Llewellyn's Astrological 2004 Calendar Llewellyn's Astrological Planner 2004 Calendar Holy Blood, Holy Grail The Templar Revelation: Secret Guardians of the True Identity of Christ How to Read Your Astrological Chart: Aspects of the Cosmic Puzzle Feng Shui Step by Step : Arranging Your Home for Health and Happiness--with Personalized Astrological Charts Planets in Synastry: Astrological Patterns of Relationships (The Planet Series) 2004 Astrological Appointment Book (Jan-Jun) Chiron and the Healing Journey: An Astrological and Psychological Perspective

Wikipedia talk:Special characters

From Wikipedia, the free encyclopedia.

Moved from Wikipedia:Village pump:

Recently I have seen some pages being created with special characters in their title (eg. Istiklâl Marsi, François Paul Jules Grévy, Émile Saisset, etc.). I'd like to know if this is accepted policy. If so, should the special character be typed in straight from the keyboard (eg. "é"), by its HTML entity name (eg. "é") or by some other means? After all, and strictly speaking, the guy's name is not Francois but François. D.D. 11:26 Jan 29, 2003 (UTC)

Some people only want to allow 7 bit ASCII for page titles while a growing majority want to allow full 8 bit extended ASCII which allows for accented and other Latin-1-based characters. Some older browsers (esp. on Mac OS 9x) destroy keyboard-generated é characters. But this is becoming less and less of a problem for at least Latin-1 characters (the situation for UTF is a HUGE mess with MANY browsers breaking characters). However many English-speaking users don't know how to create these extended Latin-1 characters at all so the only way they can link directly to François Paul Jules Grévy is to copy the characters. This is bad and very unwiki. It would therefore be very nice if the software treated Francois Paul Jules Grevy as a synonym of the accented name. That way users who don't know how to create the special characters can easily still link to article titles which have them. Of course, if and when there is widely accepted Anglicization you really should use that instead. --mav

We could freely add redirects with unaccented titles to article with accented titles, like I just did for François P.J. Grévy, and suggest people to do so. (BTW, is putting all a person's first names in the title really useful and recommended ? I'd answer "no" to the first question, it's a bit of a nuisance for me. I'm afraid the current informal policy answers "yes" to the second question.) --FvdP 20:32 Jan 29, 2003 (UTC)

But this only solves part of the problem (people may want to link to articles for which the redirect is not done yet). I would agree with automatic redirects, when no article exists under the unaccented title. --FvdP 20:38 Jan 29, 2003 (UTC)

Thanks for the answers. I think those redirects are a sensible solution to the problem. I'll stick with that for the time being. D.D. 09:07 Jan 30, 2003 (UTC)

Can we have the same auto-redirects for mis-capitalisation? If there's no article under the uncapitalised title, of course.

How will these "auto" redirects work? It is easy enough to strip the cedilla off of François, but in Norwegian, sometimes the Å (a separate letter in the Norsk alphabet) is rendered as Å and sometimes as Aa. The Aa is probably what Norwegians would expect to see in English, but maybe not. See Talk:Åfjord for a brief, slightly informed discussion. By the time you went through all these cases in all languages you'd have a pretty big table and likely some conflicts. Aren't handmade redirects like Aafjord safer? Ortolan88

Why not make it as simple as possible, and if there are more difficult cases, just ignore them. Surely that way we get a system better than our current one which is simple, easy ot maintain and unlikely to run into problems. Maybe when you click on an edit link for something like united states of america, in the explanation on how to edit the page it could give suggestions as what this page may be trying to link to? i.e. United States of America Smelialichu 16:19 Jan 31, 2003 (UTC)

Moved from Wikipedia:Village pump:

When I write an "a" with an acute accent is it best to use the code (ie. "á") or just the letter itself ("á")? Or is there no difference? -- sannse 13:40 Feb 26, 2003 (UTC)

The letter itself, see Wikipedia:Special characters. - Patrick 14:28 Feb 26, 2003 (UTC)

I beg to differ. Quoting from the above mentioned Wikipedia:Special characters page

Use an HTML named character entity reference like "á" . This is the most reliable method, and is unambiguous even when the server does not announce the use of any special character set, and even when the character does not display properly on some browsers.

which seems to imply, and it would be my experiance that the HTML named character entity reference like á is FAR better than inserting an #0160 directly. Rick Boatright 14:55 Feb 26, 2003 (UTC)

But Wikipedia:Special characters also says "For the purpose of searching, a word with a special character can best be written using the first method." (c&p etc.) "If the second method is used a word like Odiliënberg can only be found by searching for Odili, euml and/or nberg" So now I'm confused again. -- sannse 15:42 Feb 26, 2003 (UTC)

bear in mind that many pages including that one were written when the Wikipedia software couldn't handle accents at all. I'm sure one of the devs (eg Brion :-) will be along to tell us ... - Tarquin 17:20 Feb 26, 2003 (UTC)

That page is obsolete, pay it no mind. ;) Seriously though, let me lay it out:

• á or the uglier, but just as correct   will work in darn near any browser, for both viewing and editing.
  • but under our current system, it won't turn up correctly in searches
• inserting à should work correctly in darn near any browser, for both viewing and editing.
  • A small number of oddly-configured browsers might munge it on editing. However this is fairly rare for the latin-1 wikis. (There is more trouble with UTF-8 and older browsers, see m:Meta.wikipedia.org technical issues, but the English Wikipedia wiki is currently latin-1)
  • Search works!

Note also that titles are another matter. We do allow, and some of us encourage, accented chars in titles. There is a very small portion of the browser world that chokes on these (namely, Konqueror prior to 3.1, I'm told it's now fixed). Named characters references in the text of a link should be automatically be converted to the appropriate character for linking purposes (but there are some funnies when using numeric references; currently these go to UTF-8 as their primary use in linking is the interlanguage links... someday this'll all be smoothed out. :) --Brion 17:39 Feb 26, 2003 (UTC)

Okay, so let me see... Both work in almost any browser, but both have problems in some cases. So, erm, which should we prefer? I think I missed the actual answer there, but I'm probably just being dim... :) -- Oliver P. 17:48 Feb 26, 2003 (UTC)

Àààààààààààààààààààààààààààààà! --Brion 17:53 Feb 26, 2003 (UTC)

Heh! thanks, I was just trying to figure out the best way of asking again myself :) I think I've got it - I should use the characters themselves in the text to facilitate searching. (I also find it a lot easier to see what I've written that way).

As for titles: I've seen a few pages where the unaccented characters are used in a redirect page. So if someone searches for the unaccented word they don't miss a page that uses accents in the title or text. This seems useful, or is it not? sannse 17:58 Feb 26, 2003 (UTC)

That's mainly to catch links by people who can't/won't/don't think to use the accented character. But yeah, it catches searches too. --Brion 18:53 Feb 26, 2003 (UTC)

IE 6.0 does not show the red diamond above.

1. Test with

   <font face="Sans-serif" color="red">&diams;</font>

   for your browser: ♦

works for IE 6.0, does not work for Menchi's browser.

Any solution? - Patrick 08:01 May 8, 2003 (UTC)

The black diamond (without all the extra code) shows up fine in IE 6.0 (on XP) for me. -- Minesweeper 08:26 May 8, 2003 (UTC)

Okay, this is very interesting. The black diamond shows up for me in the default skin, but with the Blue Cologne skin, the diamond does not show up in IE 6. Of course everything shows up fine in Moz 1.3. - Minesweeper 06:54 May 9, 2003 (UTC)

I use IE 5.5 on WinNT 4, and the ♦ symbol shows up fine for me. -- Erzengel 08:29 May 8, 2003 (UTC)

I stuck in an extra copy, with what I suspect were some of those MS-y "smart quotes" replaced with regular quotes. How's that? They both work for me, Galeon-1.2.10/Mozilla-1.3. -- John Owens 08:33 May 8, 2003 (UTC)

I removed my test (accidently with smart quotes in the text, but not in the test itself) to avoid confusion. The black diamond (without the extra code) I do not see, even though I also have IE 6.0 on XP. - Patrick 08:59 May 8, 2003 (UTC)

1. Limiting page titles to Latin-1 is a problem; It is very (and I mean very) annoying for a non-English speaker to see a word in their language (or the name of a city in their country) stripped of all the diacritical marks (example: Lodz); it makes a very unprofessional, anglocentric and ignorant impression. (I certainly hope nobody is thinking of replying to say Polish speakers shouldn't be reading the English version of Wikipedia.)
2. Inputting non-Latin-1 characters is more than cumbersome, if not impossible for some users, even when they have the software (e.g. Chinese IME). The outlined "trick" for CJK characters does not work for me; I just keep getting question marks. So, to input CJK characters, I would have to look up the Unicode value of every single one of them and input them as numerical entities -- how much more difficult than necessary is that? Additionally, it wastes your disk space, but I suppose that's not much of a problem just now.

In comparison to that, the minor trivial problems some unpopular outdated browsers have with UTF-8 seem negligible. So... why does this site not use UTF-8? I don't know what your plans for the future are, but certainly you're not planning to stick with Latin-1 forever? You'll probably have to switch to UTF-8 some time in the future, so you might as well do it now because you have less data to convert now. -- Timwi 02:03 18 Jun 2003 (UTC)

Latin-1-centricism also cause difficulty in translating English and French articles to CJK Wikipedias. As we'd have to manually change the Unicode# back to true characters again. Or we'd have to copy the text from the webpage (not the editing page), but that'd lose all Wikification. --Menchi 02:07 18 Jun 2003 (UTC)

Cyrillic characters

Does anybody know if there is a way to copy and paste Cyrillic characters? Whenever I try, I just get a row of question marks - is it possible at all? -- Cordyph 18:49, 29 Jul 2003 (UTC)

It works for me with Mozilla - I did it several times with russian cosmonauts before, see e.g. Vladimir Komarov. Maybe your browser does not convert the russian characters to the Unicode numbers like Б. -- andy 21:23, 29 Jul 2003 (UTC)

Wait a moment, I just activate my Mozilla browser - yes, that works fine. I just tried this in Opera 7, but next time I will activate my Mozilla for correct character conversion - thanks a lot for the hint. (Hephaestos just told me that he has entered all the unicode values directly - appears to be a lot of unnecessary work) -- Cordyph 21:44, 29 Jul 2003 (UTC)

Un-typeable characters

Does the table on Wikipedia:Special_characters help? —Frecklefoot 18:17, 8 Aug 2003 (UTC)

And what about Start/Programs/Accessories/System Tools/Character Map, if you have it? (Might need switching to Unicode.) (Or, just install a lot of keyboard layouts.) Ксип Cyp 18:21, 8 Aug 2003 (UTC)

Charmap is fine for one or two characters otherwise a bit tedious to use. For several European languages you'll find a 'World Keyboard' at Bable Fish. Erik Zachte 23:14, 8 Aug 2003 (UTC)

Plan to move this page to meta soon

We need one place to develop MediaWiki documentation so I'm starting a project-neutral MediaWiki User's Guide in meta and think it would be a good idea to cross wiki redirect this page to meta:MediaWiki User's Guide: Creating special characters as soon as I'm done with the conversion (before that I will change each link to this page to a direct one to meta so that it appears correctly as an external link). --mav 04:11, 7 Sep 2003 (UTC)

Unicode

I hate them too. It makes it harder to edit, especially when there is a long string in a row. Don't most modern browsers support just pasting in the characters. What is the actual problem with doing that? If there is anything wrong with that, maybe the characters could be parsed so that when editing they look like normal characters, but when the page appears for display they are in Unicode. Dori | Talk 16:52, Dec 3, 2003 (UTC)

I agree except for — (—) which I think is better than using two dashes (--). Here's the transform (perl code) I've been using to fix special characters. This (well, the guideline, not the code) should go into the manual of style if it's not there already.

    # smart quotes
    s/[\x93\x94]+/\"/gs;
    s/[\x92\xb2\xb9]+/\'/gs;
    s/[\xb3]+/\`/gs;
    s/[\x96]+/-/gs;
    # HTML escapes
    s/\–/-/gs;    # –
    s/\—/—/gs;
    s/\‘/\`/gs;   # ‘
    s/\’/\'/gs;   # ’
    s/\“/\"/gs;   # “
    s/\”/\"/gs;   # ”
    # unwanted HTML escapes
    s/\–/-/gs;
    s/\‘/\`/gs;
    s/\’/\'/gs;
    s/\“/\"/gs;
    s/\”/\"/gs;

Daniel Quinlan 17:40, Dec 3, 2003 (UTC)

— and – do not display on some older browsers. So I use Unicode for them. This may be why you are seeing others as well. Fernkes 21:30, Dec 3, 2003 (UTC)

— should be used in articles. If it's an issue for display in HTML, then the Wikipedia software could do a transform to the Unicode character when generating HTML from the Wiki source, but I doubt that Unicode characters actually work on a higher percentage of browsers and systems. For –, it is far easier (and well-accepted) to use a simple "-" character. Daniel Quinlan 22:29, Dec 3, 2003 (UTC)

Dashes are annoying, everyone seems to have their own idea. People use -, --, — and —. The latter apparently works in more browsers than —, but nonetheless people go through articles chaotically changing each of the four styles to any other of the four styles, the target style being determined by the phase of the moon and various other astrological indicators. Let's just implement render-time automatic conversion from -- to some decreed standard and save everyone the hassle. -- Tim Starling 04:06, Dec 4, 2003 (UTC)

I thought about requesting that some time ago, but I had a bad feeling that "--" is sometimes used in other ways where it doesn't want to be —. Using — in article source may be safer. Anything to avoid numeric codes in article source, though! (If — is really more reliable, we should at least convert — to that when producing HTML.) Daniel Quinlan 04:17, Dec 4, 2003 (UTC)

What else is -- used for? I don't think it's used for anything where the meaning would be obscured by converting it to —. But if there is such a case, it can be escaped: --. -- Tim Starling 04:26, Dec 4, 2003 (UTC)

It is used in C sources, so any article that contains a C source would be displayed most incorrectly. Nikola 06:52, 5 Dec 2003 (UTC)

If I understand your regexps, you're adding the backtick, by the Wikipedia:Manual of Style is still not a Good Thing... Dysprosia 04:08, 4 Dec 2003 (UTC)

I only replace a Unicode/HTML backtick with a plain-old backtick. I add nothing. Much more manual editing is required to fix backticks since you need to figure out what the original editor intended or what is most appropriate in that context, so this script does additionally warn that a backtick was found. Daniel Quinlan 04:17, Dec 4, 2003 (UTC)

I'm no good at explaining things :) I'll let the MoS do it for me - not curved (smart) ones or the "backtick": Dysprosia 04:22, 4 Dec 2003 (UTC)

You're replacing left quotes with backticks, aren't you? Not backticks with backticks. -- Tim Starling 04:26, Dec 4, 2003 (UTC)

Ah, gotcha. Thanks for the correction. Daniel Quinlan 04:42, Dec 4, 2003 (UTC)~

For reference I mean this character ` <-- backtick. Dysprosia 04:33, 4 Dec 2003 (UTC)

Yeah, for some reason I thought &lsquo; was the same, but it's not. I'll default to converting &lsquo to ' now. Daniel Quinlan 04:42, Dec 4, 2003 (UTC)

Llewellyn's Astrological 2004 Calendar		Llewellyn's Astrological Planner 2004 Calendar		Holy Blood, Holy Grail		The Templar Revelation: Secret Guardians of the True Identity of Christ
How to Read Your Astrological Chart: Aspects of the Cosmic Puzzle		Feng Shui Step by Step : Arranging Your Home for Health and Happiness--with Personalized Astrological Charts		Planets in Synastry: Astrological Patterns of Relationships (The Planet Series)		2004 Astrological Appointment Book (Jan-Jun)
Chiron and the Healing Journey: An Astrological and Psychological Perspective