138

This has been confusing me for some time. With the advent of UTF-8 as the de-facto standard in web development I'm not sure in which situations I'm supposed to use the HTML entities and for which ones should I just use the UTF-8 character. For example,

  • em dash (–, &emdash;)
  • ampersand (&, &)
  • 3/4 fraction (¾, ¾)

Please do shed light on this issue. It will be appreciated.

Matt
  • 9,068
  • 12
  • 64
  • 84
allesklar
  • 9,506
  • 6
  • 36
  • 53

8 Answers8

130

Based on the comments I have received, I looked into this a little further. It seems that currently the best practice is to forgo using HTML entities and use the actual UTF-8 character instead. The reasons listed are as follows:

  1. UTF-8 encodings are easier to read and edit for those who understand what the character means and know how to type it.
  2. UTF-8 encodings are just as unintelligible as HTML entity encodings for those who don't understand them, but they have the advantage of rendering as special characters rather than hard to understand decimal or hex encodings.

As long as your page's encoding is properly set to UTF-8, you should use the actual character instead of an HTML entity. I read several documents about this topic, but the most helpful were:

From the UTF-8: The Secret of Character Encoding article:

Wikipedia is a great case study for an application that originally used ISO-8859-1 but switched to UTF-8 when it became far too cumbersome to support foreign languages. Bots will now actually go through articles and convert character entities to their corresponding real characters for the sake of user-friendliness and searchability.

That article also gives a nice example involving Chinese encoding. Here is the abbreviated example for the sake of laziness:

UTF-8:

這兩個字是甚麼意思

HTML Entities:

這兩個字是甚麼意思

The UTF-8 and HTML entity encodings are both meaningless to me, but at least the UTF-8 encoding is recognizable as a foreign language, and it will render properly in an edit box. The article goes on to say the following about the HTML entity-encoded version:

Extremely inconvenient for those of us who actually know what character entities are, totally unintelligible to poor users who don't! Even the slightly more user-friendly, "intelligible" character entities like θ will leave users who are uninterested in learning HTML scratching their heads. On the other hand, if they see θ in an edit box, they'll know that it's a special character, and treat it accordingly, even if they don't know how to write that character themselves.

As others have noted, you still have to use HTML entities for reserved XML characters (ampersand, less-than, greater-than).

Martin
  • 22,212
  • 11
  • 70
  • 132
William Brendel
  • 31,712
  • 14
  • 72
  • 77
  • This answer helps tremendously. But to clarify, for my own understanding: there's nothing risky or invalid about using `&entity;` syntax within an HTML document with a declared UTF-8 charset, correct? While plain UTF-8 characters are better for the reasons you've listed, there's no issue still having some HTML entities alongside them in the same document? – Jacob Ford Jun 18 '16 at 15:54
  • @JacobFord Right, mixing HTML entities with UTF-8 characters is not risky or invalid, just potentially confusing to someone reading the source. – William Brendel Jul 13 '16 at 20:28
  • I agree that it can be easier to type `¾` than to find the ¾ character in some unicode lookup table somewhere. – pbarney May 11 '23 at 18:56
96

You don't generally need to use HTML character entities if your editor supports Unicode. Entities can be useful when:

  • Your keyboard does not support the character you need to type. For example, many keyboards do not have em-dash or the copyright symbol.
  • Your editor does not support Unicode (very common some years ago, but probably not today).
  • You want to make it explicit in the source what is happening. For example, the   code is clearer than the corresponding white space character.
  • You need to escape HTML special characters like <, &, or ".
Jonathan Leffler
  • 730,956
  • 141
  • 904
  • 1,278
JacquesB
  • 41,662
  • 13
  • 71
  • 86
  • 1
    Very helpful. Thanks. I use a helpful program to get unusual characters. It's called popchar and is made by Ergonis but is only for OS X. – allesklar Jan 12 '09 at 19:52
  • 3
    Side note: Wikipedia still mandates ` ` instead of the actual whitespace character, partly because Firefox converts U+00A0 into U+0020 in forms. So using the entity in that case is the only way to ensure that the source doesn't get messed up every time a Firefox user edits it. – Joey Aug 07 '10 at 12:34
  • 3
    A fine summary, but regarding the last point, it’s `<` that often needs escaping, never `>` (and `"` needs escaping only rarely inside attribute values). – Jukka K. Korpela May 14 '13 at 19:38
  • Another reason to keep   is so that you can display multiple spaces on an HTML page. – zylstra Oct 24 '16 at 18:18
  • So `&` should always be used instead of `&`? Is there a reason for this? – Hashim Aziz May 30 '19 at 01:06
  • @Hashim: Otherwise an unescaped `&` would be ambiguous since it could either be a literal `&` or the start of an entity. In some cases it it not ambiguous, e.g. if the & is immediately followed by whitespace, but it is safer to just encode it always. – JacquesB May 30 '19 at 10:16
5

Entities may buy you some compatibility with brain-dead clients that don't understand encodings correctly. I don't believe that includes any current browsers, but you never know what other kinds of programs might be hitting you up.

More useful, though, is that HTML entities protect you from your own errors: if you misconfigure something on the server and you end up serving a page with an HTTP header that says it's ISO-8859-1 and a META tag that says it's UTF-8, at least your &mdash;es will always work.

Jim Puls
  • 79,175
  • 10
  • 73
  • 78
  • 6
    You could make the opposite argument though - `&mdash`es showing up correctly even if the headers are misconfigured makes it *harder* to detect there's a problem. – Pekka Mar 05 '13 at 20:00
5

I would not use UTF-8 for characters that are easily confused visually. For example, it is difficult to distinguish an emdash from a minus, or especially a non-breaking space from a space. For these characters, definitely use entities.

For characters that are easily understood visually (such as the chinese examples above), go ahead and use UTF-8 if you like.

Ned Batchelder
  • 364,293
  • 75
  • 561
  • 662
5

Personally I do everything in utf-8 since a long time, however, in an html page, you always need to convert ampersands (&), greater than (>) and lesser then (<) characters to their equivalent entities, &amp;, &gt; and &lt;

Also, if you intend on doing some programming using utf-8 text, there are a few thing to watch for.

  • XML needs some extra lines to validate when using entities.
  • Some libraries do not play along nice with utf-8. For instance, PHP in some Linux distributions dropped full support for utf-8 in their regular expression libraries.
  • It is harder to limit the number of characters in a text that uses html entities, because a single entity uses many characters. Also there's always the risk of cutting the entity in half.
Marco Luglio
  • 3,503
  • 1
  • 21
  • 27
4

HTML entities are useful when you want to generate content that is going to be included (dynamically) into pages with (several) different encodings. For example, we have white label content that is included both into ISO-8859-1 and UTF-8 encoded web pages...

If character set conversion from/to UTF-8 wasn't such a big unreliable mess (you always stumble over some characters and some tools that don't convert properly), standardizing on UTF-8 would be the way to go.

mjy
  • 2,727
  • 2
  • 20
  • 22
2

If your pages are correctly encoded in utf-8 you should have no need for html entities, just use the characters you want directly.

Otávio Décio
  • 73,752
  • 17
  • 161
  • 228
2

All of the previous answers make sense to me.

In addition: It mostly depends on the editor you intent to use and the document language. As a minimum requirement for the editor is that it supports the document language. That means, that if your text is in japanese, beware of using an editor which does not show them (i.e. no entities for the document itself). If its english, you can even use an old vim-like editor and use entities only for the relative seldom &copy; and friends. Of course: &gt; for > and other HTML-specials still need escapes. But even with the other latin-1 languages (german, french etc.) writing ä is a pain in you know where...

In addition, I personally write entities for invisible characters and those which are looking similar to standard-ascii and are therefore easily confused. For example, there is u1173 (looking like a dash in some charsets) or u1175, which looks like the vertical bar. I'd use entities for those in any case.

blabla999
  • 3,130
  • 22
  • 24