XML in a Nutshell, 3rd Edition

Browser Support for XHTML

Many web browsers, especially Internet Explorer 5.0 and earlier and Netscape 4.79 and earlier, deal inconsistently with XHTML. Certainly they don’t require it, accepting as they do such a wide variety of malformed, invalid, and out-and-out mistaken HTML. However, beyond that they do have some problems when they encounter certain common XHTML constructs.

The XML declaration and processing instructions

Some browsers display processing instructions and the XML declaration inline. These should be omitted if possible.

Few, if any, browsers recognize or respect the encoding declaration in the XML declaration. Furthermore, many browsers won’t automatically recognize UTF-8 or UCS-2 Unicode text. If you use a non-ASCII character set, you should also include a meta element in the header identifying the character set. For example:

<meta http-equiv="Content-type"
      content='text/html; charset=UTF-8'></meta>

Empty elements

Browsers deal inconsistently with both forms of empty element syntax. That is, some browsers understand <hr/> but not <hr></hr> (typically rendering it as two horizontal lines rather than one), while others recognize <hr></hr> but not <hr/> (typically omitting the horizontal line completely). The most consistent rendering seems to be achieved by using an empty-element tag with an optional attribute such as class or id—for example, <hr class="empty" />. There’s no real reason for the class attribute here, except that its presence keeps browsers from choking ...