Wikipedia talk:Manual of Style (dates and numbers)/archive dash discussion
This is an archive of past discussions about Wikipedia:Manual of Style (dates and numbers). Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page. |
Dashes (again)
Hi, Chameleon.
See Wikipedia:Manual of Style#Dashes.
chocolateboy 20:39, 22 Jul 2004 (UTC)
- Here is a quote from Wikipedia:Manual of Style (dashes)#Dash guidelines for Wikipedia editors
- In the interests of Wikipedia:Wikilove and pending the planned update of the Wikimedia software that will automatically convert strings of hyphens into the appropriate correct en- and em dashes, editors are encouraged to be accepting of others' dash preferences and not to modify a chosen style arbitrarily, in the same way as they would refrain from arbitrarily changing "artefact" to "artifact" (or vice-versa). The following five dash styles are currently in use on Wikipedia: of these, three formats are endorsed and two are deprecated. Please do not change them to reflect your preference, except as indicated below.
- As far as I can tell, the usage that you changed is included in the list, so I do not think you are justified in making the change.older≠wiser 20:47, 22 Jul 2004 (UTC)
I think that we should recommend and illustrate the best practice while accepting (and quietly correcting) the common practice. Gdr 20:53, 2004 Jul 22 (UTC)
I hadn't spotted the movement away from hyphen and towards — and – on the talk page. FWIW, I dislike the HTML entities (particularly without surrounding whitespace), and hyphens are certainly the de facto standard for date ranges. Nevertheless, there clearly is some momentum behind the use of the traditional forms, though I think Gdr's compromise is more consistent with Wikipedia:Manual of Style#Dashes and current (if not necessarily "best") practice.
chocolateboy 21:19, 22 Jul 2004 (UTC)
- Dashes have already been discussed ad nauseam. You need to revert the article back to proper punctuation. Not liking HTML is not an excuse. — Chameleon My page/My talk 00:35, 23 Jul 2004 (UTC)
I don't need to do anything. The policy is still under construction and hyphens are still the default.
chocolateboy 00:49, 23 Jul 2004 (UTC)
- Articles with bad grammar and factual inaccuracies are also the default, but not necessarily therefore desirable. — Chameleon My page/My talk 02:10, 23 Jul 2004 (UTC)
Fascinating. But what has that to do with this discussion?
- ... the necessity of using HTML entities makes the wiki markup more difficult to read, and these do not display consistently in all browsers. (Please see the talk page.) Wikipedia:Manual of Style (dashes)
chocolateboy 02:52, 23 Jul 2004 (UTC)
- Don't be childishly sarcastic. You see what it has to do with it.
I've cited three authorities: Wikipedia:Manual of Style#Dashes, Wikipedia:Manual of Style (dashes), and the status quo. You've pontificated ("you need to...", "don't be...", "you see what...").
- Dashes display correctly in all modern browsers. Hyphens display incorrectly in all browsers.
- ---
- ... the necessity of using HTML entities makes the wiki markup more difficult to read, and these do not display consistently in all browsers. (Please see the talk page.) Wikipedia:Manual of Style (dashes)
chocolateboy 11:47, 23 Jul 2004 (UTC)
Chocolateboy, you have cited three authorities which do not unambiguously support the changes that you made earlier ([A]lthough you seem to be under the impression that hyphens are some sort of de facto standard or the status quo, this is most definitely not the case—there are countless articles with many varieities[sic] of dashes[.]). The opinion you quote[,] "the necessity of using HTML entities makes the wiki markup more difficult to read" is just one side of the guideline and discussion—there is no consensus one way or the other and both html entities and hyphens are supported. older≠wiser 13:55, 23 Jul 2004 (UTC)
- They are the de facto standard for date ranges (the subject of this article). Otherwise, I agree with everything you say, in particular "there is no consensus". I modified the article a) to reflect this fact and b) to ensure consistency with the other guidelines cited above.
- chocolateboy 07:23, 24 Jul 2004 (UTC)
- The content at Wikipedia:Manual of Style (dashes) represents the current consensus on dashes after long debate. It is still open to tweaking, but the core of it is clear. To summarise[,] there are various hacks involving hyphens ( - ), (--) etc., which are products of the fact there is no dash key on most keyboards. Contributors are welcome to write articles with such hacks, but should expect them to be changed to proper punctuation by Wikipedians carrying out clean-up. Three types of acceptable punctuation were agreed upon, old-fashioned unspaced em-dashes, new-fangled spaced en-dashes, and (my personal favourite) spaced em-dashes. It was agreed to change hyphen hacks to proper dashes, but not to change from one type of dash to another, in the interest of Wikilove.
- Dashes display correctly in all modern browsers (NS4 may have problems, for example, but then the entire site is unreadable in NS4, so it's not really an issue), whereas hyphen hacks do not display correctly in any browser. Numerical HTML entities are tricky to understand, but word-based ones are highly intuitive. Once the concept of starting with an ampersand and ending with a semi-colon is mastered, they couldn't be simpler, e.g.: ρ for the letter [r]ho (ρ), € for the euro sign (€), — for the em-dash (—). Since nobody is being told to use such entities, but merely to leave them alone when coming across them, one has to question the oft-used complaint that they are difficult. They are no more difficult, and disrupt the flow of the source code no more, than other methods we use for marking up text or inserting content, such as the image code ([[Image:Name of the image.png|left|thumb|110px|Comment]]) or the numerous colons, asterisks, square brackets, curly braces and apostophes[sic] that litter the source code for various purposes and are a necessary evil.
- It will soon no doubt be possible to enter such correct punctuation without the need for entities. All it would take is the adoption of UTF-8, and we could replace all entities by directly-entered characters—the important thing is that proper punctuation should appear in the article. On the same topic, it has long been planned to implement automatic conversion of hyphen hacks into proper dashes. Indeed, it was once implemented, but had to removed due to problems with table syntax (which uses "--"). Until that time, the tiny inconvenience of seeing the occasional entity in the source will remain.
- As for the concept of a "status quo", the current situation is that there are in excess of half a dozen different ways of attempting to represent various dashes. It has been decided that three of these are acceptable. It is our duty to change the bad ones to the good ones. Changing good ones to bad ones is rather unhelpful.
- A "default" is something that happens "because of a lack of opposition or positive action" according the OED. The proliferation of hyphen hacks is indeed because until now people haven't been strict and positive enough about proper punctuation; this is not an argument for such hacks. — Chameleon Main/Talk/Images 10:49, 24 Jul 2004 (UTC)
Hi.
While hyphens may not display "correctly" as dashes, they clearly do display "correctly" as far as the vast majority of editors and readers are concerned. Changing those hyphens to HTML entities serves only to alienate and annoy those contributors who take their lead from web standards of punctuation adopted by such sites as the BBC and H2G2.
Until consensus is genuinely reached (the other cited policy documents show that it has not been, and opinion is more or less evenly divided on Wikipedia talk:Manual of Style (dashes)), and, more importantly, the MediaWiki software is upgraded so that dashers and hyphenators can coexist peacefully, there is no reason to suppress recognition of the customary and official policy for dates on this site.
Wikipedia policy emerges bottom-up, taking its cue from what is done; it is not imposed top-down by the appeal of a handful of editors to standards of "propriety" that obtain in different media. That argument is explicitly addressed and rejected in Wikipedia is not paper and Wikipedia:Policies and guidelines.
The pending MediaWiki update (assuming it renders " - " as " – " and "--" as "—") actually argues strongly against the current deprecation (albeit incoherent) of hyphens, as they will, in future, become the de jure standard, while the currently obsolescent HTML entities will finally be rendered obsolete.
chocolateboy 15:24, 24 Jul 2004 (UTC)
- Use of "-" in date ranges, for example, will almost certainly not become standard in any future upgrade. It may be that "--" will become standard for n-dash and "---" for m-dash. If so, then at that time one might expect to see – no longer used in new edits. But it would mostly left in old edits because it works, just as things like é appear in many articles. They are not "obsolete", especially if a particular editor is used to those entities and doesn't know any other easy way to get such a character directly from the keyboard.
- You will find much use of &ndash on Main Page and Wikipedia:Community Portal. Use of "–" in date ranges and other ranges and use of either "–" or "—" or "--" for other dashes is the current standard in Wikipedia, arrived at after much discussion, when those like yourself who honestly don't like — and – were outvoted. Actually I don't think anyone likes the entities much, but a majority preferred the results that use of those entities produces over ASCII work-arounds. Similarly I don't think anyone likes three apostrophes to indicate bold emphasis. They use it because they like the result.
- But the current standards on dashes [and on other matters] were not imposed top-down. They were decided by discussion. In the case of dashes it became an issue when some users, notably one called Wik, started unilaterally changing other people's use of "—" and "–' to " - ". There were bottom-up protests against this. There was no top-down suppression. More people who cared about the issue one way or the other opted for use of HTML entities as the norm. This was recently modified to allow "--" also, after notice on the pump and discussion on a talk page. Please abide by consensus in this matter. Otherwise it is Chocolateboy who is alienating and annoying others, who is suppressing others who are following what is here the preferred style and often the individual preferred style of the users whose work he is changing.
- Not all points of Wikipedia style are my own preferred style. But I try to follow Wikipedia style when editing here.
- As to standards being under construction ... all standards here are always under construction and subject to modification at any time, if there is consensus among those who care about any standard usage to modify a particular standard or to extend it. That's how the current policy on dashes emerged. Unilaterally changing someone else's standard usage to non-standard usage without discussion is normally not acceptable on any site or in any publishing entity, regardless of what standards that site or publishing entity has chosen.
- Even three years ago it was very reasonable in English web publishing to keep within the cage of the subset of characters defined by the intersection of Latin-1, MacRoman, and Latin-9. It even made sense to stick to ASCII as character set conversions were sometimes buggy. So of course web sites continued to use the normal ASCII/typewriter kludges to represent dashes despite the spread of Unicode because of difficulties with older systems and older browsers. Many large sites still follow those standards even after the reason for them has gone. Changes take time and the hand of convention rules. But the cage door is open now. HTML Unicode entities and UTF-8 are supported on free browsers available on almost every system and there is longer need to restrict oneself to old-fashioned limitations inherited from typewriter technology. If you won't get out of the prison yourself, don't expect others to stay there with you or agree with you holding them back.
- Jallan 00:54, 30 Jul 2004 (UTC)
- I like those pseudapostrophes for style. lysdexia 21:55, 13 Nov 2004 (UTC)
Hi, Jallan. Thanks for your comments. I agree with some of them, and, for the record, would be quite happy to see – and — in the rendered HTML.
- If so, then at that time one might expect to see – no longer used in new edits. But it would mostly left in old edits because it works, just as things like é appear in many articles. They are not "obsolete"
They are obsolete. They would be removed by conscientious Wikipedians just as the jejune and redundant use of HTML to achieve formating that is directly catered for in wikitext is removed (usually with a gloss of "wikified").
- They still appear at [[1]] which is the help page accessed from an edit page. Presumably they are there to help editors not clear on how to get the characters from their keyboards. A gloss of "Wikified" does surely not necessarily mean that such characters are removed or that, for example HTML tables are removed, but rather that titles mainly that[sic] titles are now in standard Wiki format, the introductary paragraph follows Wiki standards, links have been made to other articles and other stated standards of the style sheets have been followed—encluding[sic] use of the proper dash characters. Jallan 03:30, 10 Aug 2004 (UTC)
- Use of "–" in date ranges and other ranges and use of either "–" or "—" or "--" for other dashes is the current standard in Wikipedia
That is incorrect. See below.
- arrived at after much discussion
That ("arrived at") is also incorrect. There is no consensus. Or rather, the consensus (the solution that displeases no-one) is that hyphens should be used in wikitext and that they should be converted to – and — in the resulting HTML. See below.
- More people who cared about the issue one way or the other opted for use of HTML entities as the norm.
This is incorrect. The majority of Wikipedians are unaware of the attempts of a vocal minority to overthrow standard web usage. (It doesn't help that the discussion is spread over several talk pages and articles.) Those who do care enough to contribute to the discussion are almost unanimously in favour of the hyphens-in-wikitext/dashes-in-HTML solution referred to above.
- How do you know what the majority of Wikipedians are aware of? The discussion is spread over many talk pages and articles, but the guidelines are not. A further change to the standard (allowing "--") was advertised on the pump for about a month. No-one has tried to keep anything secret. As to "standard web usage", that is changing also. Jallan 03:30, 10 Aug 2004 (UTC)
- Here are some sites that I found quite quickly. The link is to a sample subpage when the main page does have a dash example:
- Oxford University Press: Online
- University of Chicago Press: In the News
- Enyclopedia Britannica: Henry V
- Bartleby: Columbia Encyclopedia: Shakespeare
- InfoPlease: Columbia Encyclopedia: Shakespeare
- Grolier Encyclopedia American: George Washington
- Jewish Encyclopedia: David
- University of California Press: Press Release: August Frugé Dies age 94
- Johns Hopkins University
- University of Texas at San Atonio[sic: Editorial Style Guide: Visual standards for representing UTSA in print, Web and new media]
- Plumtree Brfanding[sic and Style Guide]
- New York Times Must register (free) to see articles.
- Washington Post Special dashes can be seen heading the bylines.
- Some of this is probably quite old. One site was even using a graphic kludge to emulate a dash. But a lot is new. Do you really think any of these pages and sites will revert back to hyphens? Do you really think that many other sites are not going to change in the next couple of years.
- You might try telling the following that they don't know anything about using the web:
- Microsoft: About this site
- Apple: Hot New[sic]
- Adobe Systems: About Adobe
- Digital Web Magazine: Firefox 0.9 They demand that submitted articles have the proper dashes. See Guidelines for Contributors
- It is quite possible that some of these sites, like Wikipedia, also have hyphens for dashes in the majority of pages. But not in the new pages. Though I did find sites I did not list which were mixing "--" and "—" even on the same pages. New technology means new standards are possible. And people are beginning to take advantage of the possiblities, moving away from a standard forced by necessity rather than by consideration of was really wanted. Jallan 03:30, 10 Aug 2004 (UTC)
- Here are some sites that I found quite quickly. The link is to a sample subpage when the main page does have a dash example:
- It's an old standard and technology in MacOS that opt-- and shift-opt-- make en and em dashes. If people weren't so ignorant of superior products, we wouldn't be having this debate. lysdexia 21:55, 13 Nov 2004 (UTC)
- Please abide by consensus in this matter.
I would advise you to do the same. I have gone to some trouble to support my argument with statistics rather than dismissive references to individual Wikipedians ("Chocolateboy", "Wik"), which, IMHO, are offtopic.
- That thousands of old articles (and many new ones) do not follow Wikipedia standards in many different ways is a fact. So standardize them. Jallan 03:30, 10 Aug 2004 (UTC)
- Not all points of Wikipedia style are my own preferred style. But I try to follow Wikipedia style when editing here.
I prefer title case headings to sentence case, as do many other websites (and, for that matter, print style guides). I don't go through Wikipedia systematically changing headings to title case. Nor do I complain or revert when an editor (or bot) tweaks my headings to conform to the de facto Wikipedia standard. Given the large number of witting and unwitting Wikipedians who favour title case headings (see Lady Lysine Ikinsile's list for example), I doubt it would be difficult to corral a troupe of contributors to vote for a title-case overthrow of the relevant policy page. I choose not to do that because I prefer not to fly in the face of the de facto standard both here and on other websites. If there is an argument for a minority-backed imposition of a dash policy that flouts customary Wikipedia usage, then there is an equally strong case for the overthrow of all de facto Wikipedia standards.
- I thought about the same sentiment, but title caps is too much for an encyclopædia article, much less the sections. I'd use title caps if the contribution is proportionally-large to the overall work, the medium, which is also title capped. lysdexia 21:55, 13 Nov 2004 (UTC)
- Ah! customary usage. ASCII only was at one time customary on the web if you wanted to be safe. "Customary usage" was exactly the excuse Wik claimed when he began arbitrarly changing other people's ussage[sic]. But of course it was customary. Use of special characters outside of[sic] Latin-1, and even outside of[sic] ASCII, used to be somewhat chancy, and still is for many of the newer Unicode characters. Not for dashes though. No reason any more not to use them. The world has moved on. Try changing the main page in Wikipedia to get rid of the dashes and see what happens. Jallan 03:30, 10 Aug 2004 (UTC)
- Unilaterally changing someone else's standard usage to non-standard usage without discussion is normally not acceptable on any site or in any publishing entity, regardless of what standards that site or publishing entity has chosen.
There is no unilaterism here. There are ample arguments on both sides, as your previous paragraph backhandedly acknowledges. The position with the broadest support advocates the use of hyphens in wikitext, and dashes in the rendered HTML.
- Of course. And until then ... it advocates use of entities, or "--". Read the discussion carefully. There is no way that a hyphen in a range will be automatically changed by software from a single hyphen to an en-dash. That's why use of – in ranges is particularly urged. How does the software know whether "150-355" is a range or an identification code of some kind? Of course linked date ranges could probably be identified, leaving the entities to other cases. But current suggestion seems to be that en-dash in a range should be represented by "--" and em-dash by "---" following TeX standard. Who knows? It's all pie in the sky. Meanwhile, the entities work, and they will continue to work regardless of what might be done, just as do HTML tables. Or perhaps a convention to UTF-8 will happen first, in which case people can simply enter the proper dashes if they want without messing around with either HTML entities or silliness like "--" and "---". Jallan 03:30, 10 Aug 2004 (UTC)
- using prepositions and date labels in pattern matching lysdexia 21:55, 13 Nov 2004 (UTC)
- Of course. And until then ... it advocates use of entities, or "--". Read the discussion carefully. There is no way that a hyphen in a range will be automatically changed by software from a single hyphen to an en-dash. That's why use of – in ranges is particularly urged. How does the software know whether "150-355" is a range or an identification code of some kind? Of course linked date ranges could probably be identified, leaving the entities to other cases. But current suggestion seems to be that en-dash in a range should be represented by "--" and em-dash by "---" following TeX standard. Who knows? It's all pie in the sky. Meanwhile, the entities work, and they will continue to work regardless of what might be done, just as do HTML tables. Or perhaps a convention to UTF-8 will happen first, in which case people can simply enter the proper dashes if they want without messing around with either HTML entities or silliness like "--" and "---". Jallan 03:30, 10 Aug 2004 (UTC)
Contrary to the repeated appeal to a "vote", there is no consensus on the Dashes talk page. There are around fifty different contributors to that page. A handful of recurrent print-oriented pundits (most of whom have tried to railroad this and that page into supporting their position) have repeatedly lobbied for entities in the wikitext. A similar number have objected to this on the grounds that it obfuscates the source. The rest have either made peripheral observations, or have simply endorsed the status quo, which is that hyphens should be used in the source and rendered (in the resulting HTML) as – and —. Most contributors to Wikipedia are manifestly unaware of any of the dash policy pages; they contribute to this discussion by example. In contrast, the – and — advocates project a seriously skewed impression of Wikipedian policy by systematically organizing their putsch here and on other policy pages. The hyphen advocates are the silent majority. The entity advocates are the vocal minority.
- The "silent minority" fallacy. All the people out there who agree with you but somehow don't speak up. Not credible. True, a lot don't know (and a lot don't care.) And there is a group who don't like the idea of having to mess around with dashes, just as there is a group who don't like mixed spellings from one article to another, and a group who loaths[sic] tables, and a group who would ban almost all lists, and a group who would ban most articles about fictional characters, and a group who sees no reason why "how[-]to" articles shouldn't be included and so forth. Jallan 03:30, 10 Aug 2004 (UTC)
Hyphens are the de facto standard on Wikipedia, not only for dates, but for "dashes".
- Yes, they are standard for old articles, except when changed, in a sense. There is a lot of changing to do, and seemingly no-one feels strongly enough to start doing this as a project. So the changes happen gradually, along with other edits. HTML tables are probably the de facto standard on Wikipedia also. Jallan 03:30, 10 Aug 2004 (UTC)
(The file normalized.txt referenced below is a version of the 20040727 cur table dump with talk pages removed (script available on request). Due to stack overflow issues in Perl's recursive regular expression engine, a few longer articles are also excluded from these statistics.)
Globally, hyphens are at least 15 times more common than ndashes or mdashes combined:
- I am not surprised. Maybe about six months ago they were 40 or 50 times more common. Who knows? Jallan 03:30, 10 Aug 2004 (UTC)
- grep '–' normalized.txt | perl -pe '$_ = join ($/, /–/g) . $/' | wc -l
- > 14663
- grep '—' normalized.txt | perl -pe '$_ = join ($/, /—/g) . $/' | wc -l
- > 16526
- grep ' - ' normalized.txt | perl -pe '$_ = join ($/, / - /g) . $/' | wc -l
- > 494155
Likewise, hyphens are approximately 40' times more popular than dashes for date ranges:
- grep '\]\] – \[\[' normalized.txt | perl -pe '$_ = join ($/, /\]\] – \[\[/g) . $/' | wc -l
- > 2698
- grep '\]\]–\[\[' normalized.txt | perl -pe '$_ = join ($/, /\]\]–\[\[/g) . $/' | wc -l
- > 2599
- grep '\]\]-\[\[' normalized.txt | perl -pe '$_ = join ($/, /\]\]-\[\[/g) . $/' | wc -l
- > 59366
- grep '\]\] - \[\[' normalized.txt | perl -pe '$_ = join ($/, /\]\] - \[\[/g) . $/' | wc -l
- > 160911
As you can also see from those stats (which exclude some date ranges and include some non-date-ranges: patches welcome!), spaced hyphens are used approximately 3 times more often than unspaced hyphens.
- So ... there's a lot of work to do to get Wikipedia (and the web) out of the ASCII kludges once necessary for technical reasons, kludges based on old typewriter technology (which is a paper print technology as much as professional typography, but a cheap kludge technology). But that HTML tables are more common in Wikipedia than pipe tables doesn't mean that one shouldn't use pipe tables does it? Of course it really doesn't matter with tables. HTML or pipe, they look the same when viewed. Not so with dashes. But you are playing with the word standard, using it to mean what happens to exist, as though to say that Internet Explorer HTML is "standard" HTML. In that sense, it is standard HTML. It is very standard. In another sense it is not. So get the dashes off the main Wikipedia page and the Community portal page if they aren't what you think should be standard and you can get people to agree with you. Otherwise, leave other people's dashes alone. Jallan 03:30, 10 Aug 2004 (UTC)
chocolateboy 16:05, 8 Aug 2004 (UTC)
It's a shame that you felt the need to grievously disfigure my comments in the course of your reply. I hope you'll learn to quote (rather than interlineate) in future. If my laboriously researched facts are going to be ignored (as they have been, systematically), then I see little point in taking the trouble to refute your inky op-ed (I congratulate you on plucking a handful of semi-dashed sites out of the seething majority of hyphenated sites: most of them are mechanically-converted print resources; and the Washington Post uses hyphens in the articles). As stated before, if there's a case for the style-guide approved dashes in wikitext (no-one objects to them appearing in the rendered HTML), then there's an equally strong case for a militant overthrow of the sentence-case headings policy, which "merely" has consensus (rather than "correctness") on its side. If the censorship of the de facto dash policy on Wikipedia prevails, then I look forward to the contributions of your title-case vandalbot in the forthcoming headings consistency (with this policy, not with other Wikipedians) war.
chocolateboy 00:19, 18 Aug 2004 (UTC)
- Em and en dashes are considered the proper punctuation in many cases. Using em and en dashes can be considered similar–albeit outdated in HTML–to putting one space between sentences: two spaces are a relic of the old days of fixed-width typewriters, needed to make the page readable. I notice that you use one space. Now, with the advent of computers, we can use dashes—even true minus signs (not dashes or hyphens)—without having them be the same width as the rest of the text. Oh, and I'm sure that websites with bad HTML are about 200 times more abundant than ones with good HTML, but that doesn't mean that I'm going to code badly. And sentences that end with prepositions are about twenty times more common than their "proper" equivalents...I could go on and on. *L* — TTD Bark 05:01, 2004 Sep 11 (UTC)
- Ending with prepositions is only forbidden in Latin. lysdexia 21:55, 13 Nov 2004 (UTC)
I hope this is the right place to bump in with my opinion, too. :) Personally, I'd support any system-wide conversion away from hyphens to dashes. Dashes are certainly a part of proper typography, which is [nearly] as important as proper grammar. I gather that one of the aims of Wikipedia is to be as professional as possible, yes? I'm sure we have plenty of badly-constructed sentences, but that doesn't make them the right thing to keep, either. While we can't plausibly include every correct typographic convention (oh, beautiful curved quotes, how I miss you) IMHO we should include them when possible. Miss Puffskein 23:48, Aug 10, 2004 (UTC)
- As to curved quotation marks, I quite agree, though no-one including myself is avidly pushing for them. For one thing, there are reasonbly[sic] good algorithms to produce "smart quotes" which can allow this to be done on the fly when users download without anyone having to bother about them while editing. I expect that will be done here eventually. "Smart quote" algorithms aren't perfect, but are good enough for about 95% of usage. Things like "Spirit of '76" come out with the wrong mark as well as things like 20º 7' 3" where the straight quotation marks should not be turned into curly quotation marks. Also one would not want any translation done in fixed[-]with[sic] text, e.g. text of computer source code. The answer is to explicitly hardcode the proper curly quote symbols and proper prime symbols only for those cases where the alogrithm does not convert properly. And provide some way, perhaps \' and \", to represent straight quotation marks that are to be left alone (as well as leaving them alone in fixed-width text. It would quite doable and not very complicated. It is unfortunate that dashes cannot be handled so simply through an algorithm that handles about 95% of the cases without difficulty or there would have been no debate. And the algorithm which was tried messed up pipe table format. Jallan 20:50, 12 Aug 2004 (UTC)
- The angle measurements you mention:[sic] 20º 7' 3":[sic] they[sic] should be converted, but into prime and double-prime. Once again, A List Apart has an excellent discussion about miscellaneous [confusing punctuation marks (including dashes!). — TTD Bark 05:01, 2004 Sep 11 (UTC)
Summary of debate so far
I note that everyone in this debate is agreed that dashes in date and number ranges should appear as en-dashes when displayed or printed. The disagreement is on what method or methods should be recommended in the Manual of Style to achieve this. The proposals are
- Recommend –. (Advantage: achieves the desired result immediately. Disadvantage: awkward to type, offputting to newcomers.)
- Recommend hyphen. (Advantage: common practice. Disadvantage: never achieves the desired result, because it there's no way to automatically convert it.)
- Recommend double-hyphen. (Advantage: easy to type. Disadvantage: doesn't achieve the desired result immediately. A wiki markup feature that turned double hyphen into en-dash would introduce a small number of errors, for example in the discussion of the -- operator in the C programming language.)
Note that one day the English Wikipedia will allow all Unicode characters, so then surely the recommended approach will then just be to type an en-dash. Gdr 13:04, 2004 Aug 13 (UTC)
- It's easy to automatically convert both the spaced hyphen (to –) and the spaced double hyphen (to — (not –)). There's always ­ (soft hyphen) for (rare) isolated pre- and post-decrement operators. As you can see from this discussion, it's not a complicated issue :-) Otherwise, I agree with your summary.
- chocolateboy 23:39, 17 Aug 2004 (UTC)