it might be helpful to split up the Talk onto 4 separate pages (main page + one page for each of the 3 coloured boxes). That would certainly help me. Thunderbird2 (talk) 16:27, 1 June 2008 (UTC)

Change to GB ref

During May, "(decimal)" was added: "A typical advertisement for a PC in 2008 might specify 2 GB memory (binary) and a 160 GB HDD (decimal)." I'm just checking that it means something; I have no idea. I guess I'll add non-breaking spaces between the units and values. TONY (talk) 03:40, 1 June 2008 (UTC)

The binary means that the giga- is the binary sense, 230 (10243); the decimal that it's the decimal, 109 (10003). JIMp talk·cont 04:33, 2 June 2008 (UTC)
Thanks, Jim. Of course. I'm a dummy. TONY (talk) 08:28, 3 June 2008 (UTC)

Suffix of ordinal numbers

What is the reasoning behind the policy to not make the suffix of ordinal numbers superscript? As far as I can tell, styles like 23rd and 496th are the standard outside of Wikipedia. --Arctic Gnome (talkcontribs) 03:31, 2 June 2008 (UTC)

I looked in the archives and found this. Hope that helps.
I'm here because I started a discussion at Template talk:Th; I'm thinking about removing all the references to {{st}}, {{nd}}, {{rd}}, and {{th}}. —LOL (talk) 05:36, 2 June 2008 (UTC)
  1. Slightly harder to read.
  2. Extra work in the edit box.
  3. Shifts the line down a little, with untidy visual effect. TONY (talk) 08:20, 3 June 2008 (UTC)
  4. But the number is much more readable and it looks much better. ● 8~Hype @ 22:57, 3 June 2008 (UTC)

Since the only objection in the past 3¾ days is an WP:ILIKEIT-esque argument and there isn't any objective material to challenge the Columbia, Davidson and MIT style guides linked from the archive, shall I proceed with replacing all the ordinal suffix templates? For what it's worth, I also found the Apple Style Guide, University of Tennessee Editorial Style Guide, and BCIT writing style guide while browsing the first page of Google results for ordinal superscript style guide. CMSD reads "In Inside District News, superscript the ordinal endings", but "Inside District News" is probably not notable because of the sub-10 Google results. CMoS provides evidence of superscripts in ordinals; however, it seems to only apply to "monarchs and such" and Wikipedia does not use ordinal suffixes in sovereign titles (WP:NCNT#Sovereigns). —LOL (talk) 22:40, 5 June 2008 (UTC)

RFC discussion of User:Greg L

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

June of 2008 or June 2008

Which is preferred, "June of 2008", "June 2008", or either format, when providing a date with only a month and year? --Silver Edge (talk) 03:28, 6 June 2008 (UTC)

Wikipedia:Manual_of_Style_(dates_and_numbers)#Longer_periods says "June 2008". Gimmetrow 03:31, 6 June 2008 (UTC)
Thanks, I guess I missed that. Perhaps it should be added to the table at WP:DATE#Dates, beside "October, 1976". --Silver Edge (talk) 03:36, 6 June 2008 (UTC)

Curly quotation marks are not permitted by MOS

  • If anyone persists in using them in MOSNUM, I'll revert immediately. If you have a problem, go argue it out at the main page of MOS, which says: "The exclusive use of straight quotes and apostrophes is recommended. They are easier to type in reliably, and to edit. Mixed use interferes with searching (a search for Korsakoff's syndrome could fail to find Korsakoff’s syndrome and vice versa)".

At the moment, MOS requires words as words to be rendered in italics, not quotes. There's now a clash between this and the (disputed) point here that SI symbols must always be in roman face. This needs to be sorted out. TONY (talk) 04:47, 1 June 2008 (UTC)

  • Yes, you're right. The inconsistency between MOS and MOSNUM should be tackled. It is important for unit symbols to be upright though; in mathematics and numerical sciences, there is a long tradition of reserving italics for variable names. Suggestions anyone? Thunderbird2 (talk) 16:20, 1 June 2008 (UTC)
  • Headbomb: I see that you've italicised some of the unit symbols. How is that compatible with

In accordance with the rules of CGPM, NIST, National Physical Laboratory (UK), unit symbols are in upright, roman type.

and the (hidden) comment i.e. they are never italic; where they could be mistaken as symbols for dimensions, variables or constants? Thunderbird2 (talk) 18:04, 2 June 2008 (UTC)
  • They were italicized because they needed to be highlighted. They were between quotes before, but in some places quotes were cumbersome. The usage is similar to writing "A triquark is a particle made of three quarks". The italics on kg is compatible with MOS because it is not used as a symbol, but as a word. The only place I kept quotes was in the point about italics, so people could distinguish what to do from what not do to. Headbomb (ταλκ · κοντριβς) 05:51, 3 June 2008 (UTC)
  • Headbomb, I don't understand your explanation. We need to think this through more carefully, and attempt to address Tony's concern about consistency with MOS. Thunderbird2 (talk) 18:17, 3 June 2008 (UTC)
  • I don't think I give a shit about CGPM, NIST and the NPL. We've had this argument before, and it was unresolved. I go with Noetica's view that there's no good reason to make some god-given exception to our usage of italics everywhere else on WP (see MOS on italics), just because a few outside authorities, without stated reason, want SI units to be exclusively roman. I used quotes in my recent clean-up of MOSNUM in deference to the current rule, which was inserted, I believe, without proper consensus. You can see how ugly and hard to read they are, and in the case of degree symbols (the little superscript circles), quotes make it almost impossible to see the symbol. We have to mark "words as words" some way, and italics seems the obvious solution, as prescribed by MOS. TONY (talk) 08:25, 3 June 2008 (UTC)
  • Tony, there is a reason for it, which I gave in my first post just above. Whether the reason is considered good enough is another matter. In what sense was the change here made without consensus? Thunderbird2 (talk) 18:13, 3 June 2008 (UTC)
  • Curly (typographers) quotes look much better than straight quotes; that’s why they’re also called “typographers’ quotes.” All professionally produced literature use them. They are not routinely used on Wikipedia only because the Windows operating system makes it so cumbersome to employ them. Further, many volunteer editors to Wikipedia wouldn’t recognize the difference between an hyphen and an emdash so the wise thing to do is to just let them pound the ol’ " key. So it makes sense to permit the use of straight quotes on Wikipedia, especially in articles that are currently in a great state of flux (either revisions or expansion). So…

    I would propose that MOS and MOSNUM policy be harmonized and also updated with the additional caveat that when an article has reached a level of completeness and polish that it is undergoing little in the way of substantial rewrite or addition, that typographers’ quotes are permissible. This will keep Wikipedia easy to edit when articles are in a state of growth and flux, but will also put Wikipedia on the slow track towards looking more like a professional-grade publication. Greg L (talk) 18:47, 7 June 2008 (UTC)

Superscript characters

Both the Basic Manual of Style and the Mathematical Manual of Style recommend that proper superscripts be used for exponents (<sup>2</sup>, <sup>3</sup>) instead of mixing in the legacy Unicode characters ² and ³. This is for consistency (see 1²³4 vs. 1234), for accessibility (the Unicode superscripts are impossible to read on some monitors at the default size), and because the Unicode standard recommends they not be used when markup is available.

For consistency, this guide should reiterate what the other guides state, that the two characters ² and ³ should not be used in articles except when necessary (such as in templates or links that require them). —Werson (talk) 04:44, 5 June 2008 (UTC)

Actually I believe that <math>2^{32}</math> (rendered as 2^{32}) is the preferred method when dealing with mathematical expressions. -- KelleyCook (talk) 13:22, 6 June 2008 (UTC)
You're right, that would be more correct. Problem is I'm seeing a lot of articles using the legacy characters as shorthand, especially for units (see Area, for instance). —Werson (talk) 21:39, 6 June 2008 (UTC)
  • It wasn’t too long ago that full-tilt superscripts on Wikipedia caused extra leading on the line that included them. It resulted in ugly page layout and, where there was lots of lines using them in a short paragraph, you could hardly see when one paragraph ended and the next one began. I hadn’t noticed exactly when, but recently, Wikipedia fixed this problem so now leading doesn’t shift around like that. IMO, that’s one reason why the Unicode mini-superscripts were popular (besides the fact that they used to be included in the Insert:Symbol pallet). I see that the m² and m³ characters are still there for the choosing and I see no reason why they shouldn't be permitted. But for a consistent, harmonious appearance in articles, I would think the proper thing to do is use full-tilt number<sup>2</sup> superscripts. Greg L (talk) 05:19, 8 June 2008 (UTC)

Archiving of the rewrite

I'm currently archiving the rewrite. Please give me an hour to do so. I will post here again when I'm done. Headbomb (ταλκ · κοντριβς) 17:56, 7 June 2008 (UTC)

Everything was archived at Wikipedia talk:Manual of Style (dates and numbers)/Archive/Complete rewrite of Units of Measurements (June 2008). I understand that usually only old threads are archived, but there was just so much that can't be separated that I couldn't only archive half of it coherently. I've left the votes here for now so Woodstone can update his vote. I'll update the archive accordingly.

You can head over there to check if your links still work.

Discussion is not closed, if you want to talk about anything from there, just bring back whatever text is relevant as needed. Headbomb (ταλκ · κοντριβς) 17:56, 7 June 2008 (UTC)

Archive completed. Headbomb (ταλκ · κοντριβς) 19:31, 7 June 2008 (UTC)

Archived part

I've put archive tags around it. This was a rather old part of the discussion, wasn't it? Propose to make a fresh start, if someone feels like reactivating that part of the discussion. --Francis Schonken (talk) 21:58, 7 June 2008 (UTC)

  • I can’t find this in the archives. Where is it? And if it is in an archive, why is it here again? Greg L (talk) 23:15, 7 June 2008 (UTC)
    • Don't know. I neither took it away from this page, neither brought it back. I only put the archive tags around it. --Francis Schonken (talk) 06:35, 8 June 2008 (UTC)

Sequence of events:

Thunderbird2 (talk) 07:33, 8 June 2008 (UTC)

  • I archive it to /Archive 102 leaving a transclusion inside the discussion box ... and stick it in a lovely lime show-hide box ... hope this is okay for everyone. JIMp talk·cont 01:49, 9 June 2008 (UTC)

If it's archived, then it's archived. No need to have it here too. Headbomb (ταλκ · κοντριβς) 02:28, 9 June 2008 (UTC)

So be it. JIMp talk·cont 02:45, 9 June 2008 (UTC)

Discussion of events leading up to 7 June 2008

  • Well, what can I say? One good thing has come of all this nonsense: at least I can edit the page. I will take advantage of my new-found ability to load it into my browser, and make my views known for the record:
  • No attempt has been made to address my and Woodstone's concern about unjustified deprecation of IEC units, despite a vote as recently as March 2008 showing 11 votes to 0 against precisely this form of deprecation. (the record shows 11-1, but only a sockpuppet favoured deprecation)
  • During the final voting Throughout the entire process the wording of the oppose vote was biased; and when Woodstone attempted to remove the bias, his changes were reverted by Fnagaton and by Headbomb. Headbomb "permitted" Woodstone's neutral version only after he had uploaded the new text.
  • No attempt was made to deal with the sheer size of the page to make it easier to load it. I was unable to access the page for most of the afternoon; there may have been other editors in the same position.
  • I have been the target of repeated accusations of dishonesty [1][2][3][4] by Fnagaton during the past few days.

This is not how consensus is built. Thunderbird2 (talk) 19:35, 7 June 2008 (UTC)

    • I take back my remark about the wording being biased. I do so because that remark was based on a misunderstanding that oppose votes were to be treated differently to support votes (The way the table was laid out gave me the incorrect impression that only oppose votes were to be ignored). Nevertheless, it is unusual to have such a category of votes in the first place. It is even more unusual for a voter to have to fight to get his vote accepted in the ‘included’ category. Woodstone’s vote makes it clear that he wished it to be considered (and why wouldn’t he?), but for reasons I have not investigated he had to reinstate it not once[5], not twice [6], but five times in total [7][8][9]. Why did he need to do that? Thunderbird2 (talk) 19:39, 8 June 2008 (UTC)
  • Every attempt was made to address valid concerns that had substantive reasons, you would be correct to conclude that what you posted were not substantiated reasons but instead they were statements without valid argument, so they were disregarded. Headbomb and others asked you many times over the past week to provide substantive arguments, you did not. The individual vote you cite does not represent consensus because consensus is made with good arguments, not by the number of people who vote on one issue. The oppose vote wording is not biased at all. There were many votes, so it is disingenuous to only pick one and claim that means there is no consensus. You helped cause the large size of the page by copying large amounts of text from talk archives. Also you seem to have edited anyway so it is a non-issue and would not have affected consensus anyway. Again I see you repeat your utterly untrue and false accusations because you used the word dishonest in relation to yourself first, don't try to deny it because the archive is proof that you did. The fact you keep on misrepresenting and repeating that completely untrue accusation means you intend it to be a personal attack, stop, right now. Consensus is built by good arguments, you have failed to answer questions and you have refused point blank to answer questions, you only have yourself to blame for removing yourself from the debate by refusing to make good valid arguments. If you wanted to help with the consensus then all you had to do was contribute in a meaningful way, but all you did was claim "there isn't consensus" (despite the proof showing there is consensus) while refusing to explain why, which is nonsense. Just stop trying to misrepresent the truth, it is archived and it proves where your version of events is completely untrue. Please, for your own self respect, before you lose what you have remaining, just stop trying to re-write the truth to fit your personal opinion. It won't work and it just makes you look bad. Fnagaton 19:49, 7 June 2008 (UTC)
  • Response:
  • First, there were plenty of attempts to address your concerns. It is you who ignored our repeated attempts to have you clarify your position, and to have you explicit the reasons of your concerns. It is only very recently that you first gave us one reason (on my talk page), and that reason was that there was a vote 2-3 months ago opposing the deprecation. We went through the archives, and found no reason other given for opposition than "I like it IEC units" and "Some organisation uses them". The fact that some organisation uses IEC prefixes is made completely irrelevant in the face of the fact that virtually no organization uses of them (Google test said less than 0.1% of scientific literature uses them, and about 0.6% of "general" litterature). See also WP:IDONTLIKE, WP:UNDUE, WP:Crystal Ball. As if we go by votes alone, what was 11 vs. 0 against deprecation has turned into 10 vs 2 (or 11 vs. 3, depending on how you count) so if your opposition is that 2 months ago people were against deprecation, this crumbled in the face of the fact that now people are for deprecation.
  • Second, how was the wording biased? Options were consensus/no consensus, support/opposition for personal reasons. Reasons were completely symmetric for either sides. I lumped Woodstone's vote into personal reasons because guess what... it was based on personal reasons and still is. My reverts were more concerned about maintaining the options as they were. I "permitted" the vote so Woodstone would stop complaining about it and so I could archive everything, but it is still fundamentally a personal reason vote. That this happened after the upload is inconsequential, as it was not a swing vote at 10 vs. 2, and the opposition's reasons are just as weak as before which is the fundamental reason why this was uploaded. If he were able to separate his personals feelings on the issue from your ability to judge whether or not there is consensus none of this would have happened.
  • Third, most of the problem of you not being able to see the page was caused by your humongous upload of texts from the archives rather than simple links on the talk page so you only have yourself to blame for that.
  • And fourth, it's completely inconsequential if you feel your pride is more important that settling issues. I've had my loads of personal attacks from just about everyone here, and you didn't see me get on a high horse and play the "If you don't apologize, then I'm not going to address the concerns of people" game. If you have a problem with that, then RfC. I've accused both Greg and Fganaton of being sockmasters [Which I’ve never done. Greg L (talk) 20:48, 7 June 2008 (UTC)], I've berated Greg L for his conduct here [“That is twenty demerit points Sgt. John Spartan.Greg L (talk) 20:48, 7 June 2008 (UTC)], I've opposed Jimp's argument in the previous FCL greenbox, and I still think that Fganaton was the sockmaster of DPH, but did that stop any of them from working with me and answer my concerns? If we can all rise above petty bickering, then so should you. Headbomb (ταλκ · κοντριβς) 20:12, 7 June 2008 (UTC)
  • Re the last point about accusations of dishonesty: TB2's untrue accusations have already been added to the RfC by myself and my comments have been endorsed by at least two other editors.Fnagaton 20:17, 7 June 2008 (UTC)
  • Thunderbird2: Your above claims are beyond fallacious. All of them. …[use my] new-found ability to load it into my browser, and make my views known for the record… Who do you think you’re trying to kid here?? You were as active here on Talk:MOSNUM over the last several months as anyone else. If your views were ambiguous, it was only due to your own fault. Why? Because, it seems to me that one of your debate tactics at first was to give the appearance of someone sitting on the fence on this issue, trying to influence opinion by appearing to be someone who was wrestling with the issue of the binary prefixes. Your ham-handed effort to disambiguate Mac Pro without using your precious IEC prefixes, and your complaint here about how your success in your efforts to write articles without the IEC prefixes were “spotty”, only reinforced this view of your method of operation. But in the last month, your position had become clear as glass. We all heard your arguments. So please spare us your “No one is listening and responding to my arguments” spiel. We have listened. And we understood. And we responded (hundreds, if not thousands of times). We simply reject your position as being the unwise thing to do. What Wikipedia will be doing now will be to follow the way the rest of the computing world works. That’s good. As for your allegation that this was not how consensus is built, Headbomb’s effort should serve as the paradigm of how consensus is built. It comes as no surprise to anyone here that a principle opponent of the consensus view now alleges that there was no consensus. We reject that charge as being fallacious too. Desist please. Greg L (talk) 20:17, 7 June 2008 (UTC)
  • Anyway, now TB2's outburst has been roundly refuted I have to go to sleep. Early flight in the morning. :( Fnagaton 20:40, 7 June 2008 (UTC)

Scientific notation, engineering notation, and uncertainty's location

The section Scientific notation, engineering notation, and uncertainty could be better positioned. This section deals with the representation of numbers. Such notation can occur outside the context of measurements and is certainly something besides units. I propose moving it from Units of measurements. It could stand alone as its own section but what I propose is to move it into Numbers since that's what it deals with. JIMp talk·cont 01:26, 9 June 2008 (UTC)


You write "Makes sense. Feel free to edit." did you mean to delete it or was that a mistake? I will edit. There is no content change involved just a more logical organisation of content. It had been brought up before but I can't find it in the archives and I'm not really up for picking through diffs.

JIMp talk·cont 03:10, 9 June 2008 (UTC)

found JIMp talk·cont 03:19, 9 June 2008 (UTC)

I'm not quite sure I'm following. I deleted the collapse box containing the content of archive 102, since it's archived. What's that got to do with the moving of the Sci/eng notation section? Headbomb (ταλκ · κοντριβς) 04:04, 9 June 2008 (UTC)

Should've been nothing, or so you'd guess, but see the diff. JIMp talk·cont 04:06, 9 June 2008 (UTC)

Autoformatting date ranges

Well, which is right? I prefer to use wikidates, but for date ranges that means I have to give two full wikidates, because otherwise, as the second reference notes, these will be damaged. Using wikidates is the only way that readers with preferences set will see the date range correctly, regardless of whether they are using International Dating or American Dating format. --Pete (talk) 23:15, 2 June 2008 (UTC)

Yep & it doesn't look like getting fixed any time this decade. JIMp talk·cont 23:37, 2 June 2008 (UTC)
If you want to say September 21–29, 2002, you don't use date linking. If you want to use date linking, you have to say September 21September 29 or 21 September 200229 September 2002. Are you implying the two statements above are inconsistent? Gimmetrow 04:23, 3 June 2008 (UTC)
Yes indeed. So why use the autolemon for any date? We've tried and tried and tried to get the Wikimedia developers to decouple it from the linking mechanism, but it falls on deaf ears. Why people are quite happy to read the spelling of other varieties of English, but feel the need to put ugly bright-blue splotches on full dates just for some comforting frisson attached to seeing dates in the format of their own variety, is beyond me. North Americans easily put up with the non-American date format in every signature. I have no problem with August 26, 1979. Big deal. Just as well the autolemon is not mandatory; I discourage everyone from using it. TONY (talk) 08:18, 3 June 2008 (UTC)
This system was put in place to stop one of the earliest edit wars on wikipedia - years before you joined, Tony. You have not gained consensus for your position so please stop. Bug the developers instead. Rmhermen (talk) 18:21, 3 June 2008 (UTC)
Pete wants to make the preferences work. I want date ranges to look like normal text, instead of being big blue distractions. I say screw the preferences; if you're talking about dates that a show aired on U.S. television, use U.S. format. If you're talking about Margaret Thatcher, use European format.Cstaffa (talk) 14:36, 3 June 2008 (UTC)
Sorry, I should read more thoroughly. I should have just quoted:
Strong national ties to a topic
Articles on topics with strong ties to a particular English-speaking nation should generally use the more common date format for that nation; articles related to Canada may use either format consistently. Articles related to other countries that commonly use one of the two acceptable guidelines above should use that format. Cstaffa (talk) 23:15, 3 June 2008 (UTC)
Hmmm. There's only 18 months left before 2010. Place your bets. Of course by then we'll have another three million articles to disambiguate messy dates in. Suggest you file a bug to allow preferences to be set for plain, non-highlighted presentation of dates even if datelinked in the wikisource. I'd even suggest that for anon readers the default pref be that. Myself, I still want to see maintainable, unambiguous wikisource, so I'll continue to prefer ISO 8601 dates wherever they can be used (especially in citations). Is today 05/06/08 or 06/05/08? I'll just call it 2008-06-05 so nobody can be confused. LeadSongDog (talk) 06:26, 5 June 2008 (UTC)
It wouldn't be too difficult to have linked dates autoformat, but still end up without links in the rendered page. I think that would make Tony1 happy (no random blue dates) without affecting how autoformatting resolved the old date format controversy. Except for the problem of determining a way to force links when specifically desired, is there a strong objection to this sort of solution? Gimmetrow 03:26, 9 June 2008 (UTC)
I know that writing this here doesn't help, but I just want to say: it doesn't seem like it should be so complicated to have a parser turn [[2009-01-04]]–[[2009-02-15]] to 4 January – 15 February 2009, but turn [[2009-01-04]]–[[2009-01-09]] to 4–9 January 2009. It's pretty basic; I can't believe there's so much inertia to add slightly clever functionality to MediaWiki. —Werson (talk) 02:43, 11 June 2008 (UTC)

Archive B9

  • I've seen some people argue using the results from the poll in Archive B9, and I've seen other people argue that it's no longer applicable because significant numbers of people have changed their minds. Instead of arguing about it, I'd like to de-archive B9 into a page like Wikipedia talk:Manual of Style (dates and numbers)/Binary prefix opinions and let people change their votes if they've changed their minds, add new sections for consideration, etc. with no deadline or archival. I think this would be a valuable tool for building consensus and representing editors' (possibly changing) opinions on each issue without being lost in a lot of argumentation.

    My suggestion is to move the polls alone to the new name, and leave just the discussion in the B9 archive with a link back to it. What do you all think? — Omegatron (talk) 23:26, 7 June 2008 (UTC)

  • Let old polls be old polls. That one was made two months ago, in the context of two months ago. To change it is meaningless, as probably more than half of these people won't update their votes and we'll be stuck with an ambiguous results and votes from two time periods.

    Start a new one if you want, but that would be pretty pointless considering we just had one, and that polls mean diddly squat on their own. See WP:DEMO. Headbomb (ταλκ · κοντριβς)

  • Omegatron: The events and votes archived on B9 are not applicable in any way to what later occurred here and which eventually lead to the current contents of MOSNUM.

    Please respect the fact that many editors labored here in good faith to debate the proposed contents and develop a consensus on the current MOSNUM wording. You apparently had no interest in the goings on here, or boycotted the proceedings. But for whatever reason, you chose not to participate. The majority of editors who labored on the latest wording have zero interest whatsoever in your trying to resurrect the IEC prefix issue for further discussion. Please respect the consensus. Do all protons in the universe have to decay before you let this rest? If you dispute that there was a consensus in the latest decision, please take your complaint somewhere else. Greg L (talk) 04:52, 8 June 2008 (UTC)

  • Omegatron's suggestion is a good one. The two polls are in no way comparable, because B9 was solely about binary prefixes, whereas the new one was much broader, covering all units and their symbols. Such a poll would settle the issue of consensus on the disputed part of the new guideline. Isn't that what we are trying to achieve here? Thunderbird2 (talk) 07:43, 8 June 2008 (UTC)
  • Omegatron's suggestion is a bad suggestion because it goes against a couple of the policies he cited in the RfC complaint about Greg, notably WP:CONSENSUS because votes don't make consensus good reasons do and because when challenged no good reasons have been given, Wikipedia is not a democracy because trying to resurrect votes when consensus has just been reached is not how Wikipedia operates and Wikipedia:Polling is not a substitute for discussion. Just so there is no doubt Omegatron, it is not a good idea. Consensus has been reached so calling for yet more votes will be seen as an attempt to game the system. You had months to make a valid contribution yet you did not. Fnagaton 15:14, 8 June 2008 (UTC)
  • Francis Schonken's horse is alive and kicking:

    Because of the clear deprecation that was considered so important by some, the new guideline is being interpreted as preferring ambiguous units to unambiguous ones. User:Warren may or may not be the first editor to interpret the guideline in this way, but I am certain he will not be the last. The solution is simple: Remove the deprecation of IEC units. Thunderbird2 (talk) 14:56, 8 June 2008 (UTC)

  • No TB2, your edit does not follow the guideline, don't try to get Warren into trouble because reading the article it is obvious what the prefixes already mean due to their context. If he doesn't want to add disambiguation then OK, so if you think disambiguation is needed you should be adding disambiguation using footnotes to state the number of bytes with (power notation is good), or by adding the number of bytes to the prefixes used (power notation is good) or by adding some explanation following "Editors should specify if the binary or decimal meanings of K, M, G,... are intended as the primary meaning" in the article. You should be choosing one of those options instead of adding IEC prefixes. The guideline is very clear about this. If you don't want to add footnotes or an acceptable form of disambiguation then don't, leave it to another editor to do instead. As Greg showed in another article it isn't hard or difficult to add disambiguation that does not use IEC prefixes, so please follow the new guideline text in future. Fnagaton 15:21, 8 June 2008 (UTC)
  • Who is blaming Warren? Not me, because he has done nothing wrong. The problem is clearly with the guideline, which leaves itself open to be interpreted in this way. (And no, the meaning is certainly not obvious from the context to a non-expert). Thunderbird2 (talk) 15:37, 8 June 2008 (UTC)
  • I'm glad you say he has done nothing wrong, so please stop reverting his changes. The problem is not with the guideline if you follow the guideline and disambiguate with the examples given in the guideline. So, one question directly to you: Why are you not following the guideline ( by adding IEC prefixes) when you should be adding acceptable disambiguation instead? Fnagaton 15:47, 8 June 2008 (UTC)
  • The guideline says "Editors should use the conventional prefixes, such as kilobyte (KB) and megabyte (MB), and disambiguate where necessary." (i.e. it is not mandatory to disambiguate) and "IEC prefixes are not to be used on Wikipedia except under the following circumstances". Looking at Warren's changes, where you say he has done nothing wrong, he has correctly removed the IEC prefixes, within the scope of the guideline. He obviously doesn't think disambiguation is necessary which is also within the scope of the guideline and you did say he has done nothing wrong. Obviously from your comments you do think disambiguation is necessary, so within the scope of the guideline you should be adding acceptable disambiguation, not adding IEC prefixes back again. Fnagaton 16:09, 8 June 2008 (UTC)
  • Here is Warren's interpretation of the the new guideline:[10]
  • [computing articles] should carry the ambiguous binary prefixes unless there is a very specific reason to do otherwise
  • If that interpretation is widespread we will have more and more ambiguity introduced into computer articles. Is that the desired end result? Thunderbird2 (talk) 18:40, 8 June 2008 (UTC)
  • The guideline says "Consensus was reached that the spirit of the MoS was better reflected by having familiar but ambiguous units ... ". The guideline does not result in more ambiguity being added to computer articles if you follow the guideline and use the acceptable forms of disambiguation where necessary. I note you've not answered the question put directly to you above. Fnagaton 18:53, 8 June 2008 (UTC)
That's because I hadn't noticed it. I simply reverted a change that did not follow the spirit of MOSNUM. I did not introduce the IEC prefixes because they were there already. Thunderbird2 (talk) 21:06, 8 June 2008 (UTC)
  • As already explained above and on several talk pages, Warren's change did not go against the "spirit of MOSNUM". Your edit did go against the "spirit of MOSNUM" and did go against the letter of MOSNUM as well. You made an edit to add IEC prefixes back again and you know better than that because you know exactly what the guideline text says about using IEC prefixes. I also note, again, that you have failed to answer the question put directly to you. Fnagaton 21:21, 8 June 2008 (UTC)


  • You are asking the wrong question. You should ask yourself: which version is better: the ambiguous one or the unambiguous one (which does not follow the guideline to the letter). The answer must certainly be: the unambiguous one is better. If someone else wants to further "improve" by following the guideline that is up to him. A reversion to ambiguous units is counterproductive. −Woodstone (talk) 09:41, 9 June 2008 (UTC)
  • I'm asking a good question and he's refusing the answer it. It isn't counterproductive because it uses prefixes that the reader will be familiar with, as it says in the guideline text "Familiar units are preferred over obscure units" and "Consensus was reached that the spirit of the MoS was better reflected by having familiar but ambiguous units...". So to answer your question the better version is the version that uses prefixes that the reader will be more familiar with. Therefore, the better version is the version that Warren edited. The corollary is that the version edited by TB2 is worse than Warren's version, again this is according to the consensus in the guideline. By the way, you are asking the wrong question because you assume the article is "ambiguous" or "unambiguous" which actually isn't the case because the context of the KB/MB/GB prefixes used in the article makes it obvious, so it isn't ambiguous per se. The article might be improved by adding disambiguation that uses the methods described in the guideline if the editor thinks it is necessary to disambiguate, however the article is made worse by adding IEC prefixes. Hence my question to TB2 that has still gone unanswered. Looking at Warren's edits and the articles it is clear to me he thinks (I have to assume good faith here) from the context in the articles what the prefixes mean so the disambiguation is not necessary, look in the MOSNUM guideline text for "necessary". Now then, it is clear looking at TB2's edits he thinks that disambiguation is "necessary". But looking at TB2's edits he uses IEC prefixes which according to the guideline are "not to be used on Wikipedia" and TB2 knows this fact. What TB2 should have done is to disambiguate using acceptable methods described in the guideline, not the wholesale reversion and comments on multiple talk pages of Warren for his edits. You see, in the time it has taken TB2 to make those multiple talk page comments and the reverts he could have easily added acceptable disambiguation and helped to improve the article. Doing that would have involved a lot more good faith and much less Wiki-drama. Editors must not disrupt Wikipedia to illustrate a point. Fnagaton 10:07, 9 June 2008 (UTC)
  • By changing MiB to MB information is destroyed. A bot could easily be constructed to change 1 MiB into 1 MB (10242 bytes). However there is no bot that could ever disambiguate 1 MB dependably. So I stay with the conclusion that MiB is better than an ambiguous MB. You forget that a guideline is not mandatory. It just describes the preferred end result. Not all steps towards it need to be taken by the same editor. −Woodstone (talk) 12:11, 9 June 2008 (UTC)
  • Information is not "destroyed" by changing MiB to MB because the surrounding text in an article gives context. The guideline clearly says MB is better than MiB because MB is familiar to the reader. The end result is that the terms used in the article are familiar to the reader because we want the article to communicate to the reader and not be too confusing, hence the guideline text using that terminology. Fnagaton 13:32, 9 June 2008 (UTC)
  • So you're saying that if we put MB (220) instead of MiB, everything is peachy—just like the MOS says? And you're saying that with or without IEC prefix around, we'll still have to disambiguate the decimal MB (106), because in the real world, MB can mean either 106 or 220—just like the MOS says?

    Glad to see that you finally agree that we have a sound policy that allows us to write things unambiguously in a way that 99%+ of the planet can understand. Headbomb (ταλκ · κοντριβς) 13:19, 9 June 2008 (UTC)
  • To Fnagaton: yes destroyed, because it becomes inaccessible to a bot and can only be reconstructed by the efforts of a knowledgeable human editor. To Headbomb: I'm very surprised by that last remark. I have stated quite while ago that I agree that MB should always be disambiguated and that I have no objection to a preference for explicit powers (although personally I would prefer IEC). The only thing I still object to in this context is the explicit ban on IEC prefixes (the bullet "the IEC prefixes are not to be used" with its 3 sub-bullets). It is unnecessary, because application of the preference would phase them out. At a moment shortly before the final voting I thought we had agreement on that. There is no explicit ban on unfamiliar units like "fathom" or "cubit" and a 100 others. Why one for MiB? −Woodstone (talk) 13:55, 9 June 2008 (UTC)
  • No not destroyed at all, as already explained the context makes it clear. The advantage is that the reader sees prefixes that they are much more likely to understand. The guideline text states this preference as well, that is the consensus. Also as we have just seen we do need the explicit ban because some editors will still try to add IEC prefixes to articles when they know full well that other better methods exist and that IEC prefixes are not preferred, again that is the consensus. Not that it is really a valid point about the "fathom" or "cubit", it is a red herring, but with those units we don't have a minority of editors intent on pushing their point of view about those units into articles. We do have a minority intent on pushing their point of view with IEC prefixes, yet another reason for the explicit ban. Of course if after a few months (six months maybe?) we did not have a minority intent on pushing their point of view then I would be open to the suggestion of potentially removing the explicit ban, but until then I say we do need it. Just as the consensus showed. Fnagaton 14:04, 9 June 2008 (UTC)
  • There are no explicit ban on unfamiliar units like "fathom" or "cubit" because there is no de facto need for such a ban. When you see a fathom/cubit brigade who insist on having fathoms and cubits all around Wikipedia, then you'll see an explicit ban on fathoms and cubits. Or perhaps it's just that the fathom/cubit brigade members are reasonable enough to see that "Familiar units are preferred over unfamiliar units", conclude that meters and cubic meter are indeed more appropriate. There's no de facto need for an explicit fathom/cubit deprecation because the implicit one is respected. The IEC-brigade members however, have and will-continue to insist that IEC units are acceptable since they aren't explicitly banned. Hence the ban. Headbomb (ταλκ · κοντριβς) 15:15, 9 June 2008 (UTC)
  • Exactly. At the end of the day what the guideline does is make the articles less confusing and better understood for the readers who will be using them. Fnagaton 15:18, 9 June 2008 (UTC)
  • To Woodstone: 100% behind you
  • To Fnagaton, Headbomb: the ban is what gave Warren the licence to interpret the guideline as favouring ambiguous units. Therefore it does more harm than good.
  • Again no TB2. Warren's changes are fine because they follow the consensus by using units that readers will find more familiar and will be understood more than IEC prefixes, this is the consensus and it improves the articles. You yourself said Warren did nothing wrong. The only edits that "did more harm than good" (to use your phrase) were your edits because you reverted perfectly good edits and you did not follow the guideline. Read Wikipedia:Do not disrupt Wikipedia to illustrate a point. I again note you failed to answer the question put directly to you. Since you failed to answer that question I have another for you: When are you going to start following the guideline that has consensus? Fnagaton 17:06, 9 June 2008 (UTC)
  • The spirit and letter of the guideline is to use unambiguous units. When have I not done that? And what other question are you going on about? Thunderbird2 (talk) 17:31, 9 June 2008 (UTC)
  • No. Here is what the guideline says "Consensus was reached that the spirit of the MoS was better reflected by having familiar but ambiguous units...". The guideline actually prefers what are called "ambiguous units". And here is what the guideline says about IEC Prefixes "The IEC prefixes are not to be used on Wikipedia...". You did put IEC prefixes in your edits, which is directly opposed to the spirit and the letter of the guideline. So again: When are you going to start following the guideline that has consensus? Fnagaton 17:49, 9 June 2008 (UTC)
  • Fnagaton, debating with you takes considerable patience. I will explain one more time. My edit did not introduce the IEC prefixes, because they were there already. I simply reverted a change that introduced ambiguous units, by an editor who failed to appreciate the need for unambiguous ones. It is interesting to note that you prefer the ambiguous version. Thunderbird2 (talk) 18:04, 9 June 2008 (UTC)


  • I didn't say you "introduced IEC prefixes" I actually wrote "You did put IEC prefixes in your edits" which is different to what you just wrote, so when are you going to retract what you just wrote because it is false? What you did was was opposite to the spirit and the letter of the guideline for the reasons already given and explained to you by multiple editors here and on those talk pages you went to. The reasons for this are long but the short version is that familiar units are preferable to the unfamiliar ones. It is interesting to note that you again misrepresent ("It is interesting to note that you prefer the ambiguous version") what I actually wrote because I actually wrote "The guideline actually prefers what are called "ambiguous units" ". My personal feelings are completely irrelevant when pointing out your mistakes regarding application of MOSNUM guidelines, so do not try to make it personal with your misrepresentation. Trying to make it personal won’t work and all it does is make your point of view look even weaker than it is now. Also as already explained Warren's edits do not make the article ambiguous because of the context in the article and also as explained the guideline says "disambiguate if necessary". You did earlier write "[Warren] has done nothing wrong" so you cannot now try to twist the blame onto Warren with your misrepresentation "by an editor who failed to appreciate the need for unambiguous ones". To attempt to do so now is disingenuous. So now we have established there is no fault with Warren's edits we come to where the fault really is, your edits. You think it needs disambiguation, you have to put the effort in to use acceptable disambiguation. What you should not do is just revert someone's edits and use IEC prefixes because that is against the guideline. This is in the guideline text and is the consensus. So again we get to the question that you keep on failing to answer: When are you going to start following the guideline that has consensus? Fnagaton 18:20, 9 June 2008 (UTC)
  • Thunderbird, stop arguing; the IEC prefix debate is over. The dispute had been discussed in depth for several months now and more than enough editors have made their desires known and a clear consensus arrived at. There might be some ambiguity in current MOSNUM policy; for instance, where MOSNUM says as follows:

Editors should use the conventional prefixes, such as kilobyte (KB) and megabyte (MB), and disambiguate where necessary.

But MOSNUM is clear as glass when it states as follows:

The IEC prefixes are not to be used on Wikipedia except under the following circumstances:
   • when the article is on a topic where the majority of cited sources use the IEC prefixes,
   • when directly quoting a source that uses the IEC prefixes,
   • in articles specifically about or explicitly discussing the IEC prefixes.

Now, once again, you used (*ahem*)... non-truth to buttress one of you arguments. You wrote, in your 17:31, 9 June 2008 (UTC) post as follows: “The spirit and letter of the guideline is to use unambiguous units.” No, MOSNUM actually states as follows:

The consensus was that for the byte and bit prefixes, the spirit of the MoS is better reflected by having familiar but ambiguous units, rather than unambiguous but unfamiliar units.

Also, T-bird, you wrote in your above 18:04, 9 June 2008 (UTC) post “My edit did not introduce the IEC prefixes, because they were there already. I simply reverted a change that introduced ambiguous units, by an editor who failed to appreciate the need for unambiguous ones.” No, you are (again) wrong. Warren edited Power Macintosh 5500 precisely as MOSNUM policy prescribes; he used the familiar units (the conventional binary prefixes) that are used by the rest of Planet Earth when communicating to a general-interest audience. Furthermore, Warren used them in precisely the way that is has been written a million+ times before in the real world and will be written another million+ times in the future. Computer manufacturers and computer magazines all routinely use the conventional prefixes and perceive no need whatsoever to “disambiguate” their meanings unless they are discussing hard drive capacity. It doesn’t matter whatsoever that you feel the need for disambiguation and further feel that the IEC prefixes are the best way to do so. That ship has sailed and is now at the bottom of the ocean.
If someone edits an article to deprecate the IEC prefixes, and further, does so in a fashion that could have been lifted out of any magazine like PC World or Mac World, then it does not necessarily need to be disambiguated. If you feel such articles do need to be disambiguated, then do not accomplish that end via edits that put the IEC prefixes back into the article. I don’t care if you chose to describe your actions as “restoring wholesomeness” or whatever other specious excuses you like, it amounts to disambiguating with the IEC prefixes. There are perhaps a dozen suitable ways to disambiguate “megabyte” that are prescribed by MOSNUM and which you may avail yourself of; MOSNUM proscribes the IEC prefixes. Your refusal to accept the clear consensus and your convenient ignoring of what MOSNUM really says will not be tolerated. Before you go edit another computer-related article, I suggest that you read what MOSNUM really says. That will spare us having to read your fallacious imaginations as to what you think it says. I suggest that you begin your studies with the three passages quoted above.
Warren’s edit summary, when he reverted your reversion hit the nail on the head when he wrote “no. listen. YOU read the MOSNUM, it quite clearly states "IEC prefixes are not to be used on Wikipedia". don't edit-war over this.” You would have spared us a bunch of time in responding here to you if you had heeded that advise. Greg L (talk) 03:12, 10 June 2008 (UTC)

WP:DEADHORSE --Francis Schonken (talk) 08:03, 8 June 2008 (UTC)

This is a Clue Stick for Wikipedia editors that are as stubborn as a mule. It is a planed two by four. The name two by four comes from the dimensions of the unplaned two by four, namely 2 in × 4 in (5.1 cm × 10.2 cm). The planed dimensions are smaller, namely 1.5 in × 3.5 in (3.8 cm × 6.9 cm).
  • I agree with Fnagaton (and Francis Schonken’s point too). Warren’s edits could have been lifted out of any computer brochure or article out of Mac World or PC World. His use of the conventional binary prefixes like “megabyte” was clear enough because it was common verbiage seen a million times before in real-world usage; it wasn’t “necessary” to disambiguate them. For Thunderbird2 to have entirely undone Warren’s work was a violation of Wikipedia:Do not disrupt Wikipedia to illustrate a point. From now on, I suggest to Thunderbird2, that if you really feel that an expression like “The Windows-box 9000 comes stock with 3 GB of RAM so it runs less slug-like with Vista” truly, really, honestly confuses so many readers, then you can disambiguate said article using techniques seen here on “Mac Pro” (take a look at footnotes 10, 11, and 22). Please don’t just revert other editors’ work by restoring units of measure few readers have ever seen before. The consensus is clear that this is not what we want to see on Wikipedia any more. Whether you agree with the consensus view or not, please respect it.

    I should also add that Power Macintosh 5500 should have had the first instances of each use of “KB” and “MB” spelled out. Some of us here are assuming that the typical reader is more knowledgeable about computer terminology that should be presumed. Note that Encyclopedia Britannica spells out every instance and doesn’t use the unit symbols. Of course, that’s where the number of instances of “megabyte” in the article is limited. I’m going to revise the article to show what I mean. Greg L (talk) 20:31, 8 June 2008 (UTC)

  • P.S. There. These changes illustrate what I have in mind regarding spelling out first instances of “MB” and “MHz”, etc. It also provides link to the first instances of new units of measure, and it disambiguates the binary prefixes via footnotes using familiar wording seen in the real world. We need to remember some of the fundamentals of technical writing in an encyclopedia: some readers who come here to learn are not terribly familiar with computer terminology. Even though we have Wikilinks, the proper thing to do is spell out first instances of arcane technobabble. Greg L (talk) 21:08, 8 June 2008 (UTC)
  • Well Note 1 is clear; but Note 2 - what's a billion, I assume you mean 1024 Mbytes, or is it 1,000,000,000?Pyrotec (talk) 21:18, 8 June 2008 (UTC)
  • The edits look good to me (I take 1 billion to mean 1,000,000,000). Now that we have identified the loophole, I look forward to suggestions on how to fix the root cause of the problem. Thunderbird2 (talk) 21:31, 8 June 2008 (UTC)
  • I lifted that sentence straight off of Apple’s Web site. It’s clear enough to me and apparently, Apple computer thinks it's clear enough for England too. Note Apple’s UK iMac tech spec page. See Footnote #4. It reads as follows: “1GB = 1 billion bytes and 1TB = 1 trillion bytes; actual formatted capacity less.”. England has long used the short scale for named numbers and commerce there observes that real-world practice. The only countries still using the long scale for numbers are non-English speaking. This is en.Wikipedia. If you want to change it to prove some sort of point, be my guest; I’m not going to make a federal case out of it. Greg L (talk) 21:34, 8 June 2008 (UTC)
  • Pyrotec said he interprets 1 billion bytes as 1024 MB and I said I thought it meant 10003 B. What point is it that you think we are trying to prove? Thunderbird2 (talk) 21:41, 8 June 2008 (UTC)
  • P.S. Yeah, I caught that. Pyrotec’s point wasn’t that we had a long scale/short scale issue; Pyrotec was saying that the named number word “billion” can be interpreted as “1024 Mbytes”. I’m not going to even debate this point, other than to provide a link to Billion, and ask you to point out where 1,024,000,000 is one of its defined values. The IEC prefix wording was pored over more carefully than most other text in MOSNUM. It’s clear enough. Again, whether you agree with the consensus view or not, please respect it and abide by it. It does have the virtue of following real-world practices. If you have a problem, take it up with the world, not with editors here trying to follow real-world practices. Greg L (talk) 21:47, 8 June 2008 (UTC)
  • My point is that Note 1 says: Transistorized memory, such as RAM and cache sizes, are binary values whereby 1 KB = 210 (1024) bytes and 1 MB = 220 (1,048,576) bytes; and Note 2 says: "For hard drives, 1 GB equals 1 billion bytes; actual formatted capacities are less". The point I am making is that I would have expected Note 2 to be of the form: "For hard drives, 1 GB equals 1 billion (1,000,000,000) bytes; actual formatted capacities are less".

    If you look at Billion it can be 10^9 or 10^12, but I was not trying to make that point. (It might even be 1,024 * 1,024 * 1,024 or even 1000 * 1,048,576 - but that would be flogging a dead horse :))Pyrotec (talk) 21:52, 8 June 2008 (UTC)

  • OK. After I disambiguated “gigabyte” (using real-world language lifted straight off an Apple web page), I then disambiguated “billion” with a Wikilink to 1000000000 (number). I trust there is no need to disambiguate “1,000,000,000” means? As I pointed out with my above-mentioned edits, there are other sucking chest wounds in some of Wikipedia's computer articles (like spelling out the first instance of “megabyte” and then transitioning to the symbols (MB). Fussing over whether a “billion” really means the standard, English-language word it normally means seems to be more of a need to make a point here than trying to fix anything that really needs fixing. Can we move on now?? Greg L (talk) 22:05, 8 June 2008 (UTC)
  • It is quite simple really, and its getting it right first time, why not change Note 2 to read: "For hard drives, 1 GB equals 1 billion (1,000,000,000) bytes; actual formatted capacities are less", and move on to fix the things that need fixing?Pyrotec (talk) 22:12, 8 June 2008 (UTC)
  • Like I stated above, I would have thought that language that Apple thought was perfectly clear in both England and the U.S. is clear enough. Even Rain Man in a hurry could figure what “billion” means. So writing out “billion” means “1,000,000,000” seemed silly and unnecessary to me. It also seemed like a POINT issue in the closing days of this IEC prefix nuclear war. But maybe that’s just me. I’m not going to make a federal case out of it. Knock yourself out. Greg L (talk) 22:22, 8 June 2008 (UTC)
  • If the prefix war must go on, it must go on, but de-archiving one of the dozen binary prefix archives might just make things more confusing. Beat a dead horse ... is it dead? How many legs does a dead horse have ... are they legs any more or just lumps of meat? Ever tried dead horse? It's great raw with a little garlic or ginger. Francis' transclusion of WP:DEADHORSE looks a little different from when he posted it because I've gone and stuck some <noinclude></noinclude> tags in there so that this page isn't miscategorised and wrongly linked up with the Arabic version of the dead-horse essay and to remove the See also heading. As for me, I still stick with the logical meaning of billion but acknowledge the fact that I'm a dying breed. Until the English-speaking world sees the light ... as have many of the other-European-language-speaking worlds and go back to the original meaning of billion, trillion, etc. ... and that may never happen ... I s'pose we're stuck with this bastardised terminology in the WP article space ... but I'm off on a tangent. Consensus went in some direction today ... or which direction? ... but consensus can change. Have fun with your horse. Let's just not have the prefix war spilling all over the place again. JIMp talk·cont 02:40, 9 June 2008 (UTC)

Minor typographical question

A 100 GB (100,000,000,000 bytes) hard drive

Should this be written as

A 100 GB (100,000,000,000-byte) hard drive

to match the style of hyphenating adjectival units that aren't abbreviated? —Werson (talk) 02:32, 11 June 2008 (UTC)

It's the same thing as writing 1 km is 1,000 meters. You don't write 1 km is 1,000-meter. Headbomb (ταλκ · κοντριβς) 04:42, 11 June 2008 (UTC)
1 kilometer is 1,000 meters, but a 1-kilometer run is a 1000-meter run. —Werson (talk) 04:58, 11 June 2008 (UTC)
So? We aren't talking of a 1,000,000,000-byte hard drive, we're talking about how many bytes there are in a decimal gigabyte. The answer is that there are 1,000,000,000 bytes in a decimal gigabyte. Headbomb (ταλκ · κοντριβς) 05:14, 11 June 2008 (UTC)
  • Yes Werson, I’ve noticed this issue too over the last year; particularly as I’ve been writing examples here. Technically, I believe all the following are correct:
  1. “The computer came equipped with a 100-GB hard drive”
  2. “The computer came equipped with a 100-GB (1,000,000,000 bytes) hard drive
  3. “The computer came equipped with a 100-GB (1,000,000,000-byte) hard drive
  4. "The hard drive had a capacity of 100 GB”
  5. “Hard drive capacity: 100 GB (1,000,000,000 bytes)
I’m not so sure that even professional publications adhere to hyphenating like this but I could be wrong. It would be nice if someone researched this a bit more.
Now that I think about it, I guess here’s the way I’ve been handling it: When one has a bunch of technical specs in an article, the numeric equivalencies can be treated as specifications even though they are technically in the form of adjective modifier requiring hyphenation. My practice lately has been to use the hyphenated form when text reads mostly like near-conversational prose and has an occasional numeric equivalency imbedded within it—the form shown with the first bullet point—and to drop the hyphen when the article has lots of specs and starts to read like a data sheet. In other words, in spec-rich articles, I omit the hyphen because the #1 example above, is very naturally interpreted as the #4 example. Greg L (talk) 06:50, 11 June 2008 (UTC)
According to SI, "100-GB" is never correct; SI unit abbreviations are not hyphenated. I'm only talking about words that are spelled out. —Werson (talk) 09:24, 11 June 2008 (UTC)
  • Ahh. That’s why it looks so odd. Greg L (talk) 16:40, 11 June 2008 (UTC)
Yep. However, the standard doesn't really apply since GB is not an SI unit; it just borrows the giga- convention from the SI. And of course whether giga- should have the SI meaning or not, in that context, is controversial. -- (talk) 01:08, 14 June 2008 (UTC)

Lots of specs and you should be using symbolic/abbreviated forms, thus you'll not need the hyphen; you certainly won't be spelling it out long hand (as in the example above) each time ... generally not more than once. As for the specific example, I'd be more inclined to write "A 100-gigabyte (100,000,000,000 B) hard drive" or, more inclined still, "A 100-gigabyte (100×109 B) hard drive" (using engineering notation). JIMp talk·cont 07:57, 11 June 2008 (UTC)

I know this is not germane to the OQ... but... it really shouldn't be written either way. If you're going to write out all 12 digits, they should all be correct, but alas the chances that any "100 GB" hard drive contains 100,000,000,000 bytes are approximately zero. In fact a drive so marketed will provide somewhat more than 100 GB, less than 101 GB, but of course considerably less than 100×230. The exact number will be different for different "100 GB" hard drives, too; it's even possible (though uncommon) for it to be different among samples of the same make and model. This is not due to "formatting" or file system overhead or any such thing. I am speaking of end-user capacity available at the drive's interface connector, irrespective of hardware low-level format or partition structure or file system overhead, and not including spare sectors. There is simply nothing on a hard drive that would cause its capacity to be a small number multiplied by a power of ten, any more than there is anything to cause its capacity to be exactly a small number multiplied by a power of two. See, this is one reason that I argued previously against disambiguating with exact numbers of bytes: What would seem to be the right number isn't always. This has nothing to do with advocacy of binary prefixes, either; I am not advocating that as an answer. It's just a question of accuracy and (mis)representation: All the digits presented need to be correct. IOW I agree strongly with Jimp's preference for "100×109". Jeh (talk) 06:19, 15 June 2008 (UTC)
  • Well E notation should be avoided, so if anything, it should be 100×109 bytes or 1,0003 bytes. But the fact remains that 100 GB is 1,000,000,000,000 bytes, it's a nominal value. It's like saying the earth-moon distance is 38,000,000,00 cm or that the new budget has 5 billion dollars (5,000,000,000 $, or is it $5,000,000,000?) allocated to buying pink lawnmovers. If super accuracy is needed, then just use a footnote to point that there's always some variations and that drives don't come with EXACTLY 100 billion bytes. These are my two hundred thousand Zimbabwe dollars. Headbomb {— The greatest sin is willful ignorance.
    ταλκ / κοντριβς/Projects of the Week
    12:24, 15 June 2008 (UTC)
I disagree. Everybody knows that wordage like "100 billion dollar budget" are estimates. It might even be 105 billion. In fact that is really a form of scientific notation. But I doubt you will find many encyclopedias or news articles writing these numbers out as "100,000,000,000 dollars." Due to the confusion that already exists re. hard drive sizes we need to do what we can to avoid confusion, or at least avoid reinforcing it. When Windows (for example) displays a hard drive size by writing out the entire number, all of the digits are correct (at least as far as Windows sees - you might be looking at just one partition, and in any case partition structure is not accounted for, but it's a lot closer than "100,000,000,000"). Similarly on the drive manufacturer's spec sheet, if they write out the number, it is exact (according to my definition above, excluding spare sectors, etc.). Following current common usage, then, if we write out all the digits then they all should be correct. Failing that, engineering notation (100×109 bytes) is the best way to show which digits have confidence. Of course none of this will answer the reader's question of why the OS displays that drive's size as "93 GB". Jeh (talk) 17:47, 15 June 2008 (UTC)

Let's get the IPs banned.

Whoever that person is, it's getting pretty damn annoying to revert his/her edits on a bunch of articles. You guys are more familiar with this, so you probably know what history is associated with this.

Let's build User talk:Headbomb/Request for the I.P.s to be blocked. Headbomb (ταλκ · κοντριβς) 13:47, 11 June 2008 (UTC)

That page is already here. --Quilbert (talk) 14:15, 11 June 2008 (UTC)

Page has been updated to cover most of that IP'S activity. Please add whatever info you feel is useful, such as past verdicts, details about IP activities, as well as Sarenne etc... Headbomb (ταλκ · κοντριβς) 19:02, 11 June 2008 (UTC)

What are we going to do with editors who edit against consensus?

All: How are editors going to be able to make MOSNUM-compliant edits if other editors flout MOSNUM policy and revert the edits in a fashion that flies totally in the face of what MOSNUM calls for?

Note this paragraph of our “Itanium” article. Note also its #27 footnote. This is the result of a single edit session I made a few days ago (∆ here). Then Thunderbird2 simply reverted the article (here is the ∆ between the article before my work and after he reverted it; it was a straight reversion). He is even keeping a log (User:Thunderbird2/my sandbox) of the articles other editors try to bring into compliance with MOSNUM that he reverts so he doesn’t seem at all shy about this whatsoever. The end effect of his edits are to simply denote computer memory using the MOSNUM-banned (and highly unfamiliar) units of measure like MiB and KiB.

Note also that I did my math when I disambiguated the article. Where it says, “The bus transfers 2x128 bits per clock cycle, so the 200 MHz McKinley bus transferred 6.4 GB/s…”, I had confirmed that at 200 MHz, the value of “6.4 GB/s” was precisely 6.4 × 230 bytes per second, and thus, footnote #27 correctly described the value of “GB” for that transfer rate too.

I just do not have any idea how to resolve this. An enormous amount of discussion has occurred here on Talk:MOSNUM and Thunderbird2 seems determined to just ignore the consensus and edit-war here. Any suggestions? What is the most expedient way to get an editor to follow the rules without going through a tedious, formal RfC or something? Greg L (talk) 03:10, 14 June 2008 (UTC)

  • Simplest way would be to simply ask him to please stop to disruptively edit Wikipedia. If that fails, then there's no avoiding the "formal" measures, although there are non-formal arbitrations out there. Headbomb (ταλκ · κοντριβς) 05:15, 14 June 2008 (UTC)
    • GregL, the occasional use on Wikipedia of an IEC prefix, even though banned, is not the end of the world. [I didn’t say it was the end of the world. Greg L (talk) 17:38, 14 June 2008 (UTC) ] You won your "I don't like it" position (for now); I suggest you take your victory and be happy and stop worrying about the small number of editors will not follow the new rules. [I suggest you not take it upon yourself to pretend to counsel others as to which of Wikipedia’s rules OK to break. Wikipedia won because it no longer stands out like a sore thumb in the information world. My point is that it’s not fair to all the other editors, like Warren who take the time to do proper editing; they shouldn’t have their work undone by an intransigent editor who refuses to follow the rules. This is a collaborative writing environment; we can’t condone the actions of editors who are insistent on refusing to observe rules of conduct. Greg L (talk) 17:38, 14 June 2008 (UTC) ] Of course, if you and Fnagaton want to take it upon yourself to stalk everyone who has spoken up in favor of IEC prefixes and check up on the articles they've edited recently to be sure no "Klingons" are "practicing their IEC prefixes," that is certainly your perogative. [That’s nothing but fallacious accusations of bad faith intended to demean another editor in hopes of swaying others’ opinion your way. I will ignore it for what it is. Greg L (talk) 17:38, 14 June 2008 (UTC) ] But as for "what to do about violators" (is that the sound of wringing hands I hear?) I never heard of any sort of action being taken against someone who simply used a prefix or measurement unit not approved by MOSNUM, even persistently across many articles; such is not in and of itself disruptive. [No, you are wrong. “[editing] persistently across many articles” is indeed disruptive; and it’s a violation of Wikipedia:Tendentious editing and Wikipedia:Do not disrupt Wikipedia to illustrate a point. It’s not about “using a unit of measure not approved by MOSNUM”, it’s about editing against consensus and being purposely disruptive (if he keeps at it). Greg L (talk) 17:38, 14 June 2008 (UTC) ] I know it seems disruptive to you because, all rational arguments aside, you have described yourself as having a very strong, visceral reaction to the IEC prefixes, but really, it just isn't that big a deal.[I didn’t say it’s a “big deal”, as you put it, nor do I have a “visceral reaction” to other editors breaking Wikipedia rules. On a 1-to-10 scale, with a “10” being stuck on on a deserted island with Miranda Kerr for a year, and a “1” being receiving third-degree burns over 90% of my body, I’d give it a solid “5.0”. Greg L (talk) 17:38, 14 June 2008 (UTC) ] Now if someone insists on starting an edit war over such changes, then sure, that would be considered disruptive. But it should be handled like any other disruptive edit war; that the point being argued is a matter of MOSNUM policy doesn't change anything AFAICT, except that it gives one side a definite position of rightness. [ No. It is never acceptable to use edit waring as a tool to provoke endless debate of a MOSNUM policy, particularly a policy that was the product of four-straight months of such intensive debate. The proper remedy is formal action against the offending editor for editing against consensus after they have been amply warned about it. Greg L (talk) 17:38, 14 June 2008 (UTC) ] Jeh (talk) 10:21, 14 June 2008 (UTC)
  • The concern is about edit warring , not about Jimmy Two-Pants writing an article with KiB. Headbomb (ταλκ · κοντριβς) 15:00, 14 June 2008 (UTC)
  • I warned Thunderbird2 against disruptive editing here. Greg L (talk) 18:12, 14 June 2008 (UTC)
  • Edit warring is disruptive regardless of who engages in it, and as always, it would be looked at by an admin uninvolved in the dispute as to the particulars of the situation if a request for administrative attention were made. I'm not personally concerned by how disambiguation is to be done. I would prefer to use the binary prefixes, but if we would prefer to use exact numbers with exponential notation in a parenthetical or footnote, that's fine—exact numbers are about as unambiguous as you get. I would, however, expect there to be concern if instances of the binary prefixes were changed to the ambiguous units without an appropriate alternate form of disambiguation, as this violates the prohibition against ambiguity in unit designations. I think it's much more important that we disambiguate than the exact means by which we do so. Seraphimblade Talk to me 19:13, 14 June 2008 (UTC)
  • Using MB without at least a mention that it's used in the binary or decimal sense goes against MoS anyway, so if that happens, just specify which is meant. Headbomb (ταλκ · κοντριβς) 19:19, 14 June 2008 (UTC)
  • I don’t want to get into a situation where we have to discuss the atomic-level details of how each and every article is edited and disambiguated. Different articles are “different” and have their unique circumstances and needs. My point about “disruptive” editing is that if an article hasn’t been updated sufficiently or satisfactorily, the proper course is to correct the error using the conventional prefixes. Thunderbird2 can add or revise footnotes he thinks are absent or in error. What is not permissible is to “fix” an article by putting all the IEC prefixes back in. That is discouraging and takes the fun out of the hobby for many other editors. Their right to enjoy their hobby while following the rules overrides Thunderbird2’s right to make a point with what he feels are superior units of measure. We’ve all heard Thunderbird2’s arguments that the IEC prefixes are unambiguous and are therefore a far superior way to be precise. We acknowledge this reality but have rejected that virtue as being insufficient to override the disadvantages of the units: they haven’t seen real-world adoption and are therefore unfamiliar to our readers. This has all been discussed ad nauseam and we’re done arguing that issue; having to go all that over and over is just a matter of T-bird refusing to 'get the point'. It’s a simple fix for him to abide by MOSNUM and get along with the other editors here: Don’t “fix” articles that other editors are trying to bring into compliance with MOSNUM by reverting them back to wholesale use of the banned IEC prefixes; if he sees errors of omission or fact, he should help them by suggesting a better disambiguation that is still compliant with MOSNUM. That is not too much to ask. Greg L (talk) 18:18, 15 June 2008 (UTC)

IEC prefix debate

Hi to all – I have created this experimental sub-page as an effort towards finding a consensus. Hope it works. Cheers --Quilbert (talk) 11:18, 11 June 2008 (UTC)

I prefer it stays on this page where everyone can see it, can reference the archives and not be confused or fail to see something on a sub-page. It was tried before putting this topic on a sub-page and all that happened was that the tiny minority were able to make a lot of fallacious arguments without the majority being aware of it. Fnagaton 11:34, 11 June 2008 (UTC)
We already found consensus.Headbomb (ταλκ · κοντριβς) 13:43, 11 June 2008 (UTC)
  • I agree with Headbomb first, Fnagaton second. We’re done on this. The IEC prefixes where debated in great depth and with great passion for several months. The final vote was clear as to the consensus. Except for a few holdouts who didn’t get their way and a few editors who arrived late to the battle and want to see some action, the rest of us are sick to death of this. And even if we did want to further discuss this (*yuck*), it would be done here, not on some remote backwater page where most editors will miss it. That would be the worse possible way to develop a real and true “consensus” of any sort because the only participants would be highly animated hot-heads from two camps. Greg L (talk) 15:30, 11 June 2008 (UTC)
That specific model is only possible on a sub-page. Please have a look. Headbomb: Where exactly was the consensus established? --Quilbert (talk) 14:08, 11 June 2008 (UTC)
A clear consensus was established here. Thunderbird2 (talk) 14:43, 11 June 2008 (UTC)

And was superceded by the more recent, and this time well argued consensus found here. Opposition were challenged to substantiate their opposition, which they did not do. They referred to the "old consensus", which was not back up by sound argument and which was thus dismissed under the provision that unsubstantiated opposition can be disregarded.Headbomb (ταλκ · κοντριβς) 14:51, 11 June 2008 (UTC)

No TB2 that is not the consensus, this is. Fnagaton 14:50, 11 June 2008 (UTC)
q.e.d. no consensus --Quilbert (talk) 14:55, 11 June 2008 (UTC)
More like Q.E.D. there is consensus for what the current version is. Fnagaton 14:58, 11 June 2008 (UTC)
What Fnag said. Headbomb (ταλκ · κοντριβς) 14:59, 11 June 2008 (UTC)
Come on. You cannot establish consensus with 10 people. That the discussion is still raging means that there is none. And that diferrent people point me to totally different consensi leaves no doubt about it. --Quilbert (talk) 15:10, 11 June 2008 (UTC)
We can because the person causing the "discussion to rage" is just misrepresenting what the consensus actually is. The person pointing you in the other place is just trying to cite a really old vote and in this case votes don't make consensus, valid arguments make consensus. As Headbomb pointed out the person refused to back up his statements with sound argument so his unsubstantiated opposition can be discarded. That is reasonable. You'll note both Headbomb and myself have pointed TB2 to the same place for where the consensus really is. If you read through the archive and want to present a new stronger argument that the others have not tried then please do. I don't think there is anything to be gained by opening an entirely new page, yet again, just so the IEC proponents can restate their points of view and again fail to provide substantive argument, again. Fnagaton 15:23, 11 June 2008 (UTC)
You can read about my substantive arguments on the page given. --Quilbert (talk) 15:28, 11 June 2008 (UTC)
I did, I didn't see anything new that has not already been covered at great length by the archive we both linked. Headbomb did you see anything new? Fnagaton 15:33, 11 June 2008 (UTC)

There is wide consensus on not using the IEC prefixes. The consensus is not only on the MOSNUM talk page but in the entire publishing and computer industry. We have just spent 3 month finding an acceptable binary prefix guideline. It pass 10 (11) to 2(3). It is not just ten people, we are going to follow what the rest of the world is doing. The only problem is the Thunderbird2 will not accept the decision. We don't need an ongoing debate on a secret sub-page. -- SWTPC6800 (talk) 15:15, 11 June 2008 (UTC)

You know as well as I do that most of the outside world is ignorant and sluggish. Don’t count votes you don’t have. Also, what about Omegatron, Jimp, and the like (not to mention that mysterious vandal-IP). There are too many people not content to talk about consensus. --Quilbert (talk) 15:26, 11 June 2008 (UTC)
Omegatron also refused to take part in the debate. Jimp voted for the current version. The vandal IP is a banned user. Fnagaton 15:33, 11 June 2008 (UTC)

Alright, I see, all are sick about this. So let’s leave it as it is. I just had the impression that there is still much hostility in the air. Anyone who is interested in “substantiated arguments” for IEC prefixes can read User:Quilbert/IEC. Just for your interest, as I’m hereby retracting my commitment, as it seems unsolicited. Peace is most important. See you --Quilbert (talk) 15:41, 11 June 2008 (UTC)

I'm very relieved to hear that. Thank you for being reasonable and not dragging this into another multi-month debate. Headbomb (ταλκ · κοντριβς) 15:47, 11 June 2008 (UTC)

Dito. Greg L (talk) 16:00, 11 June 2008 (UTC)

Dito x 2. Don't feel too bad Quilbert, we've just got out of a three month long debate on this and have a huge archive to prove it. ;) Fnagaton 16:01, 11 June 2008 (UTC)
I'm not happy either. More specifically: I find it ludicrous and appalling that Wikipedia has voted (or agreed to consensus, or whatever term you like) to outright ban (not just discourage, not just deprecate, but ban) the use of a notation that is a standard accepted by the IEC and several other standards bodies in the field. I'm letting it go for now, but I don't think this consensus will last. Jeh (talk) 19:43, 11 June 2008 (UTC)
The "consensus" was not reached by a procedure that came anywhere close to a reasonable one. During discussions (as can be seen from the more detailed voting earlier) there was little support for a ban. There was even acceptance of striking the ban at some point. After it reappeared, packaged as one element in a large rewrite, it did not have enough weight for many of the participants to vote down the whole package. Negative votes were just reverted repeatedly! The bullying by a few participants surely has scared away many interested others. Therefore having a renewed debate on only the IEC issue is warranted. −Woodstone (talk) 19:54, 11 June 2008 (UTC)
To Woodstone, the procedure is fine. The support for the ban is obviously clear, TB2's attempts to cite an old vote has been completely refuted because votes don't make consensus, strong arguments make consensus. The "acceptance of striking" was only there for as long as questions were being answered, once it becamce clear that questions were not being answered and substantive arguments were not being made to keep the text removed and especially since much stronger arguments were made to unstrike the text then the text came back again. Which is how consensus works. It is only warranted to have another debate if something new and substantive is brought to the debate. Just repeating the same old refuted arguments won't help the "pro-IEC" point of view. As for the accusation of bullying it is rediculous and that is not a substantive argument relevant to this topic. For example, take a look at TB2's repeated incorrect statements. Whenever TB2 repeatedly posting incorrect statements about consensus or just reverts other editors when they follow the guideline to use familiar prefixes or when TB2 tries to cite old votes without providing substantive arguments then that actually weakens the point of view he has. When working on consensus any unsupported objections can be largely ignored, for obvious reasons. Look at the number of times Headbomb and others (myself included) directly asked questions which were ignored, it didn't just happen once or twice it happened tens of times in total. To both Jeh and Woodstone, it isn't a "ban" by any stretch of the imagination. If it was a ban then it would say "IEC prefixes should never be used", since obviously it doesn't say that because it gives clear situations when IEC prefixes can be used then calling it a ban is not accurate. What I suggest is that instead of beating this dead horse you go away and really think about a strong persuasive argument, one that has not already been tackled in the many archives on this subject. Then if or when you have that argument present it here so we can all see it. To be honest I've been willing the "pro-IEC" side to make a good attempt at providing a strong argument for several months now because it helps to raise the overall quality of the debate and produce stronger counter arguments. But so far the "pro-IEC" side have not done so, which is a shame. Fnagaton 20:46, 11 June 2008 (UTC)
You can easily say that as you can determine yourself what you consider a strong argument. I believe that the mathematical argument alone outweighs all contra arguments. But, well, everyone has different priorities. But just saying that arguments are weak without telling why is meaning- and useless. --Quilbert (talk) 21:54, 11 June 2008 (UTC)
It has been shown many times exactly why the arguments are weak, in the archive Headbomb and I linked that is clearly demonstrated. For example, the fact is when challenged by strong counter arguments the "pro-IEC" people suddenly stopped answering questions and instead started repeating statements that had nothing to do with the actual topic and were easily refuted. By definition, unsubstantiated objections are weak arguments. Fnagaton 21:57, 11 June 2008 (UTC)
I cannot see any mathematical operation on that page. So that argument isn’t there. --Quilbert (talk) 22:18, 11 June 2008 (UTC)
The mathematical operation you refer to is not an argument. It is like saying "1+1 = 2 therefore that proves my point that IEC prefixes should be used", it doesn't because the conclusion does not logically follow from the statement and, more importantly, doesn't take into account the existing policies and guidelines of Wikipedia.Fnagaton 22:23, 11 June 2008 (UTC)
Those operations show that the mathematical model used by current literature is inconsistent. Units without a good mathematical model should not be used. That is the point. --Quilbert (talk) 22:29, 11 June 2008 (UTC)
Again, showing something is inconsistent does not logically follow onto "should not be used" because the connecting arguments have not been shown. The stronger counter argument is that Wikipedia does not only follow mathematical reasons for including things in articles. English is an inconsistent language (waves to US and UK English), yet we still use it for these articles when the consensus shows us on a per article basis what to use. We don't try to reduce articles down to mathematical models because the average reader doesn't understand that. The aim for writing articles is to help the reader to understand, not to include virtually unused terms that make it harder for someone to understand. Explain please exactly what is inconsistent with saying "1024 bytes"? It is not inconsistent, it is not ambiguous. Numbers, like 1024, are certainly better understood that obscure terms like KiB, for example. By the way, if you don't see me reply for a few days it is because I have to catch a plane for a very long journey. I have no doubt that one of the others will reply on my behalf anyway. ;) Fnagaton 22:34, 11 June 2008 (UTC)
As I said: consider the unit GB/GB. If I tell you that I have used up 0.1 GB/GB of my hard drive that would mean 10.7%. Now, that technically means that 0.1 = 0.107 which is obviously wrong.
If you need a connecting argument, just take Mars Climate Orbiter#The metric mixup for an example of what a bad model can do. --Quilbert (talk) 22:49, 11 June 2008 (UTC)
Concerning your counter argument: Units are mathematical entities, languages are not (yet). --Quilbert (talk) 22:51, 11 June 2008 (UTC)
The problem is you are incorrectly mixing base ten and base two values and trying to then claim that because the ratio is not linear that means it is somehow wrong. The fault is not with the units/values per se but with the method you used in the first place to calculate the ratio of two systems. Since your initial comparison is flawed then so to is your conclusion. Also, the connecting argument you cite is irrelevant to this particular topic because that details a communication issue, which of course doesn't automatically mean IEC prefixes have to be used to solve the issue. There exist other methods that don't lead to confusion, like stating the exact number of bytes and not relying on prefixes at all. These prefixes being debated here are defined by the different use in language and industry. And now, I really do need to get going. You didn't answer my question by the way. Fnagaton 23:16, 11 June 2008 (UTC)
You mean that question: “Explain please exactly what is inconsistent with saying "1024 bytes"?”? What should be wrong with that? If we just state the exact number of bytes, I’m in immediately.
It’s not me who mixes up bases incorrectly, I’m just showing that the current consensus does so. You can also take this one: \frac{100KB}{10s}=10.24 KB/s.
That Mars mission is a good example because it demonstrates well how not adopting proper standards does lots of damage.
But whatever, we won’t agree anyway. Since you can’t answer now, this puts a natural stop to this discussion, which we shouldn’t have led in the first place. I just couldn’t stop myself when you implicitly claimed my arguments were weak. --Quilbert (talk) 23:57, 11 June 2008 (UTC)
You are mixing bases in the calculations though, which is not what the real world does when calcuating these things. Since the calculations as written are incorrect the conclusions drawn are also fallacious. There are 10 types of people in this world who understand binary, those that do and those that don't. The problem with the calculation as presented is that it is mixing two values that use different units of measure, decimal and powers of two. It doesn't matter that the letter symbols are visually the same because what matters is that the calculation incorrectly mixes these two systems. What the calculation should be doing before applying the "per second" is to convert into a mutually common unit of measure and then applying the "per second". In this case what needs to be done is to strip the prefixes and use the exact number of bytes, then agree on a mutally acceptable common form for the resulting units, then apply the calcuation. In the same way, you wouldn't take the binary number 10000000 and divide by the decimal number 10 then hope that 1000000 (binary) or 1000000 (decimal) is the correct answer for the calcuation, because obviously the answer is incorrect. What needs to be done of course is to either convert 10000000 (binary) into decimal, or convert 10 (decimal) into binary, then do the divide, usually this means choosing a sensible form for the base. Another example would be buying 5 loaves of French baguette and 2 granary. It would not be sensible to try to perform mathematical calculations based on the ratio of French baguette and granary for the volume of bread per unit of currency, unless of course the loaves were converted into a commonly acceptable form first, like breadcrumbs for example. Fnagaton 17:10, 18 June 2008 (UTC)
Continued on Fnagaton’s talk page … --Quilbert (talk) 00:07, 21 June 2008 (UTC)
I believe that arguments in the form of "I don't like it" and "They don't like it" don't need to be demonstrated as being weak. They are weak. I'm personally pro-IEC, since there is a million of problem with using KB to mean 1024 bytes, but I also realize that wikipedia is not a crystal ball. No one uses the IEC prefix, and the Jimmy Longshorts out there certainly aren't familiar with them if people working with computers for 20+ years never even heard of them. When the real world adopts them and starts using them at a rate of more than 0.1% (google scholar all years) and 0.6% (google regular for last year), then they'll be fit for Wikipedia. Wikipedia is not a crystal ball.Headbomb (ταλκ · κοντριβς) 22:03, 11 June 2008 (UTC)
And since Headbomb, who is personally pro-IEC, has not been convinced by the "pro-IEC" argument to include them in Wikipedia then that just goes to show just how weak the "arguments" have been. ;) Fnagaton 22:29, 11 June 2008 (UTC)

Binary meaning of kB, MB, ... now perhaps illegal in USA

The United States Department of Commerce has the authority to interpret the International System of Units for the USA. On May 9, 2008, the Federal Register carried a notice that "announces the publication of the latest interpretation of the SI for the United States in the 2008 Edition of NIST Special Publication 330 "The International System of Units." That publication, in turn, states on page 29, "These 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). The IEC has adopted prefixes for binary powers in the international standard IEC 60027-2: 2005 . . ."

The Federal register notice also states the Special Publication contains not only the authoritative interpretation of SI for the USA, but also "editorial style guidelines". The format of the publication does not make it absolutely clear what is authoritative and what is a guideline. So that leaves it to readers to decide whether using SI prefixes as if they had binary meanings is merely contrary to an official United States guideline, or outright illegal.

Either way, this change illustrates that this section of the Manual of Style may have to be revisited either if the IEC prefixes become more popular, or if they are definitively found to be illegal in a marketplace of significant size. --Gerry Ashton (talk) 23:41, 18 June 2008 (UTC)

You shouldn't hold your breath waiting for some computer company to be prosecuted for selling a computer with 2GB of RAM. A ton of stuff is published in the Federal Register every day. This was just an announcement that the National Institute of Standards and Technology had updated NIST Special Publication 330. [11] After the publishing industry switches to MiB, Wikipedia can follow. The Federal Register does not trump the First Amendment to the United States Constitution. -- SWTPC6800 (talk) 01:32, 19 June 2008 (UTC)
It appears that the NIST recognizes the need to follow the current literature.
NIST Special Publication 330, 2008 Edition. Page 31 [12]
"Nonetheless it is recognized that some non-SI units still appear in the scientific, technical, and commercial literature, and will continue to be used for many years. Some non-SI units are of historical importance in the established literature. Other non-SI units, such as the units of time and angle, are so deeply embedded in the history and culture of the human race that they will continue to be used for the foreseeable future. Individual scientists should also have the freedom to sometimes use non-SI units for which they see a particular scientific advantage in their work. An example of this is the use of CGS-Gaussian units in electromagnetic theory applied to quantum electrodynamics and relativity. For these reasons it is helpful to list some of the more important non-SI units, as is done below. However, if these units are used it should be understood that the advantages of the SI are lost."
-- SWTPC6800 (talk) 01:51, 19 June 2008 (UTC)
Swtpc6800 is incorrect in saying that the Federal Register notice is just an announcement that Special Publication 330 had been updated. The previous, 1998 version of the interpretation was self-contained. So before the new notice came out, Special Publication 330 was merely an informative pamphlet from the government, now it's the official definition of SI in the USA. Also, Swtpc6800's remark about the first amendment neglects to address this phrase from the U.S. Constitution (Article 1 Section 8): "the Congress shall have power to . . . fix the standard of weights and measures".
Finally, continued use of customary units cannot be compared to usurpation of prefixes that were first established by SI. --Gerry Ashton (talk) 04:30, 19 June 2008 (UTC)
The computer industry was using K in the binary sense in the 1950s. The SI was adopted in 1960, so K= 1024 predates SI. A computer term that has been in constant use for over 50 years is a "customary unit". -- SWTPC6800 (talk) 14:56, 19 June 2008 (UTC)
The adoption of SI in 1960 was merely an update to the longstanding metric system's kilogram-metre-second units. This use of the kilogram goes back at least to 1799, and was internationally accepted by the signatories to the Metre Convention from 1875. Use in computers definitely came later.LeadSongDog (talk) 15:56, 19 June 2008 (UTC)
I really shouldn't have said the prefixes were created by SI; they existed long before that. See, for example, the original U.S. law authorizing the metric system, which was passed in 1866. --Gerry Ashton (talk) 15:45, 19 June 2008 (UTC)
So? It doesn't matter what the date is. What matters is what the real world uses. Fnagaton 18:35, 19 June 2008 (UTC)
He's acknowledged the error. I'd suggest you let it go.LeadSongDog (talk) 19:17, 19 June 2008 (UTC)
  • The first rule of technical writing is “don’t confuse your readership.” Our job here is not to try to promote social change. We follow the way the computing world communicates because those commonly accepted practices are the ones our readership are accustomed to. If the rule of law governing trade and commerce has any effect whatsoever, it will cause the slow adoption of the IEC prefixes. And when there are a fair number of computer companies and computer magazines using the IEC prefixes, we can follow suit too and quickly jump on the bandwagon. Greg L (talk) 04:21, 19 June 2008 (UTC)

Reply to Greg_L's accusation of disruption

I’ve replied to Greg_L's note on my talk page, and posted most of you on your respective talk pages. Just in case I’ve left someone out, please see my detailed response here. Thunderbird2 (talk) 19:00, 15 June 2008 (UTC)

  • If you adhere to the requirements of MOSNUM and treat other editors as you would have them treat you, then there is no problem. Greg L (talk) 21:44, 15 June 2008 (UTC)

  • P.S.: I see, T-bird, that you are still actively trying to recruit editors to your way of thinking and clearly hope to overturn a policy that you obviously still disagree with (never mind the fact that all computer manufacturers and general-interest computer magazines and all professional encyclopedias have seen fit to not do as you advocate). It’s going to take more than one or two additional like-minded editors to accomplish what you desire.

    As I stated here before, I am perfectly content with conducting a h-u-g-e, Wikipedia-wide vote, with invitations to editors across all of Wikipedia’s computer and technical-related articles. For too long, MOSNUM policy has been unduly influenced by a small club of the self-appointed elite who camped out here and gamed the system to make Wikipedia conform to their desires. As happened with the IEC prefixes, all it took was one, single editor like Sarenne to spread a cockamamie policy across a hundred articles like napalm (until he was banned for life). Once done, such things are hard to undo.

    As happened in our latest vote, the litmus test in deciding the new vote will not be whether there is 100% buy-in from all editors, but whether there is a “general” consensus. And, as I proposed in the preceding link, I will seek to ensure that the discussion and voting is monitored and arbitrated by a panel of three mediators, whose ruling will be final. This particular Wikipedia policy is too important to have settled by simply wearing editors down until some 5-to-3 vote conducted on a remote, backwater page somehow manages to reverse policy. Wikipedia’s technical pages have Ph.D.s with curriculum-writing experience and similar contributing experts; most have no knowledge whatsoever about MOSNUM and Talk:MOSNUM. If this issue comes up for a vote again, there will be a much wider diversity of input than ever before. In my opinion, it’s high time for that; Wikipedia will be much better off for it.

    In the mean time, give it a rest will you? Everyone else—even the editors you are now trying to sway—are sick to death of this issue; either that, or they are wise enough to recognize the reality of the situation and don’t have the stomach for the bickering and revenge you seem to seek. Greg L (talk) 19:46, 16 June 2008 (UTC)

The Manual of Style (dates and numbers) now follows the binary unit style used by the publishing and computer industry. This was not a 10 to 3 vote but the overwhelming consensus of the real world. Greg's "Follow current literature" proposal has been removed from MOSNUM and replaced by one developed under the leadership of Headbomb. The binary prefix debate is over, can we move on? -- SWTPC6800 (talk) 03:22, 17 June 2008 (UTC)

    • I'm not happy with the tone of this discourse. TONY (talk) 13:36, 17 June 2008 (UTC)

Writing out years at start of sentences?

The MOS states, when a number is at the start of a sentence, it should be written out, unless it's a "Proper name" or "formal numerical designation." I would posit that a year counts as a formal numerical designation, and therefore, when it opens a sentence, it should be "1952 was a bad year for the studio," instead of, "Nineteen fifty-two was a bad year for the studio." This is in re an argument on RKO Pictures. So, when a year is at the start of the sentence, should it stay a number, or should it be written out? --Golbez (talk) 19:41, 24 June 2008 (UTC)

I think Golbez has found another exception. There is at least one more: The stock price fell from 101 to 9 in three hours. should not say nine. Septentrionalis PMAnderson 19:49, 24 June 2008 (UTC)
  • I agree, Golbez has found another exception where beginning a sentence with numerical digits could per permitted. I would add that it might look even better, IMO, to revise the sentence to read “The year 1952 was bad for the studio.” It’s passive voice, but beats starting with digits. At least, that’s the way it seems to me; I’m no expert on this. Greg L (talk) 22:00, 24 June 2008 (UTC)
The MOS specifically is against that, "Avoid inserting the words the year before the digits (1995, not the year 1995), unless the meaning would otherwise be unclear." --Golbez (talk) 23:40, 24 June 2008 (UTC)
  • Greg: It's not passive voice. Yes, MOS does say not to write "The year ...". I'd do it only in desperation. I personally find the spelling out of years impossibly fussy and long-winded. TONY (talk) 04:21, 28 June 2008 (UTC)
  • I see. That makes sense. No point having extra verbiage that doesn’t do anything other than pad a sentence with introductory fluff. I suppose a better way to write the sentence that avoids passive voice, and avoids padding, and doesn’t begin a sentence with digits is “The studio experienced a tough year in 1952.” A declarative statement that provides additional specificity might be better yet: “The studio released a string of poorly received movies in 1952” or “The studio lost money in 1952” (or whatever the problem was). There are probably a hundred ways to accomplish this without begging a sentence—or worse yet—a paragraph, with digits. Greg L (talk) 05:40, 25 June 2008 (UTC)
    • Recasting is always a possibility; we should mention it in general. The current problem does not seem to be resoluble that way; I tried, and was reverted and insulted for my pains. There is always another way. Septentrionalis PMAnderson 13:44, 25 June 2008 (UTC)
  • And you’ve probably got a print copy of a manual-of-style book in your possession (or at least heavily refer to an on-line one). Yes? Greg L (talk) 00:11, 26 June 2008 (UTC)
The issue would be resolved by a simple edit to existing exception 3 under Numbers as figures or words. The bullet point currently reads
Numbers that begin a sentence are spelled out; alternatively, the sentence can be recast so that a figure does not begin the sentence.
This might helpfully be edited to read
Numbers—including years—that begin a sentence are spelled out; alternatively, the sentence can be recast so that a figure does not begin the sentence.
This accords with the Chicago Manual of Style, which gives as the pertinent rule: "At the beginning of a sentence any number that would ordinarily be set in figures is spelled out, regardless of any inconsistency this may create." CMS provides an example exactly on point: "Nineteen seventy-six was the year of the nation's bicentennial celebration." Following this, there is a subordinate clause to the primary rule: "If this is impracticable or cumbersome, the sentence should be recast so that it does not begin with a number." According to CMS, the standard is clearly to spell out; clearly, beginning a sentence with a year expressed as a numeral is improper, full stop, without exception. If an extraordinary consideration of "practicality" or "cumbersomeness" applies, then recasting may be pursued; however, spelling out a year at the beginning of a sentence is clearly not regarded as usually "impracticable or cumbersome."
The CMS directive accords with the practice of the New York Times, which commonly publishes sentences that begin with spelled-out years (and, of course, never publishes a sentence—aside from direct transcription—that begins with a year expressed as a numeral). Here are a half-dozen essentially random examples: [13] (see graf 1), [14] (see graf 5), [15] (see graf 1), [16] (see graf 13), [17] (see graf 4), [18] (see graf 1 on this, the second page of the online version of the article). I've restricted my examples to articles that offer free access to all; I have also not included any of the many, many examples where a person is quoted uttering a sentence that begins with a year--which is, of course, invariably spelled out.
A few editors seem to hold the personal opinion that beginning a sentence with a spelled-out year is "pedantic" or somehow odd. The fact is that it is well-established style and one that is commonly seen in leading publications in American English. In sum, while editors should have the option of recasting where it makes a substantial difference, there is clearly no need to force recasting on sentences that demonstrate absolutely admirable, well-supported style. The change I've proposed—or a similar one elsewhere on the page—should nip such unnecessary contortions and the resulting disputes in the bud.—DCGeist (talk) 06:44, 26 June 2008 (UTC)
Well to somebody who doesn't read American newspapers, namely me, the spelled-out style looks clumsy, unnecessarily long-winded, harder to assimilate, and gratuitously at odds with how we represent dates anywhere else. I don't dispute that the sentence shouldn't start with figures - this is, e.g., the first point the Economist style guide makes on numbers generally, though it doesn't specifically discuss years [19]; but I think the sentence should preferably be re-cast. Jheald (talk) 08:46, 26 June 2008 (UTC)
I believe expressing a general preference in the MoS would be a bad move, as such general preferences are quickly interpreted as universal rules by the narrow-minded. In actual practice, some sentences lend themselves to recasting and others do not. It is instructive to look at the present case in RKO Pictures, where both of the attempts to recast the sentence have resulted in clearly inferior expression—awkward, redundant, conceptually jumbled. Some statements do naturally call for sentences that begin with a year—the evidence both of the present case and of the many instances in the New York Times supports this assertion. I hope we will agree that when stylistic preferences that are essentially visual clash with the demands of clear and felicitous verbal expression, the former must bow to the latter.
As a side—but important—point, please recognize that you are objectively incorrect when you claim that spelling out a year at the beginning of a sentence is "gratuitously at odds with how we represent dates anywhere else." Aside from your inappropriate use of "gratuitously" (I have provided more than enough evidence to demonstrate that the position you happen to disagree with is far from gratuitous), the fact is that we, just like the New York Times and the Chicago Manual of Style and every respectable English-language periodical and publishing house I know of, spell out the year at the beginning of a sentence when we're quoting someone's spoken words.—DCGeist (talk) 10:18, 26 June 2008 (UTC)
Talk about a storm in a teacup! Recasting seems like the best way to deal with this minor infelicity, as I agree spelling the year out looks poor and long-winded. --John (talk) 16:24, 26 June 2008 (UTC)
Thank you. For the record, the recommendation of CMS is, in full: "When a number begins a sentence, it is always spelled out. To avoid awkwardness, a sentence should be recast." Septentrionalis PMAnderson 01:30, 27 June 2008 (UTC)
  • Thank you PMAnderson. The tighter the rule, the better it is. The Chicago Manual of Style on this issue is, IMO, perfect and should be adopted by for Wikipedia’s MOS. Greg L (talk) 05:57, 28 June 2008 (UTC)

←Looks good. TONY (talk) 08:04, 28 June 2008 (UTC)

monthly update: request for advice

I haven't watchlisted MOSNUM for most of this month (swamped my page when I did). Advice please: almost 100 edits here during June, and from the diff it's hard to know what has settled and what is still in contention. There's such a large volume of changed text that I think it's best to link to the appropriate sections and just give a summary of the substantive changes—in some cases "Significant changes to blah [link]".

Can people assist me by pointing out the sections that are reasonably stable and substantively changed? And any particular changes you think editors at large should be explicitly warned about, please say so now. It has to be as succinct as possible, and NPOV, of course. In the May update I think I alluded to instability but quailed at grappling with specifics. The June update will be published in the Signpost on 7 June, and I'll start preparing it in the next few days. TONY (talk) 04:31, 28 June 2008 (UTC)

Here are the previous updates. TONY (talk) 04:33, 28 June 2008 (UTC)

Thanks; it's an area I'll flag as having been completely recast, with a link to the actual new text in MOSNUM. Would rather not refer them to the complex history of discussions. TONY (talk) 08:09, 28 June 2008 (UTC)

External authorities

I find people here far too willing to refer to a narrow range of American sources, particularly CMOS. Can I say that you'll get more traction here if you occasionally refer to non-American authorities, too? For all its worth, CMOS degrades itself by (1) being extraordinarily unwilling to update itself, and (2) not following its own rules, and (3) making highly questionable dictums. TONY (talk) 04:25, 28 June 2008 (UTC)

  • I realise we would have more colourful discussions here if we did. ;-) Greg L (talk) 06:03, 28 June 2008 (UTC)
Tony, what do you mean by "highl questionalbe dictums"? If you mean they adopt a position that is contrary to what most other publishers adopt, that would indeed be highly questionable. If you mean they make an arbitrary choice from several acceptable alternatives, there is nothing wrong with that, because they don't purport to describe every acceptable usage; they purport to set forth usage for the University of Chicago Press. --Gerry Ashton (talk) 12:40, 28 June 2008 (UTC)
They're alarmingly reluctant to update as the language evolves. And I'm suspicious of styleguides that make lots of money from selling hard copy, and thus not allowing instant, free, wide access. Formatting and punctuation in bibliographies is a particular gripe of mine.
But it's better than the APA guide. And don't get me wrong: there's a lot that's good in CMOS. Here, I'm finding that CMOS is mentioned almost exclusively as an external authority. This is inappropriate. TONY (talk) 11:27, 29 June 2008 (UTC)
Oh, and on the APA guide, which is treated in god-like terms by psychology and even related fields such as sociology: what gets up my nose is how much money they make by making superficial alterations every few years, getting every library and department on the planet to buy the new edition, but persisting with crappy, outdated requirements in many respects. People "love to hate it". Well, that's not good enough. TONY (talk) 00:48, 1 July 2008 (UTC)

hard-space sentence

Anderson, do you really think this is well-constructed?

Wikipedia recommends using a non-breaking space (also known as a hard space) when a line break would otherwise displace elements to a new line that could be awkward at the beginning of a line:

  • "recommends using", strictly speaking, is ungrammatical. It needed to be properly nominalised.
  • What is "otherwise" doing there?
      • Clarity rarely hurts. Septentrionalis PMAnderson 22:42, 29 June 2008 (UTC)
        • That misunderstanding is one of the reasons you're a bad writer (loose approach to redundancy). TONY (talk) 00:44, 30 June 2008 (UTC)
          • If it improves communication, it isn't redundant. Failure to recognize this is the mark of one school of incompetent writing; the type case is late Kipling. Septentrionalis PMAnderson 20:04, 30 June 2008 (UTC)
  • "line" × 3—clumsy repetition.

Now it has become:

Wikipedia recommends the use of a non-breaking space (also known as a hard space) when necessary to prevent the end-of-line displacement of elements that could be awkward at the beginning of a new line:

You've more recently added the "when necessary". Why? It says "could be awkward", and it says "recommends". Who would insert a hard space that wasn't necessary in this context? The two additional words are quite redundant. TONY (talk) 11:40, 29 June 2008 (UTC)

Lots of people. If every FAC reviewer were as reasonable as Tony, it might not be necessary to specify that this is not "always make the space before endash non-breaking". But if we do not make clear that this is not what is meant, Tony's wording will be so read; unless there is substantive disagreement, where necessary should be restored. Septentrionalis PMAnderson 22:42, 29 June 2008 (UTC)
  • My 2¢? I never would have noticed such a nuance if you two detail-oriented editors hadn’t brought it up. But logically and grammatically, I think if the “when necessary” is to be included, the text would read best if it concludes with “…would be awkward…”. If there is no “when necessary”, then “…could be awkward…”. I like the former construction here (…when necessary if it would…). In other words, either this:

Wikipedia recommends the use of a non-breaking space (also known as a hard space) when necessary to prevent the end-of-line displacement of elements that would be awkward at the beginning of a new line

or this…

Wikipedia recommends the use of a non-breaking space (also known as a hard space) to prevent the end-of-line displacement of elements that could be awkward at the beginning of a new line

I’ll so revise to the more declarative (topmost) option. If someone doesn’t like it; be my guest to revert. Greg L (talk) 00:30, 30 June 2008 (UTC)
  • "When necessary" is redundant, but occasionally it's not worth arguing with Manderson. Nothing he likes more than a fruitless argument. TONY (talk) 00:44, 30 June 2008 (UTC)

Lightbot and Lightmouse - issue with years

It would appear that there needs to be an alteration to the MOS here. Lightmouse's bot has been dewikifying years per this MOS, and I disagree that this should be happening. It is something of a contradiction that complete dates can be wikilinked as two seperate links (date and year), but on it's own the year shouldn't be wikilinked? That reads inconsistent to me. Moreover, the importance of the year articles on WP shouldn't be understated like this. I find the years very useful and I'm sure other users do as well. I suggest that the years be allowed to be wikilinked seperately, and this aspect of the MOS be changed to reflect it. !! Justa Punk !! 12:00, 17 June 2008 (UTC)

Oh dear. This debate is old news; it was fought and resolved mostly in 2005 and 2006. Year-only links are generally deprecated as nuisance links, and I don't want to go through all of the reasons here. They're in archives. Please see the statement in MOS main page on overlinking. Now, you may like to browse year article (although last time I looked, most were pitifully inadequate); but generally we don't want to dilute valuable links and degrade the readability of the text with blue-splash everywhere. Nothing, repeat nothing, is stopping any reader from tapping four keys into the search box to investigate a year article ... fine, but not linked, please. TONY (talk) 13:34, 17 June 2008 (UTC)

Agreed. Just my opinion; but I remember the first time I saw a linked year on Wikipedia (I had very little experience as a reader at that time). It was in an article on something or other and said “…in 1983, the researchers discovered…”. I thought, “oh, great! I can click on the link and read a blow-by-blow of what the researchers accomplished back then.” Obviously, that’s not what the links do. Instead you get random stuff like “the director of the Griffith Observatory wore a brown suit and mutton chop sideburns in 1983 and now we’ll see that on the cable educational channel forever.” (OK, I’m exaggerating). The wikilinks for years and dates are overrated and, in my opinion, you end up with over-linked articles. I never click on them. If a reader really wants an arbitrarily chosen chronology of notable events that occurred in a particular year, they can always type it into the search field. That’s my 2¢ on the matter; I am quite curious how others feel about this issue but am pleased to see Tony and I seem to be 100% aligned on this one. Greg L (talk)

Greg, perhaps you'll be pleased even further to hear that I'm with you too. Lone years should only be linked if they have particular relevance to the topic in question—which accounts of a very very low percentage of those lone years that one will find linked on this encyclopædia. That this guideline recommends the linking of dates (i.e. day, month and optionally year) inspite of relevance to the topic at hand is due to a flaw with Wikimedia. The flaw being that linking is required for the autoformatting function to work. This problem, inspite of the repeated calls of dozens of users, does not seem likely to ever be fixed. JIMp talk·cont 23:45, 17 June 2008 (UTC)

That is good news Jimp. Indeed, that first instance of a linked date that I clicked on as a newbie may well have been a specific date. Whether it was a link to a whole year or a particular date, it accomplished the same thing: nothing other than to over-link the page. Cheers. ;-) Greg L (talk) 02:28, 18 June 2008 (UTC)

It's time the autoformatting be fixed ... no, it's well overdue. Indeed the presence of date links for the purpose of autoformatting may be much of the cause of this notion that lone years (months, days of the week, ...) should be linked also (regardless of relevance to the article). I have some sympathy for Tony's position that it's better to leave dates unlinked and thus unautoformatted. Is it time to initiate another plea to the developers ... just to remind them that the problem is still here and causing a whole lot of strife? JIMp talk·cont 04:11, 18 June 2008 (UTC)

I share the views of Tony, Greg, and Jimp on this. We can plead for the developers to change the mechanism. We can update the guidance. These two activities can occur in parallel. Lightmouse (talk) 19:54, 18 June 2008 (UTC)

I see little point in going back to Brion Vibber at Bugzilla—he's not on-side with this one, even when faced with an 85-strong petition of WPians. No doubt we could get many hundreds on a petition now, but I wonder whether this shouldn't go to a different body first, so that they might make the representation. Is ArbCom an appropriate body to which to make a representation? Or Jimbo? I have no idea. But beyond that, frankly, I have no problem at all in accepting both major date formats without this failed mechanism, as long as they're consistent within each article. TONY (talk) 14:10, 19 June 2008 (UTC)

Just focussing on the issue of guidance... There are various elements of guidance but one is:

  • Links to date elements that do not contain both a day number and a month are not required

The neutral phrase not required could be changed to the more active should not generally be linked (to match a style used on wp:context). Would that be nearer to what people think? Lightmouse (talk) 17:02, 20 June 2008 (UTC)

  • Yes, “should not generally be linked.” Unless the date being linked has a special relevancy to the topic, such as Sept 11 within a topic on Terrorism, I don’t think dates should be linked. Linking dates to a random list of historical events is over-linking and discourages readers from exploration. If I really want to know on what date Charlie Sheen was born, I’ll type his name into the search field. Articles are optimally linked when they best anticipate what a the median reader being introduced to the subject would be interested in further exploring. Greg L (talk) 18:45, 20 June 2008 (UTC)
Seems to me some of these positions are getting rather pointy. Breaking the semantic web in order to motivate the developers for a bugfix is certainly not justified, no matter how annoyed we are with the bug. LeadSongDog (talk) 19:34, 20 June 2008 (UTC)
  • I’m not taking a position on bug fixes or templates or any of that because I don’t understand the issue nor do I use date templates just to format a date. (Should I?) When I write a date in an article, I write something like

The all-platinum kilogram prototype was presented to the Archives of the Republic in June and on 10 December 1799, the prototype was formally ratified…

…the Euro-style of date formating. My position is that I don’t advocate the linking of dates like this:

The all-platinum kilogram prototype to the Archives of the Republic in June and on 10 December 1799, the prototype was formally ratified…

…because on first sight, many novice readers would assume that by clicking on the link, they can get additional details of that particular historical event. All it takes is a couple of “gotchas” like that for most readers to learn never to click on these date links. I certainly don’t click on these links and it sounds like a lot of other editors here don’t either. It’s not a matter of “right or wrong”. This is my 2¢ on a grey area that pertains to how one could make Wikipedia better. It’s a judgement call.
It sounds like not linking dates creates problems of another sort. I think it’s too bad if that’s the case. Perhaps someone would take the time to explain why simply writing out a date in Euro-style like I did—and not linking it—is inadequate or creates problems in some way. Are there tools available for date formating that further improves the readers experience? Greg L (talk) 20:08, 20 June 2008 (UTC)
  • P.S. Here is a MacDailyNews article I just happened upon that spoke to this issue of linking dates. This is how a passage reads within that article:

MacDailyNews Take: Back on January 10, 2005, we took a bit of flack from some readers for our Take, in which we have always believed and therefore reprint here:

This is how I think readers believe linked dates will function on Wikipedia. But not here. Just a link to a list of minutia (and some trivialities), IMO. Greg L (talk) 20:29, 20 June 2008 (UTC)
I couldn't agree more with Greg L. TONY (talk) 13:25, 21 June 2008 (UTC)

Other relevant pages are wp:moslink and wp:overlink. Lightmouse (talk) 14:10, 21 June 2008 (UTC)

WP:CONTEXT ( a style guideline linked from WP:MOSNUM) says

Stand alone years do not need to be linked but some users prefer it,

I am a user that prefers it in some places, (like year of birth and year of foundation.) They are a good jumping off point for a reader to put a biography in a wider historical context. It is not the wiki way for a bot to delete these links at three a minute: I don't have a long watchlist but Lightbot is showing up frequently on it, often does excellent work on units, but it doesn't say in its edit summary whether or not it interfered with dates.

I would have to presume that the editors who maintain the year and date articles also like to have some relevant backlinks!

--Hroðulf (or Hrothulf) (Talk) 20:23, 21 June 2008 (UTC)

  • You raised a good point Hroðulf. It seems like limited year linking as you like to occasionally do (for particularly notable events like year of birth and year of foundation for historical context) could be worthwhile and informative. I can pretty much guarantee you that novice readers of Wikipedia at first anticipate date links to be ones that take the user to additional details of that particular event, as demonstrated in this example:

Tthe United States of America had been pulling away from England for many years and, on July 4, 1776, finally declared its independence.

My point is, if one is going to link dates (I tend to think there are better ways), they should function, IMO, as I showed above because this is how such links are most often done on the Web. Basically, I see linked dates like this…

…For instance, the United States of America declared its independence from England on July 4, 1776

…as a violation of WP: Principle of least astonishment. I don’t think I can come up with any suitable alternatives that fix this problem. For instance, this solution seems cumbersome:

For instance, the United States of America declared its independence from England on July 4, 1776 (click here for other notable events on July 4, 1776).

My hunch is that other editors will find that alternative as cumbersome too. But maybe not. The above technique clearly addresses the issue of least astonishment; the reader knows exactly what they will be taken to when the click. But I prefer this way:

For instance, the United States of America declared its independence from England on July 4, 1776.

In my mind, this best adheres to the principle of least astonishment and avoids over-linking and an intrusive invitation to click on a random list of events that happened during a particular date and year. Note too that linking a year (1776) might make sense “sometimes.” But the concept of linking the date (July 4) seems utterly meaningless to me. There clearly is no historical context of what happened last year on July 4th v.s. what happened on that date over 200 years ago. Arguing that linking a date provides historical context (not something Hroðulf is proposing) seems a non sequitur to me. It really amounts to nothing more than trivia. Ergo: over-linking.
The benefit of adhering to the principle of least astonishment and of anticipating what the median reader who is researching that subject will be interested in further exploring, is that we encourage exploration and learning. We do so by not teaching readers that they will often waste their time if they click on a link. Nor do we bombard them and numb them with too many blue text links. People who surf the net develop excellent “brain filters.” Someone who is researching Thermodynamic temperature doesn’t need to be bombarded with links that are as common as “the air that I breath.”
The excessive linking of dates we’ve seen over the previous year is not good, IMO. I’ve had I.P. editors come in and link every damned date and year in articles I’ve been working on. The underlying motivation for doing so seemed to be nothing more than “the tool exists, ergo I link.” Given the collaborative writing environment of Wikipedia, the tendency will always be for other editors to go ape-shit with date linking.
Perhaps I’m wrong though. Maybe a MOSNUM policy can be written that spells that what Hroðulf does—limited context linking for particularly notable events (like year of birth and year of foundation)—is OK. I certainly wouldn’t link the date (that would be pure trivia); only the year. I personally don’t see the advantages of this limited use as being worth the fuss that will no doubt arise from abuse of the tool. Greg L (talk) 22:06, 21 June 2008 (UTC)

Featured articles do not generally have links to date fragments. I am not sure what conclusion we can draw from that, but I think it is interesting. Please take a look. Lightmouse (talk) 11:55, 22 June 2008 (UTC)

As is often the case the discussion here seems largely to consist of a few people congratulating each other on how much they agree with each other about how right their opinion is. [Lighten up a little. We just got past a four-month-long contentious battle and are mending fences. Greg L (talk) 19:17, 22 June 2008 (UTC)]
Personally, I don't have a strong opinion one way or the other about dates. I do find the obsessive linking of all dates a little odd, but I also find the idea of rooting out all occurrences of linked dates in some sort of Stalinist purge equally strange.
On the whole I don't find that anything is deeply harmed (or 'diluted' or 'degraded') by these links. If fact I find them easy to ignore. I also don't find anything astonishing about a link taking you to a page about the linked text. Isn't that what every other link does? [Not necessarily. The reader is supposed to be able to properly anticipate what the link will take them to. That’s what Wikipedia’s “principle of least astonishment” means. Because of the way most other Web sites treat the linking of dates, readers who are new to Wikipedia typically expect different behavior out of linked dates until they learn what they really do (after which, they are pretty much never clicked on again). Greg L (talk) 19:17, 22 June 2008 (UTC)]
What I actually object to strongly though, is Lightbot removing all links even though the policy says they are allowed if appropriate in context. How on earth is a bot to judge context?
Moilleadóir 14:09, 22 June 2008 (UTC)
I don't know how the bot works, but I wouldn't mind much if the bot only looks for articles with a large number of lone years, a majority of which are wikilinked. I'm pretty sure we can say that such an article is overlinked, and if an editor has to manually reinstate the contextual occurences (knowing that the bot will thenceforth leave the article alone), that's a small price to pay, as it would be much less work than manually delinking all the years. Needless to say, the best solution to date overlinking would also include changing autoformatting so as to not depend on wikilinking. -- Jao (talk) 05:33, 23 June 2008 (UTC)

I have to say that I strongly disagree with what Lightbot is doing. If, say, the article says a writer moved to a new city-state in 1473, then I do want to be able to see (if I choose to) what was the situation of the world in 1473. Who ruled where? What wars were ongoing? What was happening in the 1470s generally, and what did the map of Europe look like? This is the kind of material I can find on the 1473 page, or one link away.

Certainly, an article shouldn't be overlinked. But this is a matter for editorial discretion on an article-by-article basis.

It is manifestly not appropriate for an automated 'bot to simply bulldoze all year links regardless.

My view is that WP's year pages are useful; and therefore they can be useful to link to. If somebody put the whole lot up for AfD, that proposal would be shot down in flames. So we shouldn't be removing the links to them wholesale either. Jheald (talk) 10:14, 23 June 2008 (UTC)

  1. Could Lighbot please stop blindly unlinking years and dates until this debate is resolved and a new consensus is implemented in WP:MOSNUM?
  2. I don't think editors of the Main Page Selected anniversaries series will be pleased to hear that 'in this year' and 'on this day in history' are trivia inappropriate for an encyclopedia

--Hroðulf (or Hrothulf) (Talk) 10:40, 23 June 2008 (UTC)

  • Hrothulf: You still aren’t “getting” the point, or you are mischaracterizing our position. When you state “2. I don't think editors of the Main Page Selected anniversaries series will be pleased to hear that 'in this year' and 'on this day in history' are trivia inappropriate for an encyclopedia”, no one here takes such a position. Making the information available on Wikipedia is fine. Our position, as clearly stated above, is that our problem is the way the info is linked-to within articles is link behavior that is contrary to the way visitors to Wikipedia expect such links to behave. If you still don’t understand what we’re talking about, please take the time to read my 22:06, 21 June 2008 post, above. The sooner you address the real point of what we have a problem with, the sooner we can make progress here towards a consensus. Greg L (talk) 19:54, 23 June 2008 (UTC)
  • I think I don’t disagree with limited years being allowed. A bot with Terminator-like determination and lack of selectivity may indeed be unwarranted. I certainly think the issue of how to handle years should be further discussed before a bot is let loose on them. Personally, I think the principle of least astonishment would have year links be done per my third example box in the green section, above.

    I would suggest however, that the bot be turbocharged to hunt down dates. Here, at this example of an episode of Family Guy, is a perfect example of what should be deprecated, IMO, from Wikipedia. Note the very last seven characters of the Notes section. If you click on it, you don’t go to an account of Tim Russert’s passing; you are taken to a list showing, among other things, that “[on this date in] 1525 - Martin Luther marries Katharina von Bora, against the celibacy rule decreed by the Roman Catholic Church for priests and nuns.” (*gag*) By the way, I already fixed this entry. Greg L (talk) 19:40, 23 June 2008 (UTC)

The most important thing if we're not to confuse newbies is consistency. So across Wikipedia a link that reads June 13 should consistently go to June 13. That's the way you minimise astonishment, across all the clicks on all the articles on the site as a whole. If we are going to link to an event, we should choose link text that identifies that event, eg "Tim Russert, who had passed away on June 13", and absolutely not "Tim Russert, who had passed away on June 13".
I am deeply dubious of your assertion that a newbie would expect the latter to link to Tim Russert#Death. But even if they did, overall, confusion is minimised if such links never link to articles like Tim Russert#Death. Then, at worst, a newbie might make such a mistake once, or twice. But after that, if we are consistent, he/she will realise, and then will never be confused again. Jheald (talk) 20:35, 23 June 2008 (UTC)

There is one way of measuring the acceptable rate of linking of date fragments. Just look at the rate in Featured Articles or Good Articles. It is very low. The overall quality of Wikipedia articles would rise (almost by definition) if non-GA/FA status articles had a similar low rate. Lightmouse (talk) 22:29, 23 June 2008 (UTC)

  • Jheald, (I fully agree with Lightmouse, by the way): That you’re dubious about the way date links on the Web are done doesn’t surprise me. Many Wikipedia editors get really used to the Wikipedia-Way™ of doing things and forget how the rest of the Web works. Here is a common way dates are linked:

MacDailyNews Take: Back on January 10, 2005, we took a bit of flack from some readers for our Take, in which we have always believed and therefore reprint here:

Where else but on Wikipedia does a linked date ever take a reader to a list of trivia? Yes, I agree with you that if Wikipedia were consistent, then that would fix some of this confusion. But initial confusion isn’t the entire issue. I agree with many of the others here: If a reader wants to find out what happened on a specific date, like June 13th (a date which includes such interesting tidbits like how some priest married a nun back in the 1500s), they can type it into the search field. The same goes for links for entire years. If a reader really is interested in what might have occurred on 1799 when they are reading about how a kilogram prototype was promoted to “kilogramhood”, they can enter “1799” into the search field. This isn’t about whether maintaining lists of trivia that occurred in various years or on various dates is a good idea for Wikipedia, it’s whether the typical reader is interested in clicking on them often enough to make it worthwhile adding yet another link to the link clutter. It comes as no surprise to anyone else here that there are editors who help edit and maintain this historical trivia data (thank you by the way) and you like to see links to it used in articles. I never, ever click on these links and many of the other editors weighing in here feel the exact same way. I’m done arguing over yet another issue for which there is no clear right or wrong—just shades of gray and judgement calls regarding what makes for the best user experience. Goodbye. Greg L (talk) 22:55, 23 June 2008 (UTC)
Actually, as context I think it is quite useful to know that 1799 was at the height of the Napoleonic wars, ten years into the French revolution, the year Napoleon made himself sole ruler, and also the year George Washington died.
For what it's worth, I'm not an editor of the "year" pages. But I very much appreciate that they are there, and the pages that they are gateways to. *You* may not click on such links, but there are others who do. Oh, and I do hope you are being intentionally ironic, dismissing Luther as just "some priest". Getting married may not be the absolutely most important theological schism between him and Rome; but it's a fairly sharp example of his rejection of Rome and all its traditions and authority just the same -- with reverberations that still ring down to the present day. Jheald (talk) 23:53, 23 June 2008 (UTC)
The removal of links to years (1989, 2004, etc.), centuries (15th century, 16th century, etc.), and "s" years (1800s, 1980s, etc.) that Lightbot is currently undertaking should be discouraged, in my opinion. 1) There are way more pressing needs for bots than this removal of content on wikipedia, 2) the links to dates provide a valuable function to those viewing the articles, and 3) the method that has been undertaken for this measure seems highly suspect. The actions of this bot are a net negative for wikipedia. Cardsplayer4life (talk) 22:56, 23 June 2008 (UTC)

Greg L, that you persist in your idiosyncratic interpretation of the principle of least astonishment somehow doesn't surprise me. Please at least acknowledge that a counter-argument exists, i.e. that a link taking a user to a page about the link text is both logical and unastonishing. This is how most links here work. In fact, links on the web work in a variety of interesting and wacky ways and following web practice could very well violate the principle.

Several people have taken the trouble to visit this page and say that they don't think that the current actions of Lightbot are appropriate. So far, it doesn't seem like anyone is listening. To continue overapplying an extreme interpretation of the policy in this situation shows at the very least a disinterest in working collaboratively, which is neither appealing nor encouraging.

Moilleadóir 07:17, 24 June 2008 (UTC)

What is the purpose of articles like 'Audi' having a link to 21st century? Lightmouse (talk) 00:16, 28 June 2008 (UTC)
Beats me. Your point? ☸ Moilleadóir 03:38, 28 June 2008 (UTC)
His point is that such utterly gratuitous and irritating bright-blue splashes should be delinked; please do, Lightmouse. TONY (talk) 04:13, 28 June 2008 (UTC)
Careful, some ungenerous soul might detect a hint of personal preference in that statement!
Seriously, my original point was about a bot's inability to judge context, not about the validity of every possible permutation of date-linking.
Just because I think Greg's argument about the 'astonishing' nature of date links is feeble, doesn't mean I'm a rabid date-linker. I just don't think it's such a huge threat to WP that it's worth a small number of users here upsetting a potentially large number of users 'out there'. And certainly not if the motivation is a personal aversion to the colour blue. :)     ☸ Moilleadóir 09:09, 28 June 2008 (UTC)

AFAIK there never was a broad consensus that delinking of solitary years could be performed by a bot (not even by semi-bot or script-assisted editing), irrespective of whether the application was fully automated or manually controlled. I took part in the 2006 discussions on this topic: back in those days one was left the choice: stop that kind of editing, or leave. Some left (e.g. Special:Contributions/Bobblewik).

Well, I thought maybe times have changed and consensus swept in support of such delinking activities. Apparently not, seeing the reactions to Lightbot's edits.

Re. Greg L's argument: disagree, several websites, specifically general data repositories like Wikipedia (e.g. IMDb) use date links linking to general data about what happened on the date, not the MacDailyNews way (which would be a "surprise link" in Wikipedia context). Principle of least surprise all depends on where you're coming from in this case, so I don't see how this would play any role in the discussion we're having here.

I agree with Moilleadóir above. Probably MOSNUM should make clear again (as it used to at the end of the 2006 debates) that the Wikipedia community is not receptive to these kind of automated delinking actions: this would avoid frustration for those engaging in this type of delinking on bot scale, only to run against community frustration shortly thereafter. Lightbot's task description (currently Wikipedia:Bots/Requests for approval/Lightbot) should reflect this.

Anyway, it could be argued that Lightmouse has his bot perform task not covered by the current approval (Wikipedia:Bots/Requests for approval/Lightbot), e.g. for these kind of edits Lightmouse writes on the approval page "... I have done thousands of these and I know what to check for." - well apparently not, seen the expressed disagreements. So, disengage from this part of the tasks or stop the bot.

Further, I checked Lightbot's testrun of 100 edits. I couldn't find a single solitary year delinked by the bot during its testrun. That may be accidental, I only want to say: that part of the tasks was never part of the approval process. So I'd recommend Lightmouse to submit an "additional task" request for approval specifically for the delinking of solitary years.

Also, centuries (15th century, 16th century, etc.), and "s" years (1800s, 1980s, etc.) are not "date fragments" by any conventional use of the English language, so the delinking of them is currently not covered by the 4th (nor by any other) task of Lightbot's approval. --Francis Schonken (talk) 11:33, 28 June 2008 (UTC)

I find it difficult to address a potential user complaint. It is easier to deal with real complaints and a real articles. Complaints about real articles do exist but we have not had a large proportion. Featured Article status is the best that Wikipedia can offer. It should be the measure by which all articles should be judged. So the judgment of people here can be bypassed if anyone wants. Simply put the affected article forward for Featured Article status and see if the FA reviewers say it needs links to date fragments. Lightmouse (talk) 11:57, 28 June 2008 (UTC)

Re. "I find it difficult to address a potential user complaint" – I kind of sensed that, but that is a capacity vital in someone setting up a bot. In other words, "... I know what to check for" was just bluffing. I insist that the approval of Lightbot be withdrawn until these issues are agreed upon. As for the BAG, I'd advise them not to grant approval to bots runned by those finding it difficult to address potential user complaints. --Francis Schonken (talk) 12:30, 28 June 2008 (UTC)

Francis Schonken, just let me into the secret of why you feel, on some deep emotional level, attached to year, decade and century links. I suspect that you are a very conservative person who is loath to accept any major change on this fluid fast-adapting, wonderful online product. I haven't yet heard that cogent argument from either you or Rebecca, who appears to have a dog-at-a-bone attitude to this issue yet more trenchant than your own. Now, I hope you don't take this as a personal attack, because it's not meant in that spirit: I am genuinely trying to understand what appears to be hardline inflexibility, and your postings thus far don't explain this to me, or, I suspect, others here.
I first came up against this phenomenon when you refused to allow what I saw as copy-editing improvements at ... where was it ... naming conventions (unsure now), where you may have had ownership issues. You were certainly not cooperative in discussion, I felt. But again, my purpose is not to attack, but to prompt a higher level of debate here, and to gain insights into each others' quite contrary stances.
You know where I'm coming from, I think. As a reading psychologist (that was my dissertation topic), I see bright-blue splashes all over the place, which (1) makes WP's pages slightly harder to read, (2) makes the pages messy in appearance, beyond a certain point, and (3) dilutes the high-value links, making it less likely that readers will follow them.
The situation would be a whole lot better if the default colour for links was much less obstructive—a pastel blue rather than bright blue, without underlining—with the option to change this in user preferences. But because WikiMedia itself is a conservative organisation, of not sufficiently dynamic or well-funded, technical issues seem not to be dealt with. Someone needs to donate $200K to them to upgrade their programming, frankly. So for the time being—possibly in the long term—we're stuck with an inflexible linking (and autoformatting) system, and this is why I strongly support a move away from what I see as trivial linking:
  • years, decades, centuries that very rarely lead anywhere but to a nuisance page, with the possible exception of historically early chronological items, provided the destination page is in a reasonable shape (often not), and some piped links (although there lies a problem in itself);
  • the names of anglophone countries such as the US, the UK, Australia and NZ, and of other countries that are likely to be very well known to almost all readers, such as France, India, China, and Germany (again, with obvious exceptions, such as in the article "Economy of Germany");
  • dictionary-type words.
There are compelling reasons to take a much more disciplined and considered approach to linking; to resist the temptation to link because you can (which I've fallen prey to in starting up articles, I admit). I'm referring to both the changing of our current habits and to retrospective changes. So Lightmouse should be applauded for his active, dynamic approach to lessening the problem and making WP a better place. I presume that he checks over the results carefully. TONY (talk) 12:57, 28 June 2008 (UTC)
Do we need links telling us when some random actress or actor was born?
  • Ditto everything Tony said. I would only add that if the article is on a historical subject (Napoleonic weaponry and warfare for instance), then links to years, decades, and ranges of years make sense and can certainly be left alone if they are linked judiciously. I don’t know what Lightmouse’s bot does with links like these. I couldn’t be more pleased with what his bot is doing to years (and especially dates) in regular articles on current topics. If I want to go to October 16 to discover fascinating trivia, like how Angela Lansbury “was born on this date in 1925”, I’ll type “October 16” into the search field (which I’ll never do). Greg L (talk) 23:31, 28 June 2008 (UTC)

Nah, can't speak for Rebecca (and have no hunch of her psychology), but the psychological analysis presented above misses the mark on every imaginable level. Sorry, couldn't say whether its obvious inadequacies are anywhere linked to career crash frustrations. Nor do I see any merit in discussing levels of conservatism with someone who has the old wig listed first as an inspiration. Passons. No offense intended, but that trail does not lead anywhere near the explanation you're looking for.

Wikipedia is a consensus-based project built primarily by humans who take pride in their work. That means that automated actions interfering with that work need to be based on a quite broad consensus. I only see that broad base missing for serialised delinking of solitary years. Still. I had seen it going on over the last few weeks, and didn't speak: as said I thought maybe the climate had changed. The reaction by other Wikipedians made me aware that things were still pretty much the same as when we discussed this a few years ago. --Francis Schonken (talk) 18:17, 28 June 2008 (UTC)

You could easily be accused of inflexibility too Tony, but this is hardly the point. Playing the man is generally considered bad form on Wikipedia, even if you claim afterwards that your derision of someone's position isn't "meant in that spirit".
Unless Francis has you under some kind of conservative mind control, no one is forcing you to click on those links. It is also possible to change their colour if they bother you that much. ☸ Moilleadóir 10:23, 29 June 2008 (UTC)
  • Moilleadóir, if the links were in this form:

The assembly voted to adopt the new measure on 16 October 1799 (other notable events of 1799)

…then I wouldn’t have a problem with it (if they were few and far between and limited to historical articles). I know there is a difference of opinion about the “principle of least astonishment”, and differences as to our view of the basic facts regarding what is “normal” behavior on the Web, and even whether it is reasonable to expect that all readers “learn and remember” Wikipedia’s behavior with links to dates. We’ll have to agree to disagree on these facts. But my main problem is with articles that have many dates in which I.P. users go ape shit linking every single one of them until a whole sentences are blue. But dates are not currently linked as I’ve proposed above, and further, all the current links do is take the reader to a damned list of random trivia.
Trivia ↑ (what does this have to do with an assembly vote held on 16 October 1799?) No one gives a dump.
Those of us here simply think that “trivia is too trivial” to bother cluttering up an article with links to it. You argue that “no one is forcing you to click on those links”, but, IMO, you are missing an important point. Users of the Web develop “brain filters” to focus on novelty and the information they are seeking. The brain learns to ignore extraneous information like advertisements and sidebars and in-line links to vendors, such as when they come across the term hard drive. If you bombard users with too many links, they stop seeing them. The sooner editors understand this important psychological behavior of the human mind and how one presents a properly crafted page of information that doesn’t over-stimulate the brain and invites exploration, the better off Wikipedia will be.

I honestly believe that many I.P. users who’ve gone wild linking the crap out of every damned date, have just been smitten with the power and excitement of HTML markup and Wikilinking and just need to go buy themselves a Commodore 64 so they can do some programming in BASIC and get it out of their system. Better that, than turning Wikipedia pages into one big turd of blue. Hooray for the bot! Greg L (talk) 17:56, 29 June 2008 (UTC)

No offence taken; I don't offend easily, which is a good stance in highly charged debates on a Wiki. Francis has had a good sniff around my talk page, so he will have reaslised that I have changed the colour of my links; as always, I'm more concerned about what our readers out there see. People have queried both the number and visual appearance of links on WP, unprompted, when the conversation turns to my active role on WP. They think it's odd to say the least, The old wig is among other likes and dislikes that are on the radical side;liking the music of the 18th century isn't a mark of conservatism, and nor is liking contemporary rock bands a mark of a willingness to change things.
So I still don't understand whether you people object just to the method of reducing trivial links, or are attached to the said links per se. TONY (talk) 10:44, 29 June 2008 (UTC)
Greg, love your analogy with Angela Lansbury, and I think "one huge turd of blue" needs to be carved into stone! TONY (talk) 03:39, 30 June 2008 (UTC)
Tony, not sure about the others, but I know I object far more to the method of removing links rather than the removal themselves. As I've stated below, at #Autoformatting purposes, I don't think the wikilinking of dates is a good idea, autoformatting notwithstanding. However, I can accept that the wikilinking of dates MAY sometimes be helpful, but it would take a human eye to discern those cases. So I object to a bot removing the wikilinked dates, but I do not object to human editors removing them. (However, if consensus decides that NO dates should be wikilinked at all, then obviously a bot could do that just fine. I don't think consensus will go quite that far, however.)--Aervanath lives in the Orphanage 14:29, 1 July 2008 (UTC)

date autoformatting is optional

I'd like to remind users that for some time now, the autoformatting of dates has not been required.

There are four advantages in not linking dates:

  1. Inconsistent raw formatting within an article is obvious to editors and thus less likely to escape our attention. (The autoformatting mechanism conceals the inconsistencies from us, the very people who are most likely to enforce consistency, but the raw formats are displayed in bright blue to almost all readers, who are not registered and logged in. The rules for the choice of format in an article are in MOSNUM, here); they are easily summarised as (a) be consistent within an article; (b) take account of national ties to a topic; and (c) retain the existing format unless there's a good reason not to.
  2. There are fewer bright-blue splotches in the text, which makes it slightly easier to read and improves its appearance.
  3. The following issues concerning the dysfunctional aspects of the autoformatting mechanism do not arise:
    • piped links to date elements ([[20 June|20]], [[20 June]] [[1997 in South African sport|1997]]) (several forms of piped links break the date formatting function);
    • links to date ranges in the same calendar month e.g. December 13–17 or the night of 30/31 May – the autoformatting mechanism will damage such dates (30/May 31);
    • links to date elements on disambiguation pages;
    • links to date elements in article and section headings; and
    • links to date elements in quotations (unless the original text was wikilinked).
  4. As a minor advantage, edit windows are slightly easier to read and edit.

It may be that WikiMedia can be persuaded to invest resources in revamping the mechanism to avoid or mitigate these problems, but this is unlikely to occur in the short to medium terms. TONY (talk) 14:01, 21 June 2008 (UTC)

  • Oh… I SEE. Autoformatting of dates that adjusts to user preferences. That is quite cool and I would use it if it was a well implemented tool. But… there is no doubt about; having the autoformatting irrevocably accompanied by click-to-be-disapointed blue-link text blows. So I certainly won’t use it. Whats-his-face the developer is wrong to take his “go fish” attitude in the face of so much consensus of opinion. However, I’m sure he is a volunteer contributor (just like the rest of us) and has a perfect right to participate only in what he agrees with; that is an important element of enjoying this hobby. Otherwise, it’s just toil.

    So we’re left with the choice of using the cool, oh-so-logical, Euro-way date convention (14 December 1799) or the stupid American way (December 14, 1799). Though Americans aren’t accustomed to the Euro-cool way, they are able to instantly parse it so there is zero confusion. There’s no point just flipping a coin, so I use the Euro way. A parsing template that looked towards the user’s preferences and doesn’t link to a random, irrelevant list of events would be far superior to forcing editors to make such a formatting convention choice (or using the current autoformatting and its click-to-be-disapointed™). Greg L (talk) 19:10, 21 June 2008 (UTC)

  • Greg—Just as for spelling, there are rules (in MOSNUM, actually) for which date format is most suitable for an article. They're based partly on the rules for EngVar. Brion what's-his-face was not just neutral, but slightly uncooperative. I don't think he's engaged closely in WP editing, somehow. The Bugzilla page remains open, but languishes. The American way is unfortunate in its requirement of a comma (to separate the jostling figures); but I see no reason that both systems not be used on a consistency-within-article basis, just like spelling. I think US readers might be upset to see the Chicago article in European format, and UK readers the London article in North American format. No problem, then ...
  • The only other thing we can do is take it high up the food-chain in WP. ArbCom is really only for dispute resolution; I wonder whether Jimbo might be the best port of call. The best plan might be to say to him that we got 85 signatures a year ago to get the linking and autoformatting mechanisms decoupled, and we could get hundreds now. If we did, would you be willing to make representations for us to WikiMedia's board?
  • Does that idea appeal? TONY (talk) 09:32, 22 June 2008 (UTC)
  • If you want to take the lead, I’ll help. It’s one of those things that isn’t a glaring shortcoming, but I’m pretty confident we are taking the right stance on an issue that will make Wikipedia incrementally better. As for keeping the date style relevant to the subject (it’s not 4 July 1776), I agree. I guess I’ve never added a date to a Wikipedia topic that was closely tied with particular country. I’ve focused intently to articles like Thermodynamic temperature, Kilogram, Kelvin, Parts-per notation, and Mass versus weight. Since all are universal and scientific in nature, the Euro formatting of dates seems to have been accepted and successful so far. A date tool that adjusts to user preferences would be a better option yet. Greg L (talk) 19:10, 22 June 2008 (UTC)
  • P.S. While we’re getting things done with developers, how say we also inquire into what it takes to get some character-counting parser functions made so magic words and templates can handle delimiting numeric strings without mathematics. Both {{delimitnum}} (also languishing as bz 13025) and {{val}} would benefit; both are too buggy for general use. Greg L (talk) 22:42, 22 June 2008 (UTC)

Is there already a Bugzilla request open for autoformatted dates not to show up blue? --Hroðulf (or Hrothulf) (Talk) 10:35, 23 June 2008 (UTC)

Once again, here is the link to the still-open but moribund Bugzilla link. The petition is presented some way in via a link to a WP page. BTW, someone has said that MediaWiki's developers are all volunteers: not quite true, I believe—there are one or two part-time paid developers among the volunteers. TONY (talk) 00:55, 1 July 2008 (UTC)
  • Dank55 has alerted me to the operation of a bot that is going around linking full dates. The bot text cites MOSNUM as insisting on autoformatting. I've notified the user here and invited him/her to discuss the matter here. TONY (talk) 13:23, 1 July 2008 (UTC)
  • This whole debate has confirmed in my mind that the auto-formatting, while seeming cool on first impression, is not useful and should be disabled. I have disabled mine. I think the autoformatting causes us to lose sight of what we are doing: we are creating an encyclopedia for everyone, not just logged-in users. People who have not logged in or created an account have no autoformatting option, and therefore the whole thing is moot for them. When we as editors have the option, we forget that what we see is not what everyone else sees, and therefore are one further step removed from creating an encyclopedia for everyone.--Aervanath lives in the Orphanage 15:03, 1 July 2008 (UTC)
    • Indeed, Aervanath. But disabling it would cause back-compatibility problems with millions of items all at once. And non-logged in readers (the 99%) do have to put up with the bright-blue rinse, but of course see the raw formatting. No point anyway for them. That's why I'm increasingly inclined to try to change the culture so people just don't use it, rather than trying to fix it by petition. Otherwise, give us an autoformatter for "colour/or" and ize/ise, please. All too silly; we're adults, and accept what are basically two varieties of spelling and two varieties of date formatting. Big deal. TONY (talk) 16:01, 1 July 2008 (UTC)
  • Yes. Having a feature (autoformatting) that is only for registered users is useless for the vast majority of readers. We might as well forget that the “autoformatting” feature exists. That leaves only the “virtue” (plague) of links to random events. By having such extensive detail on autoformating of dates on MOSNUM, we are essentially inviting editors to use the tool. There must be dozens of templates that are not mentioned on MOSNUM that, if used, would actually improve Wikipedia. I suggest we nearly delete the entire Autoformatting and linking section. Instead, it would be replaced with this:

Autoformatting and linking

Note that dates can be autoformatted and linked by coding as follows: [[1925-10-16]]. Depending on the preferences setting of registered users, this autoformatting feature allows some registered users to see that code displayed and linked as October 16, 1925 where still others might see 16 October 1925. However unregistered users (the vast majority of readers on Wikipedia) will simply see the code displayed as 1925-10-16. Since the autoformatting feature does not work for unregistered users, and since the accompanying linking feature almost always takes readers to articles that are lists of historical trivia, Wikipedia recommends that links to dates generally be limited to articles that are intrinsically historical in nature, such as “French Revolution,” and that such links be limited to years, year ranges, and decades, and—even then—they should be done only judiciously where the link would naturally be of likely interest to that readership. Links to dates (like 16 October) are discouraged for all but the rarest of circumstances.

That’s my suggested text for the current guideline. Greg L (talk) 01:54, 2 July 2008 (UTC)
  • Greg: nicely written indeed. It does suggest to the troops that autoformatting was designed as a link ... but when you look at Brion Vibber's opening statement at Bugzilla, he appears to confirm that he saw that as an advantage (i.e., linking years is useful, so why not couple these functions ... aargh). I hope you don't mind that I changed your "merely" to "almost always", just to avoid possible criticism. In principle, I like it. Yes, the last sentence raises the association with "On This Day" columns in newspapers and, indeed, WP's main page. OK for stimulating armchair interest, but it's diversionary entertainment, isn't it! Angela L., eat your heart out. TONY (talk) 03:25, 2 July 2008 (UTC)

Responding to the question on my talk page:

Well, first of all I'm not operating a bot. My edits are manual, assisted by AWB. Secondly, yes I am aware that some users don't like dates to be wikilinked and I know that date wikilinking is not required. I understand from MOS:SYL that it can be done optionally i.e. it is neither required, nor forbidden. I interpret this to mean that I can do it.
No, I don't know of any bots wikilinking dates in the main text of pages. I'm not, and won't be, doing this in bulk.
Now, I consider that my edits are useful, as:
  1. I'm correcting incorrectly formatted dates in template fields such as dates with ordinals, abbreviated months, dates in / format etc. which I don't believe is at all in question, as is in compliance to WP:DATE and the templates' recommended date format. I am not converting American date format to International or vice versa, nor insisting on a 'default' date format.
  2. Whether or not wikilinking of dates is wanted as part of the user date preference, by linking them as per current functionality I allow users to view dates in a consistent format. I am not converting American date format to International or vice versa, so wikilinking these dates doesn't change the displayed formats for users who aren't logged in. My edits to dates in 'cite' templates are in compliance with the recommended/supported format of the template. If it is agreed that wikilinking of dates is not wanted in Wikipedia, all these dates can be unlinked by changing a handful of templates, instead of needing to change thousands of articles.
So, if there is a Wikimedia change to date rendering, I believe my edits will assist compliance to the new format as dates I've changed will be reformatted automatically. Thanks Rjwilmsi 17:27, 1 July 2008 (UTC)
Rjwilmsi's approach highlights a conflict. Rjwilmsi seems to think that the Cite book template says to put dates in ISO format, that is sufficient authority to do so. However, this guideline says to use consistent format throughout an article, including the citations and references. The template has no parameter to specify which format is being used in an article, so there is no way it can reformat dates to match the other dates in the article. It will be correct for articles that happen to use the same format as the template, and wrong for the rest. I have started a new section to discuss this conflict. --Gerry Ashton (talk) 18:46, 1 July 2008 (UTC)

AD before or after the year is not a style choice

Using BC/AD or BCE/CE (with or without the dots) in writing historical dates is a matter of choice. However, using either system incorrectly shouldn't be. The problem in the MoS is that it states writing AD after the year is acceptable (it is not). Every style manual I've consulted (e.g. MLA, APA, Chicago,...) agrees that AD should be placed before the year. When writers use AD after the year it is usually because s/he is unaware of the convention. I hope this change appears in the Manual of Style, to educate people who care about writing correctly and to improve the quality of the Wikipedia. Omnihistor (talk) —Preceding comment was added at 07:43, 30 June 2008 (UTC)

Style guidelines should not be followed contrary to the usage of careful writers (the present politically correct term for literacy); although one does wonder why, if this isn't a stylistic choice, Omnihistor is consulting style guides. Both choices are frequent, both are commonplace, and a rule on this matter does not contribute to clarity, accuracy, neutrality or attribution. Septentrionalis PMAnderson 20:00, 30 June 2008 (UTC)
I agree with PMAnderson. Style guides' notion of "correctness" notwithstanding, both usages are frequently employed, and both are clearly understood by the reader. As long as the use is consistent within each article, both should be allowed.--Aervanath lives in the Orphanage 14:33, 1 July 2008 (UTC)


{{Convert}} has been criticised for its use of the siemens per tesla symbol for the short ton. However, the Short ton, Long ton and Tonne articles do mention "S/T", "L/T" and "M/T". These, however, were added in January. The use of a solidus for abbreviations of units where no division is involved is somewhat odd. The forms "ST" and "LT" are much clearer. As for "MT", I wonder whether we should allow this at all. It's only in North American (US only?) English that the tonne is called a "metric ton" and the NIST uses "t". I propose that these abbreviations be added to the list of abbreviation to be avoided. JIMp talk·cont 03:29, 19 June 2008 (UTC)

  • The siemens/tesla (S/T) is entirely correct and is in conformance with the SI and should not be modified or degraded in some way to avoid conflicts with “short ton” / “long ton (tonne)” issues. Besides, ST (for short ton) and LT (for long ton) aren’t the correct unit symbols; the proper symbol for tonne is lowercase t (see the BIPM’s SI brochure, Table 6 (Section 4.1). Note further that besides the uppercase S being the symbol for the siemens, the lowercase s is the unit symbol for the second. So I believe it is best to not have any form of the letter S be used as an abbreviation for “short”—both s and S are in use already. And, indeed, there is mathematical division in “S/T” if one really means one siemens per tesla (just as one joule per second is J/s, which is one watt). It will be a holy war (that I don’t want to touch with a ten-foot pole) to effect change on such a well used unit of measure, but a reading of the BIPM brochure shows the following possible solutions:
    • metric ton = tonne = t = 1000 kg
    • short ton = 2000 lb (no use of the symbol t, no abbreviations, and no use of “ton”, all of which would be ambiguous)

Note also that Mt (for megaton), while rather unfamiliar to lay people, is entirely in conformance with the SI and is commonly used as a measure of greenhouse gas emissions and other environmental contexts. My suggestion would be to never try to ban the use of an SI-compliant unit of measure or unit symbol; non-SI-compliant abbreviations should yield first. Also, much confusion would be spared if editors weren’t so quick to use unit symbols. The best practice, in my opinion, is to spell out the unit of measure on the first and maybe second instances and—if the measure isn’t used too frequently, on all occasions in an article. Only if it the unit of measure is used many many times in an article, should the unit symbol be used—and even then—only after it has been introduced in the fully spelled out form. This is the way Encyclopedia Britannica does it because it follows the basic fundamentals of technical writing.
Greg L (talk) 03:53, 19 June 2008 (UTC)

Agreed: Yes, the non-SI complant should go first ... what do we do about metric horsepower vs petasiemens? I'd be happy to be rid of all six ton abbreviations; "S/T", "ST", "L/T", "LT", "M/T" and "MT"; on WP. Let the long and short tons be spelt out in full each time (even in infoboxes) and allow only the symbol "t" for the tonne. I don't know how familiar those abbreviations are, the spelt out forms are clear. It is best, in my view also, to spell units out though there are a few other exceptions besides those which are frequently used. Abbreviations/symbols should be prefered in infoboxes and in parenthetical conversions also if the unit name is long (e.g. "miles per imperial gallon"), the abbreviation/symbol should be prefered. Of course, there is another issue I'm raising here: that of the use of the solidus in unit abbreviations/symbols. I'm suggesting that this be restricted to the indication of division (per). JIMp talk·cont 04:18, 19 June 2008 (UTC)

  • As to your last point (the solidus), that is exceedingly common and quite proper. The velocity “one meter per second” is “m/s”. You’ll find this use everywhere. Density in symbolic form is mV and in actual units is g/ml (or g/cc). The only alternative is to use reciprocal units like m·s−1 or just ms−1 but that notation is generally only used in highly technical or scientific articles (and too frequently, IMO, here on Wikipedia). Greg L (talk) 04:33, 19 June 2008 (UTC)
  • P.S. As to the rest of you questions, I would suggest the following guidelines for editors:
  1. Consider the target audience and ensure not to write over their heads.
  2. Stay compliant with the SI unless a discipline consistently uses non-SI units.
  3. Spell out units of measure as much as possible. Resort to unit symbols only if they appear exceedingly often in the article, and even then, introduce them well with several instances of spelled out units.
The only possible discipline I can think of that might use petasiemens would be the bleeding edge of superconductor research. So if Wikipedia has an advanced article directed to such a high-level audience, and if that discipline really is using this fully-SI-compliant unit of measure, I don’t see a problem. Greg L (talk) 04:45, 19 June 2008 (UTC)
Something that I had not seen until recently is the use of the solidus in labeling the axes of graphs. For example if the horizontal axis of a graph is time (measured in seconds), and in the text the variable t is being used for time, the axis might be labeled t/s. I avoid this because while a few people may be trained in this usage, many other people will be searching through the text, trying to figure out where the variable s is defined. --Gerry Ashton (talk) 04:51, 19 June 2008 (UTC)
  • I think too many contributing editors read a technical paper or article somewhere and don’t have the confidence to revise the axis legends for greater clarity for a general-interest audience. While down at WSU talking with the Ph.D. on the medical device we’re working on, I noticed he had a number of books on technical writing. Ph.D.s do a lot of that. One of his books was titled “Why not just write it clearly?” (or something to that effect). Greg L (talk) 14:27, 19 June 2008 (UTC)
  • We cannot do away with the solidus, no, I'm saying we do away with it in unit abbreviations that don't involve unit division. Long tons are long tons not longs per ton, so "L/T" should be out. Kilometres per hour, on the other hand, can go km/h.
  • I always used to label my graphs like that but I haven't added any graphs to WP.
  • No, the petasiemens vs metric horsepower ambiguity of "PS" is hardly worth troubling about.

JIMp talk·cont 05:50, 19 June 2008 (UTC)

  • Oh, I get your point. Sorry. Indeed, we should certainly loose “L/T” for “long tons”. Yes, there is no division. “L/T” breaks so many rules and conventions. According to the BIPM, SI brochure, Table 6 (Section 4.1):

In English speaking countries this unit is usually called "metric ton".

And as I mentioned above, Section 4.1 also makes it clear that the internationally accepted unit symbol for metric ton is lowercase t; homegrown abbreviations are improper. I note that Wikipedia articles often use “tonne”. Either “tonne” or “metric ton” works for me. “Long ton” should be looked upon with disfavor IMO, and the bastard “L/T” symbol should certainly be deprecated in favor of “t”. I personally prefer “metric ton” over “tonne”, per the BIPM’s acknowledgment that it is the common term in English. Why? Everyone in Europe and the rest of the world recognizes what it is. And (bonus points) Americans also know what a “metric ton” means. But many Americans don’t know what “tonne” means and often assume it is a British spelling of “ton”. So “metric ton (t)” is understood by the widest audience, has the least confusion, and is absolutely unambiguous (which is why the BIPM acknowledges it).
As for short ton, all I can figure out is what I wrote above: • short ton = 2000 lb (no use of the symbol t because it means metric ton, no abbreviations like “S/T”, and no use of just “ton”, all of which would be ambiguous). Would you agree? Greg L (talk) 14:17, 19 June 2008 (UTC)
Agree that PS should always be petaseimens. If an article really needs Pferdestärke, it can be spelled out and wikilinked.LeadSongDog (talk) 20:40, 19 June 2008 (UTC)
Yeah, sorry, I guess calling the likes of "m/s" division wasn't that clear. One problem is that the unit Pferdestärke is actually very common in automotive articles where it's always abbreviated to "PS" ... we'd have an uphill battle convincing the editors of those articles to spell it out ... in German. Another problem is that the tonne is never called a metric ton in certain dialects of English (mine for example). JIMp talk·cont 00:36, 20 June 2008 (UTC)
I've just gone through all the Whatlinkshere/PS entries to find any Pferdestärke references and pipetrick dab them. No doubt there are still many non-wikilinked uses of PS though.LeadSongDog (talk) 04:35, 20 June 2008 (UTC)
  • Is there any reason PS can’t also be used as a symbol for Pferdestärke? Just because PS is the symbol for petasiemens doesn’t mean that an automobile-related article can't also use it does it? I can’t imagine that such widely different disciplines as automobiles and superconductors could trample on each others toes (confusion-wise). Greg L (talk) 04:56, 20 June 2008 (UTC)
That would seem to make sense, so long as template:convert (and the like) stick to the petaseimens meaning.LeadSongDog (talk) 05:46, 20 June 2008 (UTC)
I'm ok with having {{convert}} always spell out short tons and long tons. The term 'metric ton' or (sometimes 'metric tonne') is an American and Canadian term. However, as GregL points out, many of us in North America will assume, as I did many years ago when I first encounter it, that 'tonne' is just the British way of spelling 'ton'; similar to gram vs. gramme. Even the folks from Jimp's neighborhood would recognize what is meant by 'metric ton' and that's why I think it's a better choice. —MJCdetroit (yak) 03:04, 25 June 2008 (UTC)
  • Thanks MJCdetroit. To all: Note that according to the BIPM: SI Brochure: Section 4.1, Non-SI units accepted for use with the SI, and units based on fundamental constants: Table 6: “In English speaking countries this unit is usually called ‘metric ton’.” I sincerely hope that referencing the BIPM would settle the issue here. There will simply be no way to convince all editors that the terminology they’ve grown up with and are accustomed to is (or is not) the “normal” way or is “correct” or is the “best” term when referring to “tonne” or “metric ton”. I do note that my current version of Safari (3.1.1), just flagged “tonne” as a misspelling (if that means jack). As MCJdetroit alluded to above, simply calling it “metric ton” while awkward to some, will result in the least confusion since by far the greatest number of readers understand what it means. And, as I stated, if we can all lighten up on how sure we are that what our ears are used to hearing has got to be the “correct and only” way, and bow to the reality that the BIPM is saying the unit is “usually” called “metric ton” in English, then that will avoid a lot of needless bickering here. In acknowledgement of this reality, I recommend “short ton” (no symbol) and “metric ton” (symbol: t). Greg L (talk) 20:29, 25 June 2008 (UTC)

I agree with MJCdetroit and Greg L. NIST's Special Publication 811 on page 63 says the symbol for metric ton is "t", and gives no symbol for short ton or long ton. While NIST is not authoritative for the whole world when it comes to SI, they are authoritative for the short and long ton, since the USA is nearly the only country still using them. --Gerry Ashton (talk) 21:28, 25 June 2008 (UTC)

I disagree, in the UK we are familiar with the ton (which you in the USA choose to call the Long Ton) and its near metric equivalent the tonne. For example, in the context of transportation and transportation of dangerous goods 30 ton / 30 tonne are almost equivalents (within 2%); and most UK readers would not be familiar with the term Long ton. The BIPM appears to be referring to US-readers when it refers to "English speakers". So just for the benefit of the US, were are to expect the ton (Short ton) and the metric ton. My recommendation is the "ton" (long ton, no symbol), the “short ton” (no symbol) and “metric ton or tonne” (symbol: t).Pyrotec (talk) 21:51, 25 June 2008 (UTC)
Let me support Pyrotec on this. "Metric ton" might have been used in the 1970s, but it's simply no longer current in the UK. Consider the following google searches: - ton 827 - tonne 697 - metric ton 0 - ton 1,590 - tonne 2,410 - metric ton 1 - ton 102,000 - tonne 28,800 - metric ton 157 - ton 52 - tonne 103 - metric ton 0 - ton 2,840 - tonne 347 - metric ton 1 - ton 4,300 - tonne 1,360 - metric ton 25
And of course nobody in the UK would have any idea what a "long ton" would be referring to.
(Note the "ton" figures also include a lot of instances of the word being used as slang for the number 100; and in expressions such as "a ton of people", meaning lots and lots of them; and other usages. So only a minority of these hits may be using ton as a unit of weight).
So I'd support Pyrotech, and recommend the "ton" (long ton, no symbol), the "short ton" (no symbol) and "metric ton" or "tonne" (symbol: t). Jheald (talk) 23:24, 25 June 2008 (UTC)
  • I’d argue that the key test shouldn’t be which term people use in different countries (which is obviously going to vary from country to country), but which term is “understood” by the widest audience. Are editors from England really trying to say that readers there don’t understand what “metric ton” means? The trouble is that Americans are entirely unfamiliar with what “tonne” means. I’m automatically skeptical of proposal that would have Wikipeda flouting what the BIPM says. Greg L (talk) 23:37, 25 June 2008 (UTC)
I think a fair number of people would find it very unfamiliar. The quasi-SI unit is the tonne, and it is not just in the UK that that is what is used: - ton 2,190 - tonne 2,950 - metric ton 2 - ton 650 - tonne 799 - metric ton 2 - ton 2,910 - tonne 6,310 - metric ton 72 - ton 314 - tonne 282 - metric ton 6 (Canada) - ton 1,720 -- tonne 1,430 - metric ton 40 - ton 20,500 - tonne 3,050 - metric ton 431
It would be a big mistake to mandate that every instance of "tonne" must be changed to "metric ton". Jheald (talk) 00:40, 26 June 2008 (UTC)
  • Of course the official name for unit is “tonne”; that much isn’t in dispute. We’re also getting sideways on multiple issues here…

    One issue is whether the symbol for “tonne” or “metric ton” should be called L/T and various other variations on that theme (which is, after all, what is embodied in the very title of this section). Clearly, I would say ‘no’ to the use of any such unit symbols or abbreviations. According to the BIPM, the official symbol for the unit is lowercase t. We should abide by what the BIPM says and not stray from that policy unless there is a compelling reason to do so. In rare circumstances, official MOS policy flouts the BIPM. For instance, according to the BIPM’s SI brochure: Subsection 5.3.3, Formatting the value of a quantity, a space is always used to separate the unit symbol from the numeric value and this rule applies to the use of the percent symbol. Fortunately, in this case, MOS wisely ignores that BIPM rule and instructs editors to write “75%”, not “75 %”. Why is this wise? Because we are following commonly accepted and recognized conventions—the way the real world works—and don’t want to confuse our readers. I’m not seeing a compelling reason to ignore the BIPM here with regard to the unit symbol for “tonne/metric ton”; according to the BIPM, all the language-variations of Wikipedia should standardize on lowercase t for the unit symbol; no “L/T” or “LT” or “L-T”.

    The next thing to decide is how what unit name to call it by. The BIPM says “In English speaking countries this unit is usually called ‘metric ton’.” I can guarantee you that for America, that is true; spell checkers don’t even recognize “tonne” here (there’s a red squiggly line under it right now). But that isn’t the only reason I am advocating the use of “metric ton.” I am arguing that the term, albeit not the primary name in all English-speaking countries, is recognized (understood) by the greatest number of English-speaking readers. Your argument is that the BIPM’s assertion is wrong and/or outdated. But clearly, America—with a population of 300 million—has no idea what the unit “tonne” means; most assume it must be a “British specialised spelling for short ton”. The term “metric ton” is instantly recognized by Americans. And even though many British, New Zealand, Australian, and Canadian readers might chafe over how the term “metric ton” may not be standard in their area, I am confident that vast majority instantly understand what “metric ton” means when they encounter it (please point it out if I am wrong on this assumption). Once again, it’s not what unit name is a first choice preference, it’s what unit name is most widely understood.

    Finally, with regard to your statement, that it would be unwise to “mandate that every instance of ‘tonne’ must be changed to ‘metric ton’ ”, I am taking no such position. I don’t pretend to understand all the intricacies and implications of such a move. I would say that, certainly for new articles and edits where convenient, the wisest thing to do is use the least confusing “metric ton” (symbol t). Further, I am saying that to avoid confusion and ambiguity, the 2000-pound “ton” should be called “short ton” and should not use any symbols or abbreviations like “S/T” (and certainly not t, which is reserved). I advance that this all results in the least ambiguity and confusion, and further has the virtue of being entirely consistent with the BIPM’s resolutions and statements on the subject. Greg L (talk) 02:40, 26 June 2008 (UTC)

  • I absolutely agree with Greg_L that the symbol for the metric ton a.k.a. tonne is t. Not only does the BIPM say so, it says so in the official interpretation of SI for the USA (on page iii). I also agree with using the most widely understood term, even if it is a bit awkward for some. And I promise to never use the word "ton" by itself unless it is in a direct quote, and then I'll use a footnote to indicate which ton. --Gerry Ashton (talk) 02:55, 26 June 2008 (UTC)

We certainly should strive to use widely understood terms, yes. However, certain terms are just out of place in certain varieties of English. The term "metric ton" is clear and unambiguous but completely foreign in certain dialects—regardless of the assesment of the BIPM. According to WP:ENGVAR if the article has a strong tie to a given dialect, we are to use it, and if not any dialect is permitted. An article written in Australian English—it may be about uranium mining in the Northern Territory—should use "tonne" not "metric ton". The word is unambiguous, any confusion caused by certain readers' mistaking it for some quirky spelling of "ton" can be dispelled with a link or footnote (or conversion). I would therefore propose the following recommendation:

  • 1000 kg → "tonne" or "metric ton" & abbreviated to "t"
  • 2240 lb → "long ton" not abbreviated
  • 2000 lb → "short ton" not abbreviated

As for permitting the long ton to be written simply as "ton", this is ambiguous, inherently ambiguous, not just a result of some readers' unfamiliarity with spelling. An editor who comes across the word "tonne" in an article and who knows what a tonne is can quickly and easily clear the issue up for our American friends by providing a link/footnote/conversion. An editor who comes across the word "ton" in an article and who knows what a ton is, on the other hand, knows that clearing the issue up might take a good deal of digging around in sources and/or making assumptions and knows that there may even be no clearing the issue up at all because of the term's ambiguity. JIMp talk·cont 03:34, 26 June 2008 (UTC)

  • “According to WP:ENVAR if the article has a strong tie to a given dialect, we are to use it, and if not any dialect is permitted.” That makes sense to me, Jimp. We’re trying to write for maximum clarity and not all readerships are going to be the same for all articles. If it’s going to be an article on, for instance as you say, uranium mining in the Northern Territory of Australia, then editors should use terms and dialect appropriate for the readership that will most likely be coming to that article. I realise it should be “tonne” then. Greg L (talk) 06:36, 26 June 2008 (UTC)
The Economist (which has well over half its circulation in the United States) seems to think that "tonne" is sufficiently widely understood that no qualifier "metric" is understood (Economist styleguide). [Well, I read the provided link. The Economist discusses international issues. And what they wrote in their style guide is “In most non-American contexts, prefer hectares to acres, kilometres (or km) to miles, metres to yards, litres to gallons, kilos to lb, tonnes to tons, Celsius to Fahrenheit, etc.”. In other words, they have the same philosophy as we have here on Wikipedia: use SI units, only they aren’t as pro-SI as Wikipedia is because they apparently use U.S. Customary units for American-context articles. The reference you provided only reinforces my position that “tonne” is not understood in America (which it is absolutely not). Even spell checkers here in America (which have enormous capacities) don’t even recognize the word. My copy of MS Word, which has a 1999 Encarta World English Dictionary doesn’t recognize it, and my up-to-date copy of Safari (3.1.1) flags “tonne”. It would be a bad guideline, in my opinion, to routinely use the term “tonne” in our articles when the word wouldn’t be understood by such a huge number of English speakers; particularly when the term “metric ton” is understood by everyone. Greg L (talk) 13:59, 26 June 2008 (UTC)]
I think you miss my point. In all contexts apart from specifically US-national ones, the Economist recommends to use tonne. Thus, so many tonnes of global CO2 emissions. The Economist does not recommend to use metric tons; and it does not appear to be worried that its majority US audience might not understand. Jheald (talk) 10:25, 30 June 2008 (UTC)
As for "long ton", the problem is this really won't be understood outside the United States. [Are you trying to refute anything I wrote? I am in complete agreement that “long ton” should not be used. Greg L (talk) 13:59, 26 June 2008 (UTC)] The simplest recommendation would be to prefer metric units. Of course, in some cases the quantity isn't sufficiently precise that it makes much difference; or the usage is historic -- eg "several tons [or tonnes] of rubble"; "4-ton truck". If the measured quantity is exact (or at least given more accurately than +/- 20%), and there is a overwhelming reason that a non-metric ton be used, then we should perhaps insist on a metric conversion, because that would clarify exactly which ton was in question.
How serious do we think the problem is? How many examples do we think there are of cases where the word "ton" is being used, and could cause serious ambiguity that is not already readily resolvable from the context? How does WP currently use the word "ton" ? Jheald (talk) 09:05, 26 June 2008 (UTC) [It’s not an issue, in my mind, of how “serious” an issue is. I think editors get too wrapped up in validating the “correctness” or “rightness” of their country’s way of doing things and (furthermore) dislike looking at a name for a unit of measure they are entirely unaccustomed to. That’s natural. The issue here isn’t “serious” like an army general making a decision that could determine whether a thousand additional soldiers might die. The issue at hand is nothing more than whether there is a better guideline to adopt for Wikipedia’s manual of style so our articles are least confusing for the maximum number of readers. This is, however, more important than a “color v.s. colour” dispute. My position is simple: it would be unwise to use a name for a unit of measure that utterly baffles the vast majority of 300 million English-speaking readers if there is another name that baffles no one. This is particularly true, IMO, when the BIPM already says that “in English-speaking countries, the unit is normally called ‘metric ton’ ”. It’s not complex. If someone feels that “metric ton” is not only unused in their English-speaking country, but also the readers in their country don’t even understand what “metric ton” means, then speak up. I just really doubt this is the case at this juncture. Greg L (talk) 00:07, 27 June 2008 (UTC)]

1. As an aside: Version 7 of the SI brochure said:

Version 8 of the SI brochure says:

I think the change from some to usually is surprising, particularly since one of the main editors is British.
2. A web search attributes the following text to ANSI/IEEE Std 268-1992 and to the U.S. Geological Survey:

  • The term "metric ton" should be restricted to commercial usage

3. I am sympathetic to the idea that understandability/accessibility for all is more important than familiarity for the narrower interests of those within a particular domain or region. I think this is why Wikipedia prefers micrometres/nanometres over microns/angstroms.
4. Mass units are ambiguous in US-related articles but not elsewhere and the metric system is regarded as an anomaly that needs to be highlighted. This is why a special prefix word is widely used there but not elsewhere.
5. I do not know which way to go on this one. Lightmouse (talk) 01:27, 27 June 2008 (UTC)

Metric ton should be clear enough to everyone but I wouldn't go so far as prohibiting the word tonne, if it's not recognised by half the English-speaking world, this is a serious problem, but there are solutions which need not require the use of a term which has no place in the given dialect an article is written in. JIMp talk·cont 01:38, 27 June 2008 (UTC)
Regarding Lighmouse's points:
2. Without the context, we can't tell what ANSI/IEEE Std 268-1992 and the U.S. Geological Survey want instead of "metric ton" in non-commercial use; maybe they want megagram.
4. The tonne isn't formally ambiguous anywhere, but it is confusing for many Americans everywhere. If the American does not know what it means, the context of the article won't help. --Gerry Ashton (talk) 01:50, 27 June 2008 (UTC)

Yes, for point (2), they may have been giving Mg priority over tonne/metric ton. I hope so. The world is more complicated because of the low usage of units like Mg and Mm, as our current discussion demonstrates .
For point (4), I meant ambiguity between just two mass units (short or metric tons). Lightmouse (talk) 02:04, 27 June 2008 (UTC)

Incidentally, the difference between long ton and tonne is only 2%. Whereas the difference between short ton and metric ton is 10%. This should have an effect on how seriously people in the US and in the British Commonwealth treat the ambiguity of the old units. Lightmouse (talk) 02:12, 27 June 2008 (UTC)

For what it's worth Googling Australian pages turns up hits for tonne vs metric ton in a ratio of over 240 to 1. For NZ: over 160 to 1. For Ireland: almost 70 to 1. For the UK: over 80 to 1. Even for Canada the same ratio is almost 50 to 1. Worldwide: about 4 to 1. Interestingly enough, for US it's over 2 to 1 ... if I've searched correctly. JIMp talk·cont 03:14, 27 June 2008 (UTC)

What is the correct way to apply prefixes to 'metric ton'? Lightmouse (talk) 10:37, 27 June 2008 (UTC)

The aim should be to convey unambiguous information with the minimum of confusion. For most purposes I think a linked tonne would suffice. And the SI prefixes follow naturally as kilotonne (kt), microtonne (μt), etc. If metric ton is used, it is wise to avoid prefixes. The context should never be used as an alternative to disambiguation. Thunderbird2 (talk) 11:30, 27 June 2008 (UTC)

I think the preferred way to express mass is with multiples and submultiples of the gram. The tonne and metric ton are undesireable, prefixed versions of the tonne are more undesireable, and I have never heard of a two-word unit taking a prefix—megacubic metres, anyone? --Gerry Ashton (talk) 12:50, 27 June 2008 (UTC)

Depending on context, I'd consider either megastere, gigalitre, or (in a pinch) hm3 to avoid "million cubic metres" LeadSongDog (talk) 14:48, 27 June 2008 (UTC)

In English (UK English) it is common to refer to 1,000 tonnes or 2,000 tonnes. Putting the issue of whether we call it a tonne or a metric ton to the side for one moment, a kilo-metric ton or even 2 kilo-metric tons seems daft. On the other hand 1 or 2 kilotonnes seems more reasonable; and Nuclear weapon yield's are expressed as TNT-equivalents in kilotons (or kilotonnes) and megatons (or megatonnes) - OK for "hairspliters" there a cumulative errors of about 2% between the two tons/tonnes. But, do we really need to dissambiguate a kiloton (or kilotonns) as one mega-kilogram; or one gigagramme?Pyrotec (talk) 16:31, 27 June 2008 (UTC)

Indeed, kilotons sounds like nuclear explosive yield. But then, so too does megaton and the environmental world uses megatons all the time when it comes to CO2 emissions. Different disciplines are used to different terminology. Little of what other editors above are saying above can’t be resolved simply by using language like this:

While *awkward* for some, it it confuses absolutely no one (which tonne would) and is fully SI-compliant and is fully in accordance with the teachings and advise of the BIPM. Greg L (talk) 19:13, 27 June 2008 (UTC)

The terms 'tonne' and 'metric ton' are not ambiguous. They may be unfamiliar. I do not believe there is a problem to solve here. The term 'corn' is ambiguous though...  :) Lightmouse (talk) 19:37, 27 June 2008 (UTC)
  • I think we’ve addressed the title issue: the use of “L/T S/T M/T”. I think we’ve also arrived at a consensus that “long ton” is not to be used (leaving only “tonne” and “metric ton”). I believe we have reached a consensus that the proper symbol for “tonne / metric ton” is the SI symbol t and certainly not “L/T” nor any variations of same (all of which are intended to stand for “long ton“). I think we have also determined that no unambiguous and suitable symbol for the short ton exists and the abbreviation S/T should not be used since both the S and the T are taken by the siemens and the tesla (and the lowercase s is reserved for the unit of time second). Therefore, “short ton” should be spelled out and no abbreviations or symbol used. I think we’ve also reached a consensus that the word “ton” should not be used as it is ambiguous since it can mean either 1000 kilograms or 2000 pounds. I think we may have arrived at a consensus that the choice as to whether “tonne” or “metric ton” should be used can be influenced by the subject of the article: in articles that are quite narrow in scope and very much tied to a particular country, such as uranium mining in the Northern Territory of Australia, editors should use “tonne” since it is most familiar to readers speaking the Australian dialect of English and it can reasonably be anticipated that the majority of readers will be Australian. What has not necessarily reached a consensus (but what I support) is that for general articles likely to be of interest to readers speaking the full variety of English dialects, the term “metric ton” should take preference over “tonne” because the latter is not understood by Americans and the former (metric ton)—while not a native first choice in some English dialects—is well recognized by all English dialects because of its intrinsically descriptive name. Greg L (talk) 05:52, 28 June 2008 (UTC)

MOSNUM should prefer tonne because it is unambiguous, is broadly accepted and has an internationally accepted symbol (t). If there is local consensus that "tonne" might cause confusion in a particular context, then let local editors make that decision. In a nutshell, MOSNUM should:

  • prefer tonne
  • permit metric ton, short ton and long ton where spelt out in full and linked
  • deprecate the other abominations

Thunderbird2 (talk) 12:41, 28 June 2008 (UTC)

  • All right. In my previous post, perhaps I should have said “general consensus.” I realize that there is no way all editors are going to be in agreement on this because different dialects have different practices. If it weren’t for a MOS rule, there would be edit warring over “color v.s. colour.” But at least an American understands what “colour” means (even though it looks odd to them, and visa versa). The issue at hand is much more substantial than that. “Tonne” may be unambiguous by definition perhaps, but it is certainly not unambiguous in practical effect because it is virtually unknown by 300 million Americans, many of whom assume it is a British spelling for an Imperial (pound-based) short ton. This is why even the BIPM (the organization that endorsed the name of “tonne”) says that In English speaking countries this unit is usually called "metric ton". The BIPM isn’t concerned about “being right” or “getting its way”. All the BIPM cares about is communicating without confusion; as does Wikipedia. Though it may be an unused term in other English dialects, everyone knows what “metric ton” means. That’s also why the symbol t should only be used for the “tonne / metric ton” (both have their uses), and why “ton” should be deprecated (because, indeed, it is quite ambiguous). Greg L (talk) 16:29, 28 June 2008 (UTC)
I say permit but tonne and metric ton & let editors decide which is most appropriate according to context. This is basically what has been happening on WP and thus this is where consensus really lies. One thing we should discourage is the over-kill metric tonne: there is no other tonne. Greg, stop for a moment, we're talking about three different units. Ever heard of a stone? It's a fourteen-pound unit that was once common in English-speaking countries outside of North America. When the Poms adopted the ton they warped it to fit stones in, just like they did to the mile. So a long ton is 160 stone, i.e. 2240 pounds, slightly more than a tonne. Dog, gigalitre, sure but I'd avoid the obsolete megastere, even cubic hectometre (an SI unit) would be preferable. I believe that a ton of TNT is the same as a tonne of TNT: it's not an exact measurement but a conventional definition. JIMp talk·cont 00:29, 30 June 2008 (UTC)
  • Good job Jimp; you figured out my confusion. No. I did not know we were talking about three different units. Yes. I spent a month on the British Isles (England, Scottland, Wales, N & S Ireland) and was quite surprised to be faced with stones when I stepped onto a scale. That was back in 1978 and I believe some of you still measure body weight in stones to this day, yes? And yes, I forgot all about that the long ton stood for 2240 lbs. In fact, now that I think about it, “tonne” is what I long figure stood for “long ton” (the British spelling for a 2240-pound ton). I’m a mechanical engineer who’s not too shabby with my units of measure (for an American engineer even) and I didn’t figure out that “tonne” was the official name for 1000 kg until I was in my 40s. Yes, I know that sounds sad. My point is that I can guarantee the editors here that the average American hasn’t a flying clue what “tonne” means. Given the reality of the situation, I fully agree with you that the reasonable thing to do is permit “tonne” and “metric ton” depending on what is most appropriate according to context. So…

    Where are we at? Fill in the blanks below. I’ll start:

  • Metric ton / tonne (1000 kg). Symbol: t “tonne” to be used in contexts strongly tied to countries with English dialects that use “tonne”.
  • Short ton (2000 lbs). Symbol: None? (can there really be confussion with S/T (siemens/tesla) within an article? Is there a standard symbol or abbreviation?)
  • Long ton (2240 lbs). Symbol: ?? (can L/T really be confused with anything from the SI within an article? Is there a standard symbol or abbreviation?)
  • Ton (to be deprecated as ambiguous).
Greg L (talk) 05:57, 30 June 2008 (UTC)
Thinking outside of the box...
LongT = 2,240 lb
ShortT = 2,000 lb
Therefore, ...her mother-in-law weighs 28.0 metric tons (27.6 LongT; 30.9 ShortT)
Without any "standard" symbols, that's what I came up with. Please improve it if you can... —MJCdetroit (yak) 16:29, 30 June 2008 (UTC)f
  • Does any of the above come from real-world practices? Or would it be a home-grown “invention”. And even if the former, it saves little room. I note that Encyclopedia Britannica is big on avoiding symbols and just spells out units of measure unless they repeat a great many times. If there are no symbols that are widely recognized in the real world, I would suggest we just write out “long ton” and “short ton”. Greg L (talk) 19:09, 30 June 2008 (UTC)
We the Brits do measure (human) body weight in stones and pounds; and also kilograms. However if it was a vehicle body we would weigh it in tons and hundred weight (cwt); or tonnes and kilogrammes. I see the US has 100 pounds to the cwt and the UK has 112 pounds to the cwt.

I've just added the following point to Confusing units and symbols

The terms long ton and short ton are written out in full. The the symbol t or the terms tonne or metric ton, as appropriate in context, are used for the metric unit of 1000 kg.

I believe that this reflects the general conclusion of this discussion without going beyond what we seem to have agreed upon. JIMp talk·cont 07:59, 3 July 2008 (UTC)

I believe that to be a fair summary. Lightmouse (talk) 09:41, 3 July 2008 (UTC)
  • Agreed. Very succinct and pithy. Well done. Greg L (talk) 13:31, 3 July 2008 (UTC)

Placement of Jimp's edit

Jimp placed the statement above in the former "Confusing units and symbols" section, but there was similar advise in the Disambiguation section. I feel this structure invites the creation of inconsistent advice in the two sections. Therefore, I have moved Jimp's contribution to the "Disambiguation" section, combining it with what was already there. I also renamed the "Confusing units and symbols" section to Units and symbols often written incorrectly to more precisely indicate the contents of the section. I am not particularly attached to the wording I used, and invite improvements. I would be opposed to having disambiguation advice in two different sections.--Gerry Ashton (talk) 13:58, 3 July 2008 (UTC)

Autoformatting purposes

Discussion moved from wp:context: start
I am currently engaged in a discussion with User:SilkTork who seems to feel that dates should not be linked for autoformatting purposes unless there is a further contextual need to have the date linked. I believe this is inconsistent with the current WP:MOSDATE, as well as with the following part of this style guideline (emphasis mine):

  • Dates when they contain a day, month, and year — [[25 March]] [[2004]] — or day and month — [[February 10]]should be linked for date preference formatting.

This would seem to be at odds with the "general rule of thumb" stated a little later on in the guideline:

  • Wikipedia has articles on days of the year, years, decades, centuries and millennia. As a general rule of thumb, link to one of these pages only if it is likely to deepen readers' understanding of a topic. Piped links to pages that are more focused on a topic are possible (1997), but cannot be used in full dates, where they break the date-linking function.

Just to clarify that this does not supercede the requirement that dates needing autoformatting should be linked, I made this edit, which was promptly reverted by SilkTork, who told me to take it to the talk page, since it "alters the meaning". I do not believe that this does alter the meaning of the guideline, and the fact that SilkTork is having difficulties understanding the rules concerning autoformatting clearly underscores the need to be a little firmer about this point. Is there any support here for my proposed edit? siℓℓy rabbit (talk) 15:32, 24 June 2008 (UTC)

It seems frustrating that linking and autoformatting are technically inseparable - it would be better if we could autoformat without linking, somehow. Regardless of standing policy, I personally find it objectionable to link an entirely irrelevant article just to improve the presentation of the current article. Dcoetzee 17:11, 24 June 2008 (UTC)

The ideal solution would of course be to have a technical solution in the MediaWiki software allowing for the formatting of dates without linking them. This feature request promises to be fulfilled sometime in the indefinite future. However, the present guideline seems to suggest that, in the meantime, dates must be linked if they are to be autoformatted. Somehow the meaning of the present guideline has been taken, by well-intentioned editors, to mean that autoformatted dates should be unlinked unless they are relevant to the context. This is not, and never has been, the case. siℓℓy rabbit (talk) 18:28, 24 June 2008 (UTC)

Discussion moved from wp:context: end

I have taken the liberty of bringing this discussion here. I hope nobody objects but the issue is relevant to both places and this is a more active forum. Lightmouse (talk) 16:59, 29 June 2008 (UTC)

* If User:SilkTork is making a tool that will allow a date to be input into a template that looks to user preferences to see whether they want the date displayed as “October 16, 1799” or “16 October 1799”, and doesn’t link to a list of trivia, let me know when it’s available. Greg L (talk) 18:46, 29 June 2008 (UTC)

  • I now see that all he is doing is manually removing link dates—not creating a template tool. My mistake. See Lightbot and Lightmouse - issue with years, above, where this whole issue is being discussed. As to Silly rabbit’s statement above: “The ideal solution would of course be to have a technical solution in the MediaWiki software allowing for the formatting of dates without linking them”, I agree, that would indeed be ideal IMO. However, if you read the link above, you will find that there are editors who feel as you do (in the absence of such a tool, they should be linked). I fall into the same camp as SilkTork: loose the links until the tool becomes available. Greg L (talk) 22:40, 29 June 2008 (UTC)
  • First thing: neither of those texts quoted above is actually in MOSNUM. Must be from very old versions. MOSNUM makes it quite clear ("can be autoformatted") that this regrettable practice is optional only. Second: the push to have the techs at WikiMedia decouple the linking and autoformatting mechanisms has a two-year history, and has been met with a less-than-enthusiastic attitude, even in the face of a huge petition of WPians. Don't hold your breath. What's the big deal about having two date formats, provided consistentn with an article? We do this for spelling, yes? TONY (talk) 00:56, 30 June 2008 (UTC)
  • Clarification: The above texts were quoted from WP:CONTEXT before this thread was moved here. Second of all, I still feel that my concern is not addressed (you reverted my change to the WP:CONTEXT guideline). One reading of the current guideline is that autoformatting should not be done unless the date being autoformatted is relevant to the context. If this is a correct reading of that guideline, then it should be made explicit, and we shouldn't edit-war over interpretations. siℓℓy rabbit (talk) 03:11, 30 June 2008 (UTC)
  • I thought it was linking that should be done only where useful and relevant. The fact that linking and autoformatting are scrambled together is very unfortunate, and we have to live with it. But the guidelines are different for each. TONY (talk) 03:35, 30 June 2008 (UTC)
  • My feeling is also that WP:CONTEXT really doesn't have anything to say on whether or not a date should be autoformatted. I think that this should be made clearer at WP:CONTEXT. siℓℓy rabbit (talk) 12:40, 30 June 2008 (UTC)

I think I understand the point that silly rabbit is making. Many users are confused. Autoformatting has been the emporer's new clothes that few dare question, so there is still the popular misconception that 'dates must be linked'. That confusion is partly because the software engineering contains a conceptual error. However, I think we could improve the guidance that exists on several pages. I submit that simple explanations involve simple tests that can be applied by naive editors on a remote talk page i.e. a set of bullets say that solitary years should not generally be linked, days of the week should not generally be linked, month+year combinations should not generally be linked etc. Lightmouse (talk) 19:03, 30 June 2008 (UTC)

  • I’m behind any proposal that addresses the excesses of linking I’ve seen. If the reader is researching a historical article on say, the French Revolution, it makes sense to provide the reader with a link to the “1790s”. But general-interest articles that happen to mention a date or year shouldn’t be cluttered with links to random trivia. Greg L (talk) 22:42, 30 June 2008 (UTC)
Well, maybe it makes sense, if the 1790s article is in reasonable shape and contains enough non-trivial information that is vaguely relevant to the field of the topic. But this idea of just scattergunning the chronology with links must be resisted in favour of disciplined selection for our readers' sakes (just as we select the best and most appropriate sources for them).
Here are two reasons for wanting all chronological items to be linked.
  • The first I've heard explicitly advanced: "I like to go off and browse these things—the Wiki tree is fascinating" (to be resisted as a red-herring: these people are quite capable of keying three- or four-character years or decades into the search box if that's their frisson).
  • The second is not talked about, and is the result of my assumption—that those who have invested their time and energy in constructing chronological pages get a kick out of coming up tops in the linked-to list, and feel that their work might be viewed by a lot more people because of the culture of linking these items (also to be resisted: these desires should not be at the expense of readability, appearance and focus on high-value links).
Now, I haven't heard anything from Francis Schonken or anyone else as to whether and why they're attached to the linking of these items. Reasons, please? TONY (talk) 00:43, 1 July 2008 (UTC)
Do we link years too often? Of course.
But if they are significant dates for the article, the year serves as well as a wikilink for contest as Poland or World War I, neither of which takes long to type. (Tony's argument proves too much; if it were valid we only need wikilinks at explicit cross-references, because we can always cut and paste other terms.)
The second assumption may well be true, but seems likely to be an extreme minority sntiment; rather, years are linked because editors have seen them linked.
It might help to state explicitly that years need not be linked; expecting editors to apply the full force of the distinction between can and should at one glance through MOS is probably unrealistic. Septentrionalis PMAnderson 14:06, 1 July 2008 (UTC)
  • I conducted a non-scientific straw poll of my real-life friends who, while not editors, are regular readers, and found a general feeling that the linking of dates had zero utility. Therefore, I would say that WP:CONTEXT and WP:MOSDATE should be explicitly changed to discourage the linking of dates. I believe that auto-formatting is not a good reason to link them at all, for the simple reason that the vast majority of Wikipedia readers don't bother creating an account, and therefore don't get any benefit from the auto-formatting.--Aervanath lives in the Orphanage 14:19, 1 July 2008 (UTC)
    • Well, if the groundswell of support continues, I can see this happening. I've just cleaned out a whole featured nomination that was a sea of blue. Awaiting the response of the nominator. Looks so much better. Compare previous blue clutter HERE with THIS. Scroll down each, preferably side by side. TONY (talk) 17:58, 1 July 2008 (UTC)
      • Nice to have no pointless blue links but really miss the dates in my preferred date format. A win/lose situation. See this bug report for interesting reading. One thing to think of is that without a working date preference system, lame edit wars will (again) start about whether it should be 1 July or July 1. Garion96 (talk) 21:58, 1 July 2008 (UTC)
      • Thanks for your comment, Garion. But you miss your home spelling too? "Colour/or", "ise/ize" and the countless slight variations we all accept? I think date formats are a much smaller issue (my own daily newspaper uses US date format against the prevailing norm in my country: never heard it raised as an issue). That bugzilla link you provided; um ... yeah, we're all very familiar with that. Fruitless, and fixing auto-dud in terms of decoupling it from the linking system won't at all solve the full panoply of dysfunctions. See the date after my sig → TONY (talk) 05:55, 2 July 2008 (UTC)
      • I am not a native English speaker, so color vs colour doesn't really bother me (although I prefer colour). The difference in dates is more worldwide, also for non-native English speakers. It wouldn't solve everything, but I would be happy with a way for auto date formatting without the wiki links. One never knows, I almost gave up hope on global accounts ever working working so.... Garion96 (talk) 11:35, 4 July 2008 (UTC)