Getting the HTML encoding declaration i XML output

["Leif Halvard Silli" <verktystell@russisk.no>]

Hello. Back in 2009, Michael Leslie asked the list (but received no 
answer) the following: [1]

   «Does anyone have any experience generating UTF-8 XHTML
   that can be consistently rendered in both Firefox and IE?»

And, like him, I want to use Docbook to produce HTML-compatible XHTML. 
However, as a (former) member of the HTML working group (and co-editor 
of a spec for polyglot markup - that is: XHTML that is HTML as well), I 
can say that the question has (since) been answered by the HTML5.x 
specifications: HTML-compatible XHTML(5) documents MUST NOT include the 
XML declaration, and they MUST be UTF-8 encoded, and the encoding must 
be declared using either the HTTP header, the Byte-order mark or the 
HTML encoding declaration. The latter - the HTML encoding declaration - 
comes in two variants:

 1) <meta charset="UTF-8"/>
 2) <meta http-equiv="Content-Type" content="text/html;charset=UTF-8/">

Both works equally well in Web browsers, but occationally there are some 
fringe, legacy implementations that only support the http-equiv variant.

The Docbook XSL book does also try to explain encoding issue of HTML and 
XHTML.[2] See chapter on ’Special characters’ under the heading 
«HTML encoding». However, the book fails to nail the solution that 
HTML5.x specifies.

Further more, it is (probably) well known that when the output mode of 
Docbook XSL is set to 'xml', then, by default, the HTML encoding 
declaration is not included. As a result, browsing a Docbook 
XSL-generated XHTML-file as text/html fails (e.g. by adding .html 
instead .xhtml), Web browsers receive no encoding declaration from the 
HTML document itself.

Hence, I propose that in next version of Docbook XSL, you allow the HTML 
encoding declaration (both variants) to be used. In fact, it would be 
best if, by default, the HTML encoding declaration always is included.

To solve my own problem, I have created the following customization 
(that I use with XMLmind XML editor), see below. If there is better/more 
generic way to do it, I would be thankful for your help (for instance, I 
am not sure why, in my iplementation, I had to include the namespace 
declaration - I’m sure that could have been avoided - anyway, it is 
excluded in the final output so it does not matter.)

<?xml version="1.0" encoding="UTF-8"?>
<?stylesheet-label Polyglott! ?>
<xsl:stylesheet
xmlns:xsl="http://www.w3.org/1999/XSL/Transform";
version="1.0">
  <xsl:import href="docbook5-config:xsl/xhtml5/docbook.xsl"/>
  <!-- REDEFINE PARAMETERS AND ATTRIBUTE-SETS HERE -->
  <xsl:output indent="yes" method="xml" omit-xml-declaration="yes"/>
<!--
     From Bob Stayton:
https://lists.oasis-open.org/archives/docbook-apps/201702/msg00031.html
-->
  <xsl:template name="root.attributes">
    <xsl:call-template name="xml.language.attribute"/>
    <xsl:call-template name="language.attribute"/>
  </xsl:template>
<!--
     Leif: Add the HTML encoding declaration
-->
  <xsl:template name="user.head.content">
    <meta
     xmlns="http://www.w3.org/1999/xhtml";
http-equiv="Content-Type"
   content="text/html;charset=UTF-8"
    />
  </xsl:template>
</xsl:stylesheet>

Btw - and not to stamp on too many toes, but: I had a look at how the 
TEI xsl sheets works, and they seem to have taken care of the issue: 
They output their HTML as XML but without the XML declaration.[3] And 
they include the HTML encoding declaration in both their HTML outputs as 
well as their Epub3 output - which seems very wise.[4] I hope that 
Docbook XSL follows the same lead. In fact, to me Docbook XSL’s html 
output mode seems like a waste of time. Better to simply produce 
HTML-compatible XML output.

[1] https://lists.oasis-open.org/archives/docbook-apps/200902/msg00099.html
[2] http://www.sagehill.net/docbookxsl/SpecialChars.html
[3] 
http://www.tei-c.org/release/doc/tei-xsl/profiles/default/html/to0.html#bt_src_O_S_to.xsl
[4] http://www.tei-c.org/release/doc/tei-xsl/profiles/default/html/to4.html#bt_src_T_metaHTMLS_......htmlhtml_param.xsl
--
leif halvard silli