Template talk:SpecialChars

From Wikipedia, the free encyclopedia

Contents

[edit] Producing replacement character examples, part 1

Two comments:

  • Strictly speaking, it's not the MSIE in itself, but only in combination with missing fonts or improper font settings. Without any font to render a specific char, also Mozilla or Opera will give up.
  • The use of � "�" is misleading as it renders as "broken encoding glyph" in all browsers. The character that MSIE users see, when the character can not be displayed due to not being in the selected font, is the "missing character glyph", which is strictly speaking font dependant, but mostly an open rectangle. Perhaps using ௿ "௿" would be a better fit.

Pjacobi 21:04, 10 Oct 2004 (UTC)

I made sure I didn't say "IE is crap and all character viewing problems are Microsoft's fault; death to Bill!", but I knew someone would complain anyway. OK, now I've put "not always display correctly". That is true. These characters do not always display correctly in IE (in the real world, they never display correctly in IE for the vast majority of people), and they will always display correctly in Firefox unless you do something to screw it up.
I was just being realistic.
I don't think � is misleading. In IE I get a box — the same box I see for the characters IE can't display at Pinyin. Your ௿ gives me a different, narrow box. In Mozilla, however, I get a question mark in a diamond with � and a normal question mark with your ௿ or with the characters Mozilla can't handle at Greek alphabet. I'll add a real question mark. Chameleon 21:35, 10 Oct 2004 (UTC)

I'm normally on Moz 1.7 for surfing but I have MSIE5 installed for testing, and I can test with MSIE6 and Opera7 from another computer. My only concern, is to not confuse users even, which may be already be confused due to their broken display.

Now MSIE5 displays two question marks on the warning template, but open rectangle for polytonic Greek on Greek alphabet.

OTOH it seems U+0BFF was a bad choice in any case. Let's try to do it systematically, if there is a system in it

Code Char Moz1.7 MSIE5 Opera8
� � diamond-?  ? box
஀  ? box box
Ϡ Ϡ sampi box box

But perhaps it depends on too many factors, and a more vague description is needed.

217.227.2.156 21:49, 10 Oct 2004 (UTC) Oops, logged out. That was me again --Pjacobi 22:01, 10 Oct 2004 (UTC)

[edit] Internet Explorer

This template is not really needed: the issue is not with the so-called "special characters", but rather with MSIE's broken implementation of characters the browser cannot display. Under no circumstances should people be told NOT to use the correct characters just to work around bugs in outdated software. My ¢2. {Ανάριον} 10:07, 14 Oct 2004 (UTC)

I hate IE as much as you, but I tried to make my wording NPOV. I fear your changes will simply be reverted by someone who believes it is too anti-IE.
Nobody is being told not to use the correct characters just to work around bugs. Please don't interpret my template like that. Its entire purpose is to warn people (people consulting the article, rather than editors) about encoding problems they might come across. I think it does that well.
I'll tell you why I created the thing. The other day, someone asked me to write out the letters of the Greek alphabet for him. I told him it would be quicker if he just looked up the appropriate article on Wikipedia, and he did so. An hour later, I asked how it had gone, and he told me "ah, Wikipedia haven't done the article right. It's corrupted or something, with little boxes instead of letters." Then I remembered he had refused to install Firefox on his computer. It really annoyed me that hundreds of people must come to our articles, just like him, and decide that we have made a mistake and not put the proper characters in. I set him straight, telling him the problem was IE, but what about all the people I don't get to tell? I decided to create a template that would inform IE users that, if they saw little boxes, there was a reason for it. Then, if they don't care about seeing the symbols, they can just ignore them, at least knowing that it is not a screw-up on our side; if they do care about seeing the symbols, they are informed of a way around the problem.
I think this is essential. We desperately need this template until Internet Explorer users on Xp (the vast majority of people on the internet) can see our symbols "out of the box". Chameleon 11:07, 14 Oct 2004 (UTC)
I'm going to revert:
  • Calling IE "outdated" is just going to raise the ire of IE supporters. Let's keep cool. It is clear enough that IE is the main problem.
  • Adding "௿" is not helpful. In IE, it displays a narrow box, which is not how obscure characters display in IE (they display a wider box). In Gecko-based browsers, it displays a question mark, which is already there.
  • It is not true that installing a Unicode font helps. Windows comes with Arial Unicode MS. The problem is that this is not the default font, and that fact that IE can't deal with Unicode except when a Unicode font is used (unlike Mozilla, which makes the appropriate substitutions).

Chameleon 11:07, 14 Oct 2004 (UTC)

Agreed with many points, but I still have to reword it.
Don't get me wrong: MSIE was a good browser once, but the fact its standards support has hardly been updated in the past six years has made it the new Netscape 4: a single browser holding back the entire web. I've slightly reworded it again, focus should be on the browser being the problem, not the characters.
And Arial Unicode MS does not come with Windows: this only comes with Microsoft Office. Unfortunately as you state MSIE is incapable of using Unicode fonts correctly, unlike Opera, Mozilla, Firething, etc.. {Ανάριον} 11:38, 14 Oct 2004 (UTC)

AFAIK the main MSIE problem is, that it doesn't try other fonts if the "normal" (selected by user, stylesheet, etc) font doesn't implement the character. Gecko browsers will try, and as a result you will sometimes see, that some rare latin characters (with underdot or underbar) are displayed from a different font.

For the concrete wording of the message, we should try to reproduce the "open box", as it is a very common display in MSIE. If ௿ doesn work, perhaps ଀ or Ϡ will.

Pjacobi 14:21, 14 Oct 2004 (UTC)

Any unassigned character should work, for example U+0750: ݐ
I think the main MSIE problem is that it, alone, refuses to use multiple fonts for one "range". If in Opera I mix "normal Latin" (ASCII) with a dark l, and my normal Latin font does not include it, the dark l will be taken from a font that does. Mozillas also work this way. In MSIE I would have to change my entire normal Latin font to the font that does include it, which often means all webpages will look bad (most large Unicode typefaces I find rather ugly, or they are too large/too small/too vague). Added to this problem is the impossibility to properly assign fonts to Unicode ranges in MSIE. {Ανάριον} 14:51, 14 Oct 2004 (UTC)

[edit] Producing replacement character examples, part 2

No, no, no! Stop changing the fake character! The result is that in IE you see a narrow box that does not correspond to the symbol you get when viewing Greek alphabet, etc, and in Gecko you get a question mark, so it reads ‘Bear this in mind if you encounter symbols such as "?" or "?"’, i.e. the same thing twice. My way was better. I'm reverting. Chameleon 18:10, 14 Oct 2004 (UTC)

"?" or "?" doesn't look nice, but "??" (as in the original version of the page by you) does. My change addresses that. -- Paddu 22:26, 20 Nov 2004 (UTC)

[edit] Font advice

It was suggested that using "Arial Unicode MS" doesn't help so it's not a font-related problem. That by itself doesn't prove that installing other fonts wouldn't help. In fact, "Arial Unicode MS", — or for that matter most fonts too — doesn't include glyphs for all assigned unicode codepoints. A "unicode font" is not required to contain glyphs for all assigned unicode codepoints, only that whatever glyphs it contains are mapped to proper codepoints as per the Unicode standard.

There definitely is a problem with fonts, since not all systems have fonts to display all characters. This problem affects not only IE but also mozilla, etc. and can only be solved by installing proper fonts and configuring the browser to use them.

But having said that, there certainly is a difference between how IE and Gecko (at least on linux) select fonts to display glyphs from. On linux, an Xft build of mozilla uses fontconfig to search all the fonts in specific directories for any glyph (this includes fonts not specified in mozilla preferences). I'm not sure Gecko searches all fonts in Windows, but at least IE doesn't do that, e.g. http://hi.wikipedia.org/ showed only boxes until I specifically installed a Hindi unicode font.

BTW the size of the box shown probably has something to do with whether the codepoint is representable in 16-bits or requires 21-bits. Mozilla on linux shows a box with the codepoint value written inside in hex, which means different widths for different no. of hex digits.

I've changed the template to reflect the fact that any browser/GUI environment can have this problem (e.g. missing or misconfigured fonts) and changed the "fake character" to codepoints guaranteed to be unassigned in Unicode plus a question mark. HTH -- Paddu 22:23, 20 Nov 2004 (UTC)

[edit] Category confusion

Why does this template have a category of Category:Pages with special characters. This category doesn't exist, and doesn't lead to anywhere. Is there some intended use for it? Can it be removed? --Sortior 00:33, Nov 25, 2004 (UTC)

Ugh! That category does exist (i.e. there exist pages under that category or under subcategories) and lists pages with special characters or subcategories that list some such pages. Just that there is no introductory text about the category on the caregory page itself.
Whether it can be removed — What is listed in the category can be got here too: Special:Whatlinkshere/Template:SpecialChars but most of these kinds of templates seem to have a related category page, e.g. Template:Disambig, Template:Cleanup, Template:Aero-stub. Probably people feel category pages are more organised than Whatlinkshere. Also there might still be a limitation on the no. of links shown in Whatlinkshere (It used to be there, not sure if it's still there now). -- Paddu 07:49, 25 Nov 2004 (UTC)
OK, at least Special:Whatlinkshere/Template:SpecialChars doesn't give any idea about subcategories but Category:Pages with special characters does. -- Paddu 09:10, 25 Nov 2004 (UTC)

[edit] Aesthetics

I'm not going to complain about the content of the box, and I appreciate the idea behind it. but please make it smaller. It disfigures the articles. It should be both in smaller font and with less text. It doesn't have to explain the browserwars in a nutshell, you know. just tell people that if they have trouble reading the page, click here, and take them to a page that explains the issue. dab () 19:57, 18 Dec 2004 (UTC)

I made the text smaller. I think I can live with the warning box in this size... dab () 12:41, 22 Dec 2004 (UTC)

[edit] Template is not practical

I don't think this template is practical. Every East Asia-, Middle East-, South Asia-, and Southeast Asia-related article is bound to contain special characters in a few cases. Articles like People's Republic of China, Japan, India, Thailand, South Korea, etc. all contain special characters. Since coming to Wikipedia I've worked on hundreds of China-related articles, including geography, history, language, culture, and famous people. Many of these articles already have disambiguates, maps, infoboxes, photos etc. crammed at the top, and I'm almost certain that all of them are going to contain Chinese characters in at least one place (if not several). So am I supposed to go back now and put this box into every single one of them? I think this destroys the look of a very large proportion of articles on Wikipedia. -- ran (talk) 04:54, Jan 25, 2005 (UTC)

I honestly believe that it was meant for pages with heavy use of mathematical symbols. If these symbols do not resolve then the page is useless. On the other hand, if the small amount of other language characters don't resolve in a mainly text article there is no big loss. Or else they shouldn't be a big loss, the articles in the English wikipedia should be in English.--metta, The Sunborn 06:35, 25 Jan 2005 (UTC)

Okay, then I'm removing it from the Chinese language page. That article does not contain appreciably more special characters than any number of East Asia-related pages out there (including People's Republic of China and South Korea). -- ran (talk) 23:30, Jan 25, 2005 (UTC)

I agree. When should we apply this template? Using this in every article which contains special characters is insane, but using it only in some articles bears the question which criteria we should use. At the moment the what links here section looks rather random to me. --Conti| 20:28, Mar 24, 2005 (UTC)

One reason to have the note is so that people don't delete the special characters because they show up as ? in their browser and they don't realize that they are special chars. One thing that would help would be to not have the description on the page itself as it is now structured but instead link to a central place. This way we only clutter the page with a very brief note which if you click takes you to the explaination. WilliamKF 18:45, 28 January 2006 (UTC)

[edit] Too big for top

This template is too large cumbersome for any article, even in extreme cases. To use, add a NOTE at top (see template SpecialCharsNote, and put the SpecialChars link as a FOOTER under a HEADER called "Special characters." This at least is reasonable, and the note doesnt have to be at top, but just before the TOC. Thanks. -==SV 00:45, 25 Mar 2005 (UTC)

Template:SpecialCharsNote refers to an anchor named Special_chars, but no such anchor exists in this template anymore. I was going to put it back in, but I can't figure out how to do it. <a name="Special_chars">Note:</a> goes into the rendered version with the anchor tags visible now. — mjb 05:01, 24 August 2005 (UTC)

[edit] Producing replacement character examples, part 3

Paddu's attempts to force the display of unrenderable-character replacement characters (boxes, question marks, etc.) do not work so well across different browsers. It's actually not permitted in XHTML to include, even by reference, code points ending in FFFE and FFFF (hex). U+FFFD means "replacement character" and should always work (post-Netscape 4.0), to the extent that it will be a replacement character, but we can't guarantee its appearance, and we really want it to be unambiguous. Firefox 1.0.4 on WinXP was showing "??&#xFFFE;??" for me, which just looks like sloppy code. Even if it were 5 question marks in a row, that would still not really be a very helpful example. I don't think it's necessary to attempt to show people what the replacement characters look like; a brief description is fine, so I have rephrased the content accordingly. I also tried to phrase it in such a way as to point out that having the right fonts is a critical factor, but not the only possible point of failure. — mjb 04:37, 12 July 2005 (UTC)

[edit] New anchor suggestion

Also posted in Template talk:SpecialCharsNote.

It would be possible to specify an anchor by adding the id="????" attribute to the <div> tag in {{SpecialChars}}. This would remove the necessity of adding a special header—which often looks clunkyanyway—and ensure that anywhere {{SpecialChars}} and {{SpecialCharsNote}} are used together they will work properly. I suggest using "SpecialCharsTemplate" as the anchor text. I also suggest altering the wording slightly so that people actually expect to click for an explanation.

{{SpecialCharsNote}} would therefore contain:

:This article contains special characters: see [[#SpecialCharsTemplate|explnatory note]].

and {{SpecialChars}} would start:

<div class="boilerplate metadata" style="background-color: #f3f9ff; margin: 1em 2.5% 0 2.5%; padding: 0 10px; border: 1px solid #aaa;id="SpecialCharsTemplate">

Phil | Talk 14:09, 20 October 2005 (UTC)

[edit] Template is too big even at the bottom of an article

This template is too big for the bottom of an article, especially because it is not specific. Adding the little note at the top of the page makes the combined effect very ugly. I recommend that it is stripped down the minimum with a link to a page that explains what can be done about it. There are specific pages dealing with specific character sets, like {{IndicText}}. These are far more helpful. Also, as many pages with this template are about languages, it is possible to build a small notice into the footer of {{language}}. --Gareth Hughes 15:45, 25 November 2005 (UTC)

[edit] one more font

I started with article Runic alphabet and was unable to read it (Mozilla 1.7.5/Win2k). Thus I followed the links from this template. Code2000 is stated to be shareware (sooner or later will ask for payment), Arial Unicode MS is stated to be available only through MS Office (which I do not have), and I did not managed to get something as TTF from Free software Unicode fonts. So I was stuck with all suggestions shown here. So I gave up and browsed further slightly annoyed. I think the suggestions here need improvement.

Browsing further I came across the TITUS (project), which helped me to see the runes. IMHO it might be valuable to other people, so I wrote a stub with link to the font download page. I am going to replace "Arial Unicode MS" with a link to the TITUS article, as MS Office owners will more or less have it while others will benefit from the TITUS font. -- Goldie (tell me) 10:23, 17 December 2005 (UTC)

[edit] Hint from Chinese Wikipedia

I am not sure if "special chars" include all Asian chars or vice versa. If yes, then the box should only notify readers with this information and maybe tell them to get the fonts; if not (like math symbols or chars from other languages), then maybe separate special chars templates regarding more specific types of characters (like Asian, maths, other languages, etc.) should be made. I guess readers prefer an eye-catching but small notice, so the template is better to be smaller and variated in terms of "types". And I support Phil's suggestion of a new anchor.

In Chinese Wiki, there are two templates about Unicode font extensions: "CJKA-Char" and "CJKB-Char" (region A and B extension of Chinese, Japanese and Korean characters). The template roughly reads:

Attention: This article contains Unicode CJK Unified Ideographs extension region (A/B). If your computer does not support, the respective characters will be shown as spaces, question marks or boxes. Please refer to (another page explaining further about it- "Wikipedia: Unicode extension").

See Han unification if needed. I hope it helps... - Polobird 18:08, 16 January 2006 (UTC)

[edit] Nominated for deletion

This template was nominated for deletion on March 4, 2006. The result was keep. The archived discussion can be found here. Zzyzx11 (Talk) 00:37, 11 March 2006 (UTC)

[edit] Template trimmed, but more work to be done

On 10 Feb 2006, Stevertigo trimmed the content down to just the bare minimum. Now we have a problem in that Wikipedia:Template messages/General still contains a complaint about the size of the template, and it still suggests using the Template:SpecialCharsNote at the top of the article. Please discuss where the SpecialChars template should go (top or bottom or both?) and whether SpecialCharsNote is needed, and make adjustments to Template messages/General accordingly. Thanks. —mjb 01:51, 20 March 2006 (UTC)

[edit] So noted

I removed the "technical note:" prefix from the template on the grounds that none of the top-placed templates I've ever seen state what sort of information they are, nor is this of any interest to the reader. That a statement about "technical limitations" is a technical note should be considered obvious and uninteresting. 82.92.119.11 23:34, 16 April 2006 (UTC)

[edit] Category

What exactly do we anticipate the category being used for? This is a meta-category and should properly be on the talk page, if a category is needed at all. But I would think what links here would be just as good. Christopher Parham (talk) 23:29, 1 August 2006 (UTC)

[edit] Runes

The template that is used showing the rune rows in the article Haglaz - can someone amend it to show the Armanen Futharkh also as this is a major rune row and should not be missed off. FK0071a 11:31, 22 November 2006 (UTC)

[edit] Hideable templates

This is the sort of template that pretty much only makes sense when WP is being viewed in a web browser. It doesn't make sense for print, for example - if special characters show up as boxes, that's a printing error, and there's nothing the reader can do about it. Templates like dablinks have a CSS style that hides them in print; this one should do the same IMO. Hairy Dude 13:23, 21 February 2007 (UTC)