Ripping out a@name generating code from the stylesheets [was:<link linkend> in XHTML output]

["Michael(tm) Smith" <smith@sideshowbarker.net>]

David Cramer <dcramer@motive.com>, 2008-10-02 09:13 -0500:

> You may also be able to get around the problem using a[href]{} in your
> css instead of just a{}: 
> http://article.gmane.org/gmane.text.docbook.apps/19015/

That reminds me: Wasn't there some discussion a while back about
the DocBook Project revising the stylesheets so that they never
output a@name at all? I think all browsers in wide use today
understand @id just fine and don't need a@name. Or is there some
other reason for keeping the a@name stuff that I'm missing?

  --Mike

> > -----Original Message-----
> > From: Michael(tm) Smith [mailto:smith@sideshowbarker.net] 
> > Sent: Thursday, October 02, 2008 2:21 AM
> > To: Andy Smith
> > Cc: DocBook Apps
> > Subject: Re: [docbook-apps] <link linkend> in XHTML output
> > 
> > Andy Smith <andy.docbook@zambezi.org.uk>, 2008-10-02 06:49 +0100:
> > 
> > > 2008/10/1 Lillian Sullam <lsullam@bruxton.com>:
> > > > 2. In every part of my chunked document the text reacts to any 
> > > > "hovering" of the mouse.  This should not happen, not all 
> > of the text is a link.
> > > [...]
> > > > I have a feeling this has to do with the <a> tag being 
> > closed with a "/"
> > > > instead of </a>, but why is this happening?
> > > 
> > > I think you're right, and you'll see the behaviour you describe if 
> > > your browser is interpreting the file as plain HTML rather 
> > than XHTML.
> > > In plain HTML the </a> closing tag is mandatory, and I guess the 
> > > browser is treating the <a.../> as an opening tag (and acting as if 
> > > there's a closing tag later in the document) rather than an empty 
> > > element.
> > 
> > Yep, because browsers don't parse text/html content as XML -- 
> > they parse it as HTML, and in the HTML syntax that browsers 
> > support, <a.../> means the same thing as <a...> (the HTML 
> > parsers in browsers pay no attention to that slash).
> > 
> > That's just one of several reasons why it's not a good idea 
> > to serve XSLT-generated output@method=xml content as text/html.
> > 
> > [...]
> > > If you use the stylesheets' HTML rather than XHTML output I think 
> > > you'll get closing tags and results that will generally be 
> > OK with the 
> > > text/html content type.
> > 
> > I think that's the best advice in general. For most cases, 
> > there is zero harm in generating HTML content from XSLT using 
> > output@method=html and much good in that it will prevent the 
> > <a.../> problem, and others too.
> > 
> >   --Mike
> > 

-- 
Michael(tm) Smith
http://people.w3.org/mike/